Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Mar 3, 2012, at 00:31 , Jan Lehnardt wrote: Good catch, will do in the morning. Done. Cheers Jan -- On 02.03.2012, at 20:00, Noah Slater nsla...@tumbolia.org wrote: Have you updated NEWS and CHANGES? On Fri, Mar 2, 2012 at 5:47 PM, Jan Lehnardt j...@apache.org wrote: On Mar 2, 2012, at 17:28 , Noah Slater wrote: Vote aborted. Thanks to everyone who took part. Five people have claimed ownership of the five remaining tasks: * Bob Dionne will driving COUCHDB-1424 * Benoît Chesneau will be driving COUCHDB-1426 * Jan Lehnardt will be driving the R15B patch √ http://git-wip-us.apache.org/repos/asf?p=couchdb.git;a=commit;h=776aa1d0 Landed in both 1.2.x and master. Thanks to Paul Davis for coming up with the more canonical C version. Cheers Jan -- * Robert Newson will be driving the performance work * I will be driving round three, obviously Let's kick some ass! On Fri, Mar 2, 2012 at 4:05 PM, Robert Newson rnew...@apache.org wrote: +1. Abort this, get the fixes we have in and start round 3. The performance regression issue is elusive, we might, out of necessity, have to defer action on that until post 1.2.0. On 2 March 2012 16:02, Noah Slater nsla...@tumbolia.org wrote: I think we should, at a minimum: * Abort this round * Land the R15B patch * Land COUCHDB-1426 (which seems easy) * Start round three I think we should try to: * Try to land COUCHDB-1424 * Get clarification on the performance issues For these last two items, I think we should impose a time limit. Let's say a week. I think we should also form some teams, to see if we can do some sprints to get these issues fixed. On Fri, Mar 2, 2012 at 3:49 PM, Jan Lehnardt j...@apache.org wrote: On Mar 2, 2012, at 16:47 , Dirkjan Ochtman wrote: On Fri, Mar 2, 2012 at 16:29, Jan Lehnardt j...@apache.org wrote: Proposed Action: I'd propose to release 1.2.0 as-is with the following points mentioned in the release notes (the exact wording of which is to be done): 1. Note that this release is incompatible with Erlang R15B. A patch is available at [LINK to DIFF]; it will appear in Apache CouchDB 1.2.1. 2. Also note that there are some reports of a performance regression in view building. While initial and ad-hoc tests showed an improvement in most cases, we'd like to ask our users to report any significant differences to the Apache CouchDB 1.1.1 release. While I am usually cheering on the release process, taken together it seems more prudent to abort this round, take the R15B driver and the 142{4,6} patches and then starting a new round. I'd agree if 142{4,6} were decided tickets, but they are still ongoing and potentially, in the 24-case especially, for a while. Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Thanks! On Sat, Mar 3, 2012 at 1:04 PM, Jan Lehnardt j...@apache.org wrote: On Mar 3, 2012, at 00:31 , Jan Lehnardt wrote: Good catch, will do in the morning. Done. Cheers Jan -- On 02.03.2012, at 20:00, Noah Slater nsla...@tumbolia.org wrote: Have you updated NEWS and CHANGES? On Fri, Mar 2, 2012 at 5:47 PM, Jan Lehnardt j...@apache.org wrote: On Mar 2, 2012, at 17:28 , Noah Slater wrote: Vote aborted. Thanks to everyone who took part. Five people have claimed ownership of the five remaining tasks: * Bob Dionne will driving COUCHDB-1424 * Benoît Chesneau will be driving COUCHDB-1426 * Jan Lehnardt will be driving the R15B patch √ http://git-wip-us.apache.org/repos/asf?p=couchdb.git;a=commit;h=776aa1d0 Landed in both 1.2.x and master. Thanks to Paul Davis for coming up with the more canonical C version. Cheers Jan -- * Robert Newson will be driving the performance work * I will be driving round three, obviously Let's kick some ass! On Fri, Mar 2, 2012 at 4:05 PM, Robert Newson rnew...@apache.org wrote: +1. Abort this, get the fixes we have in and start round 3. The performance regression issue is elusive, we might, out of necessity, have to defer action on that until post 1.2.0. On 2 March 2012 16:02, Noah Slater nsla...@tumbolia.org wrote: I think we should, at a minimum: * Abort this round * Land the R15B patch * Land COUCHDB-1426 (which seems easy) * Start round three I think we should try to: * Try to land COUCHDB-1424 * Get clarification on the performance issues For these last two items, I think we should impose a time limit. Let's say a week. I think we should also form some teams, to see if we can do some sprints to get these issues fixed. On Fri, Mar 2, 2012 at 3:49 PM, Jan Lehnardt j...@apache.org wrote: On Mar 2, 2012, at 16:47 , Dirkjan Ochtman wrote: On Fri, Mar 2, 2012 at 16:29, Jan Lehnardt j...@apache.org wrote: Proposed Action: I'd propose to release 1.2.0 as-is with the following points mentioned in the release notes (the exact wording of which is to be done): 1. Note that this release is incompatible with Erlang R15B. A patch is available at [LINK to DIFF]; it will appear in Apache CouchDB 1.2.1. 2. Also note that there are some reports of a performance regression in view building. While initial and ad-hoc tests showed an improvement in most cases, we'd like to ask our users to report any significant differences to the Apache CouchDB 1.1.1 release. While I am usually cheering on the release process, taken together it seems more prudent to abort this round, take the R15B driver and the 142{4,6} patches and then starting a new round. I'd agree if 142{4,6} were decided tickets, but they are still ongoing and potentially, in the 24-case especially, for a while. Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Thu, Mar 1, 2012 at 12:21 AM, Benoit Chesneau bchesn...@gmail.com wrote: Posted a patch fixing COUCHDB-1426. https://issues.apache.org/jira/browse/COUCHDB-1426 After some discussions, I made the issue none blocking but critical. I'm now +0 I guess. - benoît
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Thanks Benoit! * * * As an update to the ongoing 1.2.0 release these issues were raised: - Performance regression (A thread on dev@) - Multiple Spidermonkey version detection (COUCHDB-1426) - R15B icu_driver compatibility (git branch R15B0-driver) - R15B make check hang on Mac OS X Lion (COUCHDB-1424) I'd like to give a quick update on each of them and I'll propose a plan of action. If you disagree with any of my updates or the conclusion, please speak up :) * Performance Regression We started collecting a bunch of numbers for a few select scenarios, some of which show a regression, most of which show an improvement though. We haven't yet identified the variables that cause all this (different Spidermonkey releases made a huge difference in one case). So far the test cases don't cover much ground. E.g. we haven't had a chance to reproduce one of the reports that we don't have direct access too. I believe getting a better understanding of what is going on here will take some more time. Time that we should absolutely spend, so we are able to show some baseline performance numbers going forward to compare improvements and future releases against. * COUCHDB-1426 We are still discussing the intention and merits of the patch and I believe there's definitely something to fix eventually, but I don't think this should hold 1.2.0. Benoit, who reported it initially and made it blocking, agrees under the condition that we ship 1.2.1 soonish that includes a resolution for this ticket. * R15B0-driver This renders 1.2.0 as is incompatible with the latest Erlang R15B release. The fix is trivial, but it didn't make the release artefact. * COUCHDB-1424 I so far could show that this only happens on Mac OS X Lion and R15B but nowhere else. I'm happy to conclude that this isn't a release blocking issue. However, we should resolve it for the 1.2.1 release like COUCHDB-1426. * * * Proposed Action: I'd propose to release 1.2.0 as-is with the following points mentioned in the release notes (the exact wording of which is to be done): 1. Note that this release is incompatible with Erlang R15B. A patch is available at [LINK to DIFF]; it will appear in Apache CouchDB 1.2.1. 2. Also note that there are some reports of a performance regression in view building. While initial and ad-hoc tests showed an improvement in most cases, we'd like to ask our users to report any significant differences to the Apache CouchDB 1.1.1 release. * * * I'm happy to alter the plan of action based on your feedback, I just want to make sure we know which issues we are discussing and what their respective states are to move this forward. * * * https://issues.apache.org/jira/browse/COUCHDB-1424 https://issues.apache.org/jira/browse/COUCHDB-1426 http://git-wip-us.apache.org/repos/asf?p=couchdb.git;a=commit;h=b1af764b Cheers Jan -- On Mar 2, 2012, at 13:48 , Benoit Chesneau wrote: On Thu, Mar 1, 2012 at 12:21 AM, Benoit Chesneau bchesn...@gmail.com wrote: Posted a patch fixing COUCHDB-1426. https://issues.apache.org/jira/browse/COUCHDB-1426 After some discussions, I made the issue none blocking but critical. I'm now +0 I guess. - benoît
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
+1 On 2 March 2012 15:29, Jan Lehnardt j...@apache.org wrote: Thanks Benoit! * * * As an update to the ongoing 1.2.0 release these issues were raised: - Performance regression (A thread on dev@) - Multiple Spidermonkey version detection (COUCHDB-1426) - R15B icu_driver compatibility (git branch R15B0-driver) - R15B make check hang on Mac OS X Lion (COUCHDB-1424) I'd like to give a quick update on each of them and I'll propose a plan of action. If you disagree with any of my updates or the conclusion, please speak up :) * Performance Regression We started collecting a bunch of numbers for a few select scenarios, some of which show a regression, most of which show an improvement though. We haven't yet identified the variables that cause all this (different Spidermonkey releases made a huge difference in one case). So far the test cases don't cover much ground. E.g. we haven't had a chance to reproduce one of the reports that we don't have direct access too. I believe getting a better understanding of what is going on here will take some more time. Time that we should absolutely spend, so we are able to show some baseline performance numbers going forward to compare improvements and future releases against. * COUCHDB-1426 We are still discussing the intention and merits of the patch and I believe there's definitely something to fix eventually, but I don't think this should hold 1.2.0. Benoit, who reported it initially and made it blocking, agrees under the condition that we ship 1.2.1 soonish that includes a resolution for this ticket. * R15B0-driver This renders 1.2.0 as is incompatible with the latest Erlang R15B release. The fix is trivial, but it didn't make the release artefact. * COUCHDB-1424 I so far could show that this only happens on Mac OS X Lion and R15B but nowhere else. I'm happy to conclude that this isn't a release blocking issue. However, we should resolve it for the 1.2.1 release like COUCHDB-1426. * * * Proposed Action: I'd propose to release 1.2.0 as-is with the following points mentioned in the release notes (the exact wording of which is to be done): 1. Note that this release is incompatible with Erlang R15B. A patch is available at [LINK to DIFF]; it will appear in Apache CouchDB 1.2.1. 2. Also note that there are some reports of a performance regression in view building. While initial and ad-hoc tests showed an improvement in most cases, we'd like to ask our users to report any significant differences to the Apache CouchDB 1.1.1 release. * * * I'm happy to alter the plan of action based on your feedback, I just want to make sure we know which issues we are discussing and what their respective states are to move this forward. * * * https://issues.apache.org/jira/browse/COUCHDB-1424 https://issues.apache.org/jira/browse/COUCHDB-1426 http://git-wip-us.apache.org/repos/asf?p=couchdb.git;a=commit;h=b1af764b Cheers Jan -- On Mar 2, 2012, at 13:48 , Benoit Chesneau wrote: On Thu, Mar 1, 2012 at 12:21 AM, Benoit Chesneau bchesn...@gmail.com wrote: Posted a patch fixing COUCHDB-1426. https://issues.apache.org/jira/browse/COUCHDB-1426 After some discussions, I made the issue none blocking but critical. I'm now +0 I guess. - benoît
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Fri, Mar 2, 2012 at 16:29, Jan Lehnardt j...@apache.org wrote: Proposed Action: I'd propose to release 1.2.0 as-is with the following points mentioned in the release notes (the exact wording of which is to be done): 1. Note that this release is incompatible with Erlang R15B. A patch is available at [LINK to DIFF]; it will appear in Apache CouchDB 1.2.1. 2. Also note that there are some reports of a performance regression in view building. While initial and ad-hoc tests showed an improvement in most cases, we'd like to ask our users to report any significant differences to the Apache CouchDB 1.1.1 release. While I am usually cheering on the release process, taken together it seems more prudent to abort this round, take the R15B driver and the 142{4,6} patches and then starting a new round. Cheers, Dirkjan
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Mar 2, 2012, at 16:47 , Dirkjan Ochtman wrote: On Fri, Mar 2, 2012 at 16:29, Jan Lehnardt j...@apache.org wrote: Proposed Action: I'd propose to release 1.2.0 as-is with the following points mentioned in the release notes (the exact wording of which is to be done): 1. Note that this release is incompatible with Erlang R15B. A patch is available at [LINK to DIFF]; it will appear in Apache CouchDB 1.2.1. 2. Also note that there are some reports of a performance regression in view building. While initial and ad-hoc tests showed an improvement in most cases, we'd like to ask our users to report any significant differences to the Apache CouchDB 1.1.1 release. While I am usually cheering on the release process, taken together it seems more prudent to abort this round, take the R15B driver and the 142{4,6} patches and then starting a new round. I'd agree if 142{4,6} were decided tickets, but they are still ongoing and potentially, in the 24-case especially, for a while. Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
I think we should, at a minimum: * Abort this round * Land the R15B patch * Land COUCHDB-1426 (which seems easy) * Start round three I think we should try to: * Try to land COUCHDB-1424 * Get clarification on the performance issues For these last two items, I think we should impose a time limit. Let's say a week. I think we should also form some teams, to see if we can do some sprints to get these issues fixed. On Fri, Mar 2, 2012 at 3:49 PM, Jan Lehnardt j...@apache.org wrote: On Mar 2, 2012, at 16:47 , Dirkjan Ochtman wrote: On Fri, Mar 2, 2012 at 16:29, Jan Lehnardt j...@apache.org wrote: Proposed Action: I'd propose to release 1.2.0 as-is with the following points mentioned in the release notes (the exact wording of which is to be done): 1. Note that this release is incompatible with Erlang R15B. A patch is available at [LINK to DIFF]; it will appear in Apache CouchDB 1.2.1. 2. Also note that there are some reports of a performance regression in view building. While initial and ad-hoc tests showed an improvement in most cases, we'd like to ask our users to report any significant differences to the Apache CouchDB 1.1.1 release. While I am usually cheering on the release process, taken together it seems more prudent to abort this round, take the R15B driver and the 142{4,6} patches and then starting a new round. I'd agree if 142{4,6} were decided tickets, but they are still ongoing and potentially, in the 24-case especially, for a while. Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Fri, Mar 2, 2012 at 17:02, Noah Slater nsla...@tumbolia.org wrote: I think we should, at a minimum: * Abort this round * Land the R15B patch * Land COUCHDB-1426 (which seems easy) * Start round three I think we should try to: * Try to land COUCHDB-1424 * Get clarification on the performance issues For these last two items, I think we should impose a time limit. Let's say a week. I think we should also form some teams, to see if we can do some sprints to get these issues fixed. +1 this is closer to what I feel would be the right thing. Cheers, Dirkjan
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
+1. Abort this, get the fixes we have in and start round 3. The performance regression issue is elusive, we might, out of necessity, have to defer action on that until post 1.2.0. On 2 March 2012 16:02, Noah Slater nsla...@tumbolia.org wrote: I think we should, at a minimum: * Abort this round * Land the R15B patch * Land COUCHDB-1426 (which seems easy) * Start round three I think we should try to: * Try to land COUCHDB-1424 * Get clarification on the performance issues For these last two items, I think we should impose a time limit. Let's say a week. I think we should also form some teams, to see if we can do some sprints to get these issues fixed. On Fri, Mar 2, 2012 at 3:49 PM, Jan Lehnardt j...@apache.org wrote: On Mar 2, 2012, at 16:47 , Dirkjan Ochtman wrote: On Fri, Mar 2, 2012 at 16:29, Jan Lehnardt j...@apache.org wrote: Proposed Action: I'd propose to release 1.2.0 as-is with the following points mentioned in the release notes (the exact wording of which is to be done): 1. Note that this release is incompatible with Erlang R15B. A patch is available at [LINK to DIFF]; it will appear in Apache CouchDB 1.2.1. 2. Also note that there are some reports of a performance regression in view building. While initial and ad-hoc tests showed an improvement in most cases, we'd like to ask our users to report any significant differences to the Apache CouchDB 1.1.1 release. While I am usually cheering on the release process, taken together it seems more prudent to abort this round, take the R15B driver and the 142{4,6} patches and then starting a new round. I'd agree if 142{4,6} were decided tickets, but they are still ongoing and potentially, in the 24-case especially, for a while. Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Mar 2, 2012, at 17:02 , Noah Slater wrote: I think we should, at a minimum: * Abort this round * Land the R15B patch * Land COUCHDB-1426 (which seems easy) * Start round three I think we should try to: * Try to land COUCHDB-1424 * Get clarification on the performance issues For these last two items, I think we should impose a time limit. Let's say a week. I think we should also form some teams, to see if we can do some sprints to get these issues fixed. Sounds like a plan, let's go! On Fri, Mar 2, 2012 at 3:49 PM, Jan Lehnardt j...@apache.org wrote: On Mar 2, 2012, at 16:47 , Dirkjan Ochtman wrote: On Fri, Mar 2, 2012 at 16:29, Jan Lehnardt j...@apache.org wrote: Proposed Action: I'd propose to release 1.2.0 as-is with the following points mentioned in the release notes (the exact wording of which is to be done): 1. Note that this release is incompatible with Erlang R15B. A patch is available at [LINK to DIFF]; it will appear in Apache CouchDB 1.2.1. 2. Also note that there are some reports of a performance regression in view building. While initial and ad-hoc tests showed an improvement in most cases, we'd like to ask our users to report any significant differences to the Apache CouchDB 1.1.1 release. While I am usually cheering on the release process, taken together it seems more prudent to abort this round, take the R15B driver and the 142{4,6} patches and then starting a new round. I'd agree if 142{4,6} were decided tickets, but they are still ongoing and potentially, in the 24-case especially, for a while. Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Vote aborted. Thanks to everyone who took part. Five people have claimed ownership of the five remaining tasks: * Bob Dionne will driving COUCHDB-1424 * Benoît Chesneau will be driving COUCHDB-1426 * Jan Lehnardt will be driving the R15B patch * Robert Newson will be driving the performance work * I will be driving round three, obviously Let's kick some ass! On Fri, Mar 2, 2012 at 4:05 PM, Robert Newson rnew...@apache.org wrote: +1. Abort this, get the fixes we have in and start round 3. The performance regression issue is elusive, we might, out of necessity, have to defer action on that until post 1.2.0. On 2 March 2012 16:02, Noah Slater nsla...@tumbolia.org wrote: I think we should, at a minimum: * Abort this round * Land the R15B patch * Land COUCHDB-1426 (which seems easy) * Start round three I think we should try to: * Try to land COUCHDB-1424 * Get clarification on the performance issues For these last two items, I think we should impose a time limit. Let's say a week. I think we should also form some teams, to see if we can do some sprints to get these issues fixed. On Fri, Mar 2, 2012 at 3:49 PM, Jan Lehnardt j...@apache.org wrote: On Mar 2, 2012, at 16:47 , Dirkjan Ochtman wrote: On Fri, Mar 2, 2012 at 16:29, Jan Lehnardt j...@apache.org wrote: Proposed Action: I'd propose to release 1.2.0 as-is with the following points mentioned in the release notes (the exact wording of which is to be done): 1. Note that this release is incompatible with Erlang R15B. A patch is available at [LINK to DIFF]; it will appear in Apache CouchDB 1.2.1. 2. Also note that there are some reports of a performance regression in view building. While initial and ad-hoc tests showed an improvement in most cases, we'd like to ask our users to report any significant differences to the Apache CouchDB 1.1.1 release. While I am usually cheering on the release process, taken together it seems more prudent to abort this round, take the R15B driver and the 142{4,6} patches and then starting a new round. I'd agree if 142{4,6} were decided tickets, but they are still ongoing and potentially, in the 24-case especially, for a while. Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Mar 2, 2012, at 17:28 , Noah Slater wrote: Vote aborted. Thanks to everyone who took part. Five people have claimed ownership of the five remaining tasks: * Bob Dionne will driving COUCHDB-1424 * Benoît Chesneau will be driving COUCHDB-1426 * Jan Lehnardt will be driving the R15B patch √ http://git-wip-us.apache.org/repos/asf?p=couchdb.git;a=commit;h=776aa1d0 Landed in both 1.2.x and master. Thanks to Paul Davis for coming up with the more canonical C version. Cheers Jan -- * Robert Newson will be driving the performance work * I will be driving round three, obviously Let's kick some ass! On Fri, Mar 2, 2012 at 4:05 PM, Robert Newson rnew...@apache.org wrote: +1. Abort this, get the fixes we have in and start round 3. The performance regression issue is elusive, we might, out of necessity, have to defer action on that until post 1.2.0. On 2 March 2012 16:02, Noah Slater nsla...@tumbolia.org wrote: I think we should, at a minimum: * Abort this round * Land the R15B patch * Land COUCHDB-1426 (which seems easy) * Start round three I think we should try to: * Try to land COUCHDB-1424 * Get clarification on the performance issues For these last two items, I think we should impose a time limit. Let's say a week. I think we should also form some teams, to see if we can do some sprints to get these issues fixed. On Fri, Mar 2, 2012 at 3:49 PM, Jan Lehnardt j...@apache.org wrote: On Mar 2, 2012, at 16:47 , Dirkjan Ochtman wrote: On Fri, Mar 2, 2012 at 16:29, Jan Lehnardt j...@apache.org wrote: Proposed Action: I'd propose to release 1.2.0 as-is with the following points mentioned in the release notes (the exact wording of which is to be done): 1. Note that this release is incompatible with Erlang R15B. A patch is available at [LINK to DIFF]; it will appear in Apache CouchDB 1.2.1. 2. Also note that there are some reports of a performance regression in view building. While initial and ad-hoc tests showed an improvement in most cases, we'd like to ask our users to report any significant differences to the Apache CouchDB 1.1.1 release. While I am usually cheering on the release process, taken together it seems more prudent to abort this round, take the R15B driver and the 142{4,6} patches and then starting a new round. I'd agree if 142{4,6} were decided tickets, but they are still ongoing and potentially, in the 24-case especially, for a while. Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Have you updated NEWS and CHANGES? On Fri, Mar 2, 2012 at 5:47 PM, Jan Lehnardt j...@apache.org wrote: On Mar 2, 2012, at 17:28 , Noah Slater wrote: Vote aborted. Thanks to everyone who took part. Five people have claimed ownership of the five remaining tasks: * Bob Dionne will driving COUCHDB-1424 * Benoît Chesneau will be driving COUCHDB-1426 * Jan Lehnardt will be driving the R15B patch √ http://git-wip-us.apache.org/repos/asf?p=couchdb.git;a=commit;h=776aa1d0 Landed in both 1.2.x and master. Thanks to Paul Davis for coming up with the more canonical C version. Cheers Jan -- * Robert Newson will be driving the performance work * I will be driving round three, obviously Let's kick some ass! On Fri, Mar 2, 2012 at 4:05 PM, Robert Newson rnew...@apache.org wrote: +1. Abort this, get the fixes we have in and start round 3. The performance regression issue is elusive, we might, out of necessity, have to defer action on that until post 1.2.0. On 2 March 2012 16:02, Noah Slater nsla...@tumbolia.org wrote: I think we should, at a minimum: * Abort this round * Land the R15B patch * Land COUCHDB-1426 (which seems easy) * Start round three I think we should try to: * Try to land COUCHDB-1424 * Get clarification on the performance issues For these last two items, I think we should impose a time limit. Let's say a week. I think we should also form some teams, to see if we can do some sprints to get these issues fixed. On Fri, Mar 2, 2012 at 3:49 PM, Jan Lehnardt j...@apache.org wrote: On Mar 2, 2012, at 16:47 , Dirkjan Ochtman wrote: On Fri, Mar 2, 2012 at 16:29, Jan Lehnardt j...@apache.org wrote: Proposed Action: I'd propose to release 1.2.0 as-is with the following points mentioned in the release notes (the exact wording of which is to be done): 1. Note that this release is incompatible with Erlang R15B. A patch is available at [LINK to DIFF]; it will appear in Apache CouchDB 1.2.1. 2. Also note that there are some reports of a performance regression in view building. While initial and ad-hoc tests showed an improvement in most cases, we'd like to ask our users to report any significant differences to the Apache CouchDB 1.1.1 release. While I am usually cheering on the release process, taken together it seems more prudent to abort this round, take the R15B driver and the 142{4,6} patches and then starting a new round. I'd agree if 142{4,6} were decided tickets, but they are still ongoing and potentially, in the 24-case especially, for a while. Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Good catch, will do in the morning. Cheers Jan -- On 02.03.2012, at 20:00, Noah Slater nsla...@tumbolia.org wrote: Have you updated NEWS and CHANGES? On Fri, Mar 2, 2012 at 5:47 PM, Jan Lehnardt j...@apache.org wrote: On Mar 2, 2012, at 17:28 , Noah Slater wrote: Vote aborted. Thanks to everyone who took part. Five people have claimed ownership of the five remaining tasks: * Bob Dionne will driving COUCHDB-1424 * Benoît Chesneau will be driving COUCHDB-1426 * Jan Lehnardt will be driving the R15B patch √ http://git-wip-us.apache.org/repos/asf?p=couchdb.git;a=commit;h=776aa1d0 Landed in both 1.2.x and master. Thanks to Paul Davis for coming up with the more canonical C version. Cheers Jan -- * Robert Newson will be driving the performance work * I will be driving round three, obviously Let's kick some ass! On Fri, Mar 2, 2012 at 4:05 PM, Robert Newson rnew...@apache.org wrote: +1. Abort this, get the fixes we have in and start round 3. The performance regression issue is elusive, we might, out of necessity, have to defer action on that until post 1.2.0. On 2 March 2012 16:02, Noah Slater nsla...@tumbolia.org wrote: I think we should, at a minimum: * Abort this round * Land the R15B patch * Land COUCHDB-1426 (which seems easy) * Start round three I think we should try to: * Try to land COUCHDB-1424 * Get clarification on the performance issues For these last two items, I think we should impose a time limit. Let's say a week. I think we should also form some teams, to see if we can do some sprints to get these issues fixed. On Fri, Mar 2, 2012 at 3:49 PM, Jan Lehnardt j...@apache.org wrote: On Mar 2, 2012, at 16:47 , Dirkjan Ochtman wrote: On Fri, Mar 2, 2012 at 16:29, Jan Lehnardt j...@apache.org wrote: Proposed Action: I'd propose to release 1.2.0 as-is with the following points mentioned in the release notes (the exact wording of which is to be done): 1. Note that this release is incompatible with Erlang R15B. A patch is available at [LINK to DIFF]; it will appear in Apache CouchDB 1.2.1. 2. Also note that there are some reports of a performance regression in view building. While initial and ad-hoc tests showed an improvement in most cases, we'd like to ask our users to report any significant differences to the Apache CouchDB 1.1.1 release. While I am usually cheering on the release process, taken together it seems more prudent to abort this round, take the R15B driver and the 142{4,6} patches and then starting a new round. I'd agree if 142{4,6} were decided tickets, but they are still ongoing and potentially, in the 24-case especially, for a while. Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Jan, Sorry it took me a while to respond to this. As I said the db I was testing with is about ~200K docs, with one large ddoc, about 10 views. It takes a considerable long time to index, both on 1.1.x and 1.2.x but on 1.2.x it had not made the same amount of progress after 45 minutes on 1.2.x as did the 1.1.x run after 20 minutes, so I assumed that something has slowed, 40% maybe, that's likely a number I pulled out of my hat from the chat and reports of others. I understand this is hand-wavy, which is why I was reluctant to jump to conclusions. Please note that I did not state we should stop this release, nor did I state that any of the issues I enumerated were showstoppers. I stated, with a -1 vote that the overall experience led me to conclude the release was not sound. It's one vote and I always am happy to go with the majority. I thought we were off to a good start on Sunday when I was digging in here. Yes it's off topic as you and Noah agreed, but once you've pushed this out you might want to revisit the release process. This slow_couchdb is definitely a good smoke test, I like it, but beyond that it's sort of an instance of You're doing it all wrong when it comes to performance. I'll be happy to elaborate on this when we get to that. Best Regards, Bob On Feb 27, 2012, at 8:15 AM, Jan Lehnardt wrote: On Feb 27, 2012, at 12:58 , Bob Dionne wrote: Thanks for the clarification. I hope I'm not conflating things by continuing the discussion here, I thought that's what you requested? The discussion we had on IRC was regarding collecting more data items for the performance regression before we start to draw conclusions. My intention here is to understand what needs doing before we can release 1.2.0. I'll reply inline for the other issues. I just downloaded the release candidate again to start fresh. make distcheck hangs on this step: /Users/bitdiddle/Downloads/apache-couchdb-1.2.0/apache-couchdb-1.2.0/_build/../test/etap/150-invalid-view-seq.t . 6/? Just stops completely. This is on R15B which has been rebuilt to use the recommended older SSL version. I haven't looked into this crashing too closely but I'm suspicious that I only see it with couchdb and never with bigcouch and never using the 1.2.x branch from source or any branch for that matter From the release you should run `make check`, not make distcheck. But I assume you see a hang there too, as I have and others (yet not everybody), too. I can't comment on BigCouch and what is different there. It is interesting that 1.2.x won't hang. For me, `make check` in 1.2.x on R15B hangs sometimes, in different places. I'm currently trying to gather more information about this. The question here is whether `make check` passing in R15B is a release requirement. In my vote I considered no, but I am happy to go with a community decision if it emerges. What is your take here? In addition, this just shouldn't be a question, so we should investigate why this happens at all and address the issue, hence COUCHDB-1424. Any insight here would be appreciated as well. In the command line tests, 2,7, 27, and 32 fail. but it differs from run to run. I assume you mean the JS tests. Again, this isn't supposed to work in 1.2.x. I'm happy to backport my changes from master to 1.2.x to make that work, but I refrained from that because I didn't want to bring too much change to a release branch. I'm happy to reconsider, but I don't think a release vote is a good place to discuss feature backports. On Chrome attachment_ranges fails and it hangs on replicator_db This one is an explaining away, but I think it is warranted. Chrome is broken for attachment_ranges. I don't know if we reported this upstream (Robert N?), but this isn't a release blocker. For the replicator_db test, can you try running that in other browsers. I understand it is not the best of situation (hence the move to the cli test suite for master), but if you get this test to pass in at least one other browsers, this isn't a problem that holds 1.2.x. With respect to performance I think comparisons with 1.1.x are important. I think almost any use case, contrived or otherwise should not be dismissed as a pathological or edge case. Bob's script is as simple as it gets and to me is a great smoke test. We need to figure out the reason 1.2 is clearly slower in this case. If there are specific scenarios that 1.2.x is optimized for then we should document that and provide reasons for the trade-offs I want to make absolutely clear that I take any report of performance regression very seriously. But I'm rather annoyed that no information about this ends up on dev@. I understand that on IRC there's some shared understanding of a few scenarios where performance regressions can be shown. I asked three times now that these be posted to this mailing list. I'm not
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Sun, Feb 26, 2012 at 08:25:37PM +0100, Jan Lehnardt wrote: I ran Robert's test with a tiny doc and minimal view and saw the 25% performance drop, but using more realistic doc sizes and views show and improvement for 1.2.x (in one case 1k docs with a view that emits a complex key and an integer for the value). On Debian testing, using all Debian-supplied packages except for Couch itself, I do not see any performance regression. This is on spinning platters, 24GB RAM, 4x Opteron processors split across a few vservers, of which only one is running the couch test. Because Jan's test case relies on bash and ruby, I've not been able to run it on Windows, but I could attempt to run rnewson's approach via a CMD/PowerShell script if desired. IMO signs point to this being something OSX specific but it's just speculation at this point.
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Hi Bob, thanks for your reply. I think this is a bit of a convoluted situation now and I want to apologise for anything that might have upset you or anyone here. I didn't mean to be insulting or anything. I think the 1.2.0 release is very important for this community and the project, so I am very interested in getting it out. I'm not interested in getting it out at all costs but I do want to clarify what's needed to do the release. In the process, your -1 turned out to trigger a hold on the 1.2.0 release, even though you write you didn't intend that. Also in the process I confused your -1 to ask for the hold and as a result I got annoyed that you didn't provide more information. I also asked for clarification on the other issues you raised (e.g. running the test suite in Firefox/across multiple browsers) and you haven't responded to yet (I understand time constraints, but I'd have appreciated a gonna work on it sort of quick response). I'm sorry I got annoyed and let it out on you. All I want is to understand what it takes to release 1.2.0 and we need all your (and everybody's) support to get there. So thanks for any contributions to this discussion. Cheers Jan -- On Feb 29, 2012, at 16:27 , Bob Dionne wrote: Jan, Sorry it took me a while to respond to this. As I said the db I was testing with is about ~200K docs, with one large ddoc, about 10 views. It takes a considerable long time to index, both on 1.1.x and 1.2.x but on 1.2.x it had not made the same amount of progress after 45 minutes on 1.2.x as did the 1.1.x run after 20 minutes, so I assumed that something has slowed, 40% maybe, that's likely a number I pulled out of my hat from the chat and reports of others. I understand this is hand-wavy, which is why I was reluctant to jump to conclusions. Please note that I did not state we should stop this release, nor did I state that any of the issues I enumerated were showstoppers. I stated, with a -1 vote that the overall experience led me to conclude the release was not sound. It's one vote and I always am happy to go with the majority. I thought we were off to a good start on Sunday when I was digging in here. Yes it's off topic as you and Noah agreed, but once you've pushed this out you might want to revisit the release process. This slow_couchdb is definitely a good smoke test, I like it, but beyond that it's sort of an instance of You're doing it all wrong when it comes to performance. I'll be happy to elaborate on this when we get to that. Best Regards, Bob On Feb 27, 2012, at 8:15 AM, Jan Lehnardt wrote: On Feb 27, 2012, at 12:58 , Bob Dionne wrote: Thanks for the clarification. I hope I'm not conflating things by continuing the discussion here, I thought that's what you requested? The discussion we had on IRC was regarding collecting more data items for the performance regression before we start to draw conclusions. My intention here is to understand what needs doing before we can release 1.2.0. I'll reply inline for the other issues. I just downloaded the release candidate again to start fresh. make distcheck hangs on this step: /Users/bitdiddle/Downloads/apache-couchdb-1.2.0/apache-couchdb-1.2.0/_build/../test/etap/150-invalid-view-seq.t . 6/? Just stops completely. This is on R15B which has been rebuilt to use the recommended older SSL version. I haven't looked into this crashing too closely but I'm suspicious that I only see it with couchdb and never with bigcouch and never using the 1.2.x branch from source or any branch for that matter From the release you should run `make check`, not make distcheck. But I assume you see a hang there too, as I have and others (yet not everybody), too. I can't comment on BigCouch and what is different there. It is interesting that 1.2.x won't hang. For me, `make check` in 1.2.x on R15B hangs sometimes, in different places. I'm currently trying to gather more information about this. The question here is whether `make check` passing in R15B is a release requirement. In my vote I considered no, but I am happy to go with a community decision if it emerges. What is your take here? In addition, this just shouldn't be a question, so we should investigate why this happens at all and address the issue, hence COUCHDB-1424. Any insight here would be appreciated as well. In the command line tests, 2,7, 27, and 32 fail. but it differs from run to run. I assume you mean the JS tests. Again, this isn't supposed to work in 1.2.x. I'm happy to backport my changes from master to 1.2.x to make that work, but I refrained from that because I didn't want to bring too much change to a release branch. I'm happy to reconsider, but I don't think a release vote is a good place to discuss feature backports. On Chrome attachment_ranges fails and it hangs on replicator_db This one is an explaining away, but
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
you might want to revisit the release process You mean we here. :) B. On 29 February 2012 16:23, Jan Lehnardt j...@apache.org wrote: Hi Bob, thanks for your reply. I think this is a bit of a convoluted situation now and I want to apologise for anything that might have upset you or anyone here. I didn't mean to be insulting or anything. I think the 1.2.0 release is very important for this community and the project, so I am very interested in getting it out. I'm not interested in getting it out at all costs but I do want to clarify what's needed to do the release. In the process, your -1 turned out to trigger a hold on the 1.2.0 release, even though you write you didn't intend that. Also in the process I confused your -1 to ask for the hold and as a result I got annoyed that you didn't provide more information. I also asked for clarification on the other issues you raised (e.g. running the test suite in Firefox/across multiple browsers) and you haven't responded to yet (I understand time constraints, but I'd have appreciated a gonna work on it sort of quick response). I'm sorry I got annoyed and let it out on you. All I want is to understand what it takes to release 1.2.0 and we need all your (and everybody's) support to get there. So thanks for any contributions to this discussion. Cheers Jan -- On Feb 29, 2012, at 16:27 , Bob Dionne wrote: Jan, Sorry it took me a while to respond to this. As I said the db I was testing with is about ~200K docs, with one large ddoc, about 10 views. It takes a considerable long time to index, both on 1.1.x and 1.2.x but on 1.2.x it had not made the same amount of progress after 45 minutes on 1.2.x as did the 1.1.x run after 20 minutes, so I assumed that something has slowed, 40% maybe, that's likely a number I pulled out of my hat from the chat and reports of others. I understand this is hand-wavy, which is why I was reluctant to jump to conclusions. Please note that I did not state we should stop this release, nor did I state that any of the issues I enumerated were showstoppers. I stated, with a -1 vote that the overall experience led me to conclude the release was not sound. It's one vote and I always am happy to go with the majority. I thought we were off to a good start on Sunday when I was digging in here. Yes it's off topic as you and Noah agreed, but once you've pushed this out you might want to revisit the release process. This slow_couchdb is definitely a good smoke test, I like it, but beyond that it's sort of an instance of You're doing it all wrong when it comes to performance. I'll be happy to elaborate on this when we get to that. Best Regards, Bob On Feb 27, 2012, at 8:15 AM, Jan Lehnardt wrote: On Feb 27, 2012, at 12:58 , Bob Dionne wrote: Thanks for the clarification. I hope I'm not conflating things by continuing the discussion here, I thought that's what you requested? The discussion we had on IRC was regarding collecting more data items for the performance regression before we start to draw conclusions. My intention here is to understand what needs doing before we can release 1.2.0. I'll reply inline for the other issues. I just downloaded the release candidate again to start fresh. make distcheck hangs on this step: /Users/bitdiddle/Downloads/apache-couchdb-1.2.0/apache-couchdb-1.2.0/_build/../test/etap/150-invalid-view-seq.t . 6/? Just stops completely. This is on R15B which has been rebuilt to use the recommended older SSL version. I haven't looked into this crashing too closely but I'm suspicious that I only see it with couchdb and never with bigcouch and never using the 1.2.x branch from source or any branch for that matter From the release you should run `make check`, not make distcheck. But I assume you see a hang there too, as I have and others (yet not everybody), too. I can't comment on BigCouch and what is different there. It is interesting that 1.2.x won't hang. For me, `make check` in 1.2.x on R15B hangs sometimes, in different places. I'm currently trying to gather more information about this. The question here is whether `make check` passing in R15B is a release requirement. In my vote I considered no, but I am happy to go with a community decision if it emerges. What is your take here? In addition, this just shouldn't be a question, so we should investigate why this happens at all and address the issue, hence COUCHDB-1424. Any insight here would be appreciated as well. In the command line tests, 2,7, 27, and 32 fail. but it differs from run to run. I assume you mean the JS tests. Again, this isn't supposed to work in 1.2.x. I'm happy to backport my changes from master to 1.2.x to make that work, but I refrained from that because I didn't want to bring too much change to a release branch. I'm happy to reconsider, but I don't think a release vote is a good place to
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Wed, Feb 29, 2012 at 7:30 PM, Benoit Chesneau bchesn...@gmail.com wrote: On Fri, Feb 24, 2012 at 4:39 PM, Benoit Chesneau bchesn...@gmail.com wrote: On Thu, Feb 23, 2012 at 12:28 AM, Noah Slater nsla...@tumbolia.org wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting, N +1 make check , js tests signatures ok erlang R14B04 - osx lion snow leopard, ubuntu 11.04, 11.10, debian 6, fbsd 8.2 - benoît While testing the reported issue I found COUCHDB-1426 which means we can't build couchdb with a different spidermonkey lib when another one is installed. I'm now -1 on this release. Also where can I set the icu lib path now ? The option doesn't appear any more in the ./configure command , only for windows. - benoît Posted a patch fixing COUCHDB-1426. https://issues.apache.org/jira/browse/COUCHDB-1426
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Jason, made some more tests with larger documents (template is https://gist.github.com/1930804) and a different map function: function(doc) { emit([doc.type, doc.category], doc.nested.coords); } (patch http://friendpaste.com/5C99aqXocN6N6H1BAYIigs) Here's the results I got ( https://gist.github.com/1930807 ) Before COUCHDB-1186 fdmanana 23:21:05 ~/git/hub/slow_couchdb (master) docs=50 batch=5000 ./bench.sh wow.tpl Server: CouchDB/1.2.0a-a68a792-git (Erlang OTP/R14B03) {couchdb:Welcome,version:1.2.0a-a68a792-git} [INFO] Created DB named `db1' [INFO] Uploaded 5000 documents via _bulk_docs () [INFO] Uploaded 5000 documents via _bulk_docs Building view. {total_rows:50,offset:0,rows:[ {id:00051ef7-d735-48d7-9ba8-5a21a86e8d57,key:[dwarf,assassin],value:[{x:31227.35,y:31529.73},{x:116667.85,y:82008.25},{x:224.11,y:36652.41},{x:128565.95,y:6780.2},{x:165230.43,y:176208.63}]} ]} real5m6.676s user0m0.009s sys 0m0.010s After COUCHDB-1186 fdmanana 23:50:07 ~/git/hub/slow_couchdb (master) docs=50 batch=5000 ./bench.sh wow.tpl Server: CouchDB/1.2.0a-f023052-git (Erlang OTP/R14B03) {couchdb:Welcome,version:1.2.0a-f023052-git} [INFO] Created DB named `db1' [INFO] Uploaded 5000 documents via _bulk_docs () [INFO] Uploaded 5000 documents via _bulk_docs Building view. {total_rows:50,offset:0,rows:[ {id:00051ef7-d735-48d7-9ba8-5a21a86e8d57,key:[dwarf,assassin],value:[{x:31227.35,y:31529.73},{x:116667.85,y:82008.25},{x:224.11,y:36652.41},{x:128565.95,y:6780.2},{x:165230.43,y:176208.63}]} ]} real5m1.395s user0m0.008s sys 0m0.010s After COUCHDB-1186 + better queueing patch ( http://friendpaste.com/178nPFgfyyeGf2vtNRpL0w ) fdmanana 00:14:25 ~/git/hub/slow_couchdb (master) docs=50 batch=5000 ./bench.sh wow.tpl Server: CouchDB/1.2.0a-f023052-git (Erlang OTP/R14B03) {couchdb:Welcome,version:1.2.0a-f023052-git} [INFO] Created DB named `db1' [INFO] Uploaded 5000 documents via _bulk_docs () [INFO] Uploaded 5000 documents via _bulk_docs Building view. {total_rows:50,offset:0,rows:[ {id:00051ef7-d735-48d7-9ba8-5a21a86e8d57,key:[dwarf,assassin],value:[{x:31227.35,y:31529.73},{x:116667.85,y:82008.25},{x:224.11,y:36652.41},{x:128565.95,y:6780.2},{x:165230.43,y:176208.63}]} ]} real4m48.175s user0m0.008s sys 0m0.009s Disk model is APPLE SSD TS128C, quad core machine, 8Gb of ram. Unfortunately I don't have access to the machine I used for the tests in COUCHDB-1186 (spinning disk, Linux) before next week. On Mon, Feb 27, 2012 at 7:49 PM, Paul Davis paul.joseph.da...@gmail.com wrote: On Mon, Feb 27, 2012 at 7:18 PM, Filipe David Manana fdman...@apache.org wrote: Jason, can't reproduce those results, not even close: http://friendpaste.com/1L4pHH8WQchaLIMCWhKX9Z Before COUCHDB-1186 fdmanana 16:58:02 ~/git/hub/slow_couchdb (master) docs=50 batch=5 ./bench.sh small_doc.tpl Server: CouchDB/1.2.0a-a68a792-git (Erlang OTP/R14B03) {couchdb:Welcome,version:1.2.0a-a68a792-git} [INFO] Created DB named `db1' [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs Building view. {total_rows:50,offset:0,rows:[ {id:doc1,key:1,value:1} ]} real 0m56.241s user 0m0.006s sys 0m0.005s After COUCHDB-1186 fdmanana 17:02:02 ~/git/hub/slow_couchdb (master) docs=50 batch=5 ./bench.sh small_doc.tpl Server: CouchDB/1.2.0a-f023052-git (Erlang OTP/R14B03) {couchdb:Welcome,version:1.2.0a-f023052-git} [INFO] Created DB named `db1' [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs Building view. {total_rows:50,offset:0,rows:[ {id:doc1,key:1,value:1} ]} real 1m11.694s user 0m0.006s sys 0m0.005s fdmanana 17:06:01 ~/git/hub/slow_couchdb (master) 1.2.0a-f023052-git with patch http://friendpaste.com/178nPFgfyyeGf2vtNRpL0w applied on top fdmanana 17:06:53 ~/git/hub/slow_couchdb (master) docs=50 batch=5 ./bench.sh small_doc.tpl Server: CouchDB/1.2.0a-f023052-git (Erlang OTP/R14B03) {couchdb:Welcome,version:1.2.0a-f023052-git} [INFO] Created DB named `db1' [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Tue, Feb 28, 2012 at 4:49 AM, Paul Davis paul.joseph.da...@gmail.com wrote Yeah, I've seen the btree behave quite differently on SSD's vs HDD's (same code had drastically different runtime characteristics). In other words, can we get a report of what type of disk everyone is running on? + 1 . We actually pollute this thread about vote, and the ticket about view speedups which could be related or not :) Maybe we could open a ticket to collect all the feedback and tests we have ? Also noah, jan what is the status of this vote? Should we consider it as aborted or paused? - benoit
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Filipe, This additional patch looks good, though I haven't tested it. Interesting comment about R15B, I did notice a difference with BigCouch in terms of some of the internal race conditions we see at times. Perhaps there are some performance changes relating to that. I also recently upgraded from the Macbook pro to a MBA so who knows. I ran Jason and Bob's scripts a bit last night and saw similar slow downs between 1.1 and 1.2, though as reported elsewhere with larger docs it's less of an issue. In this patch[1] there's clearly a savings in avoiding the decode call, but I wonder how often that case obtains compared to the others. If {cmd, CMD} dominates then there is an additional overhead incurred however small it might be. Perhaps this explains why the benefits appear for larger docs only. Anyway, just speculation from the code. Regards, Bob [1] https://github.com/fdmanana/couchdb/commit/cce325378723c863f05cca21 On Feb 27, 2012, at 11:33 AM, Filipe David Manana wrote: I just tried Jason's script (modified it to use 500 000 docs instead of 50 000) against 1.2.x and 1.1.1, using OTP R14B03. Here's my results: 1.2.x: $ port=5984 ./test.sh none Filling db. done HTTP/1.1 200 OK Server: CouchDB/1.2.0 (Erlang OTP/R14B03) Date: Mon, 27 Feb 2012 16:08:43 GMT Content-Type: text/plain; charset=utf-8 Content-Length: 252 Cache-Control: must-revalidate {db_name:db1,doc_count:51,doc_del_count:0,update_seq:51,purge_seq:0,compact_running:false,disk_size:130494577,data_size:130490673,instance_start_time:1330358830830086,disk_format_version:6,committed_update_seq:51} Building view. real 1m5.725s user 0m0.006s sys 0m0.005s done 1.1.1: $ port=5984 ./test.sh Filling db. done HTTP/1.1 200 OK Server: CouchDB/1.1.2a785d32f-git (Erlang OTP/R14B03) Date: Mon, 27 Feb 2012 16:15:33 GMT Content-Type: text/plain;charset=utf-8 Content-Length: 230 Cache-Control: must-revalidate {db_name:db1,doc_count:51,doc_del_count:0,update_seq:51,purge_seq:0,compact_running:false,disk_size:122142818,instance_start_time:1330359233327316,disk_format_version:5,committed_update_seq:51} Building view. real 1m4.249s user 0m0.006s sys 0m0.005s done I don't see any significant difference there. Regarding COUCHDB-1186, the only thing that might cause some non determinism and affect performance is the queing/dequeing. Depending on timings, it's possible the writer is dequeing less items per dequeue operation and therefore inserting smaller batches into the btree. The following small change ensures larger batches (while still respecting the queue max size/item count): http://friendpaste.com/178nPFgfyyeGf2vtNRpL0w Running the test with this change: $ port=5984 ./test.sh none Filling db. done HTTP/1.1 200 OK Server: CouchDB/1.2.0 (Erlang OTP/R14B03) Date: Mon, 27 Feb 2012 16:23:20 GMT Content-Type: text/plain; charset=utf-8 Content-Length: 252 Cache-Control: must-revalidate {db_name:db1,doc_count:51,doc_del_count:0,update_seq:51,purge_seq:0,compact_running:false,disk_size:130494577,data_size:130490673,instance_start_time:1330359706846104,disk_format_version:6,committed_update_seq:51} Building view. real 0m49.762s user 0m0.006s sys 0m0.005s done If there's no objection, I'll push that patch. Also, another note, I noticed sometime ago that with master, using OTP R15B I got a performance drop of 10% to 15% compared to using master with OTP R14B04. Maybe it applies to 1.2.x as well. On Mon, Feb 27, 2012 at 5:33 AM, Robert Newson rnew...@apache.org wrote: Bob D, can you give more details on the data set you're testing? Number of docs, size/complexity of docs, etc? Basically, enough info that I could write a script to automate building an equivalent database. I wrote a quick bash script to make a database and time a view build here: http://friendpaste.com/7kBiKJn3uX1KiGJAFPv4nK B. On 27 February 2012 13:15, Jan Lehnardt j...@apache.org wrote: On Feb 27, 2012, at 12:58 , Bob Dionne wrote: Thanks for the clarification. I hope I'm not conflating things by continuing the discussion here, I thought that's what you requested? The discussion we had on IRC was regarding collecting more data items for the performance regression before we start to draw conclusions. My intention here is to understand what needs doing before we can release 1.2.0. I'll reply inline for the other issues. I just downloaded the release candidate again to start fresh. make distcheck hangs on this step: /Users/bitdiddle/Downloads/apache-couchdb-1.2.0/apache-couchdb-1.2.0/_build/../test/etap/150-invalid-view-seq.t . 6/? Just stops completely. This is on R15B which has been rebuilt to use the recommended older SSL version. I haven't looked into this crashing too closely but I'm suspicious that I only see it with couchdb and never with bigcouch and never using the 1.2.x
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Tue, Feb 28, 2012 at 11:05 AM, Benoit Chesneau bchesn...@gmail.com wrote: On Tue, Feb 28, 2012 at 4:49 AM, Paul Davis paul.joseph.da...@gmail.com wrote Yeah, I've seen the btree behave quite differently on SSD's vs HDD's (same code had drastically different runtime characteristics). In other words, can we get a report of what type of disk everyone is running on? + 1 . We actually pollute this thread about vote, and the ticket about view speedups which could be related or not :) Maybe we could open a ticket to collect all the feedback and tests we have ? N / Y ?
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
I'm running my script on a EC2 node with spinning media, the numbers come out the same for 1.1.1 vs 1.2. The only time I've seen a slowdown with a scripted approach is my original one which didn't use bulk docs. :/ B. On 28 February 2012 11:33, Bob Dionne dio...@dionne-associates.com wrote: Filipe, This additional patch looks good, though I haven't tested it. Interesting comment about R15B, I did notice a difference with BigCouch in terms of some of the internal race conditions we see at times. Perhaps there are some performance changes relating to that. I also recently upgraded from the Macbook pro to a MBA so who knows. I ran Jason and Bob's scripts a bit last night and saw similar slow downs between 1.1 and 1.2, though as reported elsewhere with larger docs it's less of an issue. In this patch[1] there's clearly a savings in avoiding the decode call, but I wonder how often that case obtains compared to the others. If {cmd, CMD} dominates then there is an additional overhead incurred however small it might be. Perhaps this explains why the benefits appear for larger docs only. Anyway, just speculation from the code. Regards, Bob [1] https://github.com/fdmanana/couchdb/commit/cce325378723c863f05cca21 On Feb 27, 2012, at 11:33 AM, Filipe David Manana wrote: I just tried Jason's script (modified it to use 500 000 docs instead of 50 000) against 1.2.x and 1.1.1, using OTP R14B03. Here's my results: 1.2.x: $ port=5984 ./test.sh none Filling db. done HTTP/1.1 200 OK Server: CouchDB/1.2.0 (Erlang OTP/R14B03) Date: Mon, 27 Feb 2012 16:08:43 GMT Content-Type: text/plain; charset=utf-8 Content-Length: 252 Cache-Control: must-revalidate {db_name:db1,doc_count:51,doc_del_count:0,update_seq:51,purge_seq:0,compact_running:false,disk_size:130494577,data_size:130490673,instance_start_time:1330358830830086,disk_format_version:6,committed_update_seq:51} Building view. real 1m5.725s user 0m0.006s sys 0m0.005s done 1.1.1: $ port=5984 ./test.sh Filling db. done HTTP/1.1 200 OK Server: CouchDB/1.1.2a785d32f-git (Erlang OTP/R14B03) Date: Mon, 27 Feb 2012 16:15:33 GMT Content-Type: text/plain;charset=utf-8 Content-Length: 230 Cache-Control: must-revalidate {db_name:db1,doc_count:51,doc_del_count:0,update_seq:51,purge_seq:0,compact_running:false,disk_size:122142818,instance_start_time:1330359233327316,disk_format_version:5,committed_update_seq:51} Building view. real 1m4.249s user 0m0.006s sys 0m0.005s done I don't see any significant difference there. Regarding COUCHDB-1186, the only thing that might cause some non determinism and affect performance is the queing/dequeing. Depending on timings, it's possible the writer is dequeing less items per dequeue operation and therefore inserting smaller batches into the btree. The following small change ensures larger batches (while still respecting the queue max size/item count): http://friendpaste.com/178nPFgfyyeGf2vtNRpL0w Running the test with this change: $ port=5984 ./test.sh none Filling db. done HTTP/1.1 200 OK Server: CouchDB/1.2.0 (Erlang OTP/R14B03) Date: Mon, 27 Feb 2012 16:23:20 GMT Content-Type: text/plain; charset=utf-8 Content-Length: 252 Cache-Control: must-revalidate {db_name:db1,doc_count:51,doc_del_count:0,update_seq:51,purge_seq:0,compact_running:false,disk_size:130494577,data_size:130490673,instance_start_time:1330359706846104,disk_format_version:6,committed_update_seq:51} Building view. real 0m49.762s user 0m0.006s sys 0m0.005s done If there's no objection, I'll push that patch. Also, another note, I noticed sometime ago that with master, using OTP R15B I got a performance drop of 10% to 15% compared to using master with OTP R14B04. Maybe it applies to 1.2.x as well. On Mon, Feb 27, 2012 at 5:33 AM, Robert Newson rnew...@apache.org wrote: Bob D, can you give more details on the data set you're testing? Number of docs, size/complexity of docs, etc? Basically, enough info that I could write a script to automate building an equivalent database. I wrote a quick bash script to make a database and time a view build here: http://friendpaste.com/7kBiKJn3uX1KiGJAFPv4nK B. On 27 February 2012 13:15, Jan Lehnardt j...@apache.org wrote: On Feb 27, 2012, at 12:58 , Bob Dionne wrote: Thanks for the clarification. I hope I'm not conflating things by continuing the discussion here, I thought that's what you requested? The discussion we had on IRC was regarding collecting more data items for the performance regression before we start to draw conclusions. My intention here is to understand what needs doing before we can release 1.2.0. I'll reply inline for the other issues. I just downloaded the release candidate again to start fresh. make distcheck hangs on this step:
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Tue, Feb 28, 2012 at 10:05 AM, Benoit Chesneau bchesn...@gmail.comwrote: Also noah, jan what is the status of this vote? Should we consider it as aborted or paused? As far as I can tell, we have not identified, for sure, a release blocking issue. Once we are sure that there is a release blocking issue, I will abort the vote. But I am not aborting it simply because it's taking a while to get clear on the issues. :)
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Jan, Thanks for the clarification. I hope I'm not conflating things by continuing the discussion here, I thought that's what you requested? I just downloaded the release candidate again to start fresh. make distcheck hangs on this step: /Users/bitdiddle/Downloads/apache-couchdb-1.2.0/apache-couchdb-1.2.0/_build/../test/etap/150-invalid-view-seq.t . 6/? Just stops completely. This is on R15B which has been rebuilt to use the recommended older SSL version. I haven't looked into this crashing too closely but I'm suspicious that I only see it with couchdb and never with bigcouch and never using the 1.2.x branch from source or any branch for that matter In the command line tests, 2,7, 27, and 32 fail. but it differs from run to run. On Chrome attachment_ranges fails and it hangs on replicator_db With respect to performance I think comparisons with 1.1.x are important. I think almost any use case, contrived or otherwise should not be dismissed as a pathological or edge case. Bob's script is as simple as it gets and to me is a great smoke test. We need to figure out the reason 1.2 is clearly slower in this case. If there are specific scenarios that 1.2.x is optimized for then we should document that and provide reasons for the trade-offs Cheers, Bob On Feb 26, 2012, at 4:07 PM, Jan Lehnardt wrote: Bob, thanks for your reply I wasn't implying we should try to explain anything away. All of these are valid concerns, I just wanted to get a better understanding on where the bit flips from +0 to -1 and subsequently, how to address that boundary. Ideally we can just fix all of the things you mention, but I think it is important to understand them in detail, that's why I was going into them. Ultimately, I want to understand what we need to do to ship 1.2.0. On Feb 26, 2012, at 21:22 , Bob Dionne wrote: Jan, I'm -1 based on all of my evaluation. I've spent a few hours on this release now yesterday and today. It doesn't really pass what I would call the smoke test. Almost everything I've run into has an explanation: 1. crashes out of the box - that's R15B, you need to recompile SSL and Erlang (we'll note on release notes) Have we spent any time on figuring out what the trouble here is? 2. etaps hang running make check. Known issue. Our etap code is out of date, recent versions of etap don't even run their own unit tests I have seen the etap hang as well, and I wasn't diligent enough to report it in JIRA, I have done so now (COUCHDB-1424). 3. Futon tests fail. Some are known bugs (attachment ranges in Chrome) . Both Chrome and Safari also hang Do you have more details on where Chrome and Safari hang? Can you try their private browsing features, double/triple check that caches are empty? Can you get to a situation where you get all tests succeeding across all browsers, even if individual ones fail on one or two others? 4. standalone JS tests fail. Again most of these run when run by themselves Which ones? 5. performance. I used real production data *because* Stefan on user reported performance degradation on his data set. Any numbers are meaningless for a single test. I also ran scripts that BobN and Jason Smith posted that show a difference between 1.1.x and 1.2.x You are conflating an IRC discussion we've had into this thread. The performance regression reported is a good reason to look into other scenarios where we can show slowdowns. But we need to understand what's happening. Just from looking at dev@ all I see is some handwaving about some reports some people have done (Not to discourage any work that has been done on IRC and user@, but for the sake of a release vote thread, this related information needs to be on this mailing list). As I said on IRC, I'm happy to get my hands dirty to understand the regression at hand. But we need to know where we'd draw a line and say this isn't acceptable for a 1.2.0. 6. Reviewed patch pointed to by Jason that may be the cause but it's hard to say without knowing the code analysis that went into the changes. You can see obvious local optimizations that make good sense but those are often the ones that get you, without knowing the call counts. That is a point that wasn't included in your previous mail. It's great that there is progress, thanks for looking into this! Many of these issues can be explained away, but I think end users will be less forgiving. I think we already struggle with view performance. I'm interested to see how others evaluate this regression. I'll try this seatoncouch tool you mention later to see if I can construct some more definitive tests. Again, I'm not trying to explain anything away. I want to get a shared understanding of the issues you raised and where we stand on solving them squared against the ongoing 1.2.0 release. And again: Thanks for doing this thorough review and
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
I think it's worth reminding everyone that the Futon test suite is only supported on FireFox. The reports of Chrome failing on attachment_ranges illuminates nothing; it's a Chrome bug. It would be news if it *didn't* fail on Chrome. :) The performance regression, I think, is the only thing holding up the release, can we focus on that? Particularly, narrowing down the commits that are suspected and a discussion on the methodology used to measure this so far (is it a fair test? etc). My take is that Bob has measured a significant regression on a real world data set and that Jan has seem an improvement on a different real world data set. B. On 27 February 2012 11:58, Bob Dionne dio...@dionne-associates.com wrote: Jan, Thanks for the clarification. I hope I'm not conflating things by continuing the discussion here, I thought that's what you requested? I just downloaded the release candidate again to start fresh. make distcheck hangs on this step: /Users/bitdiddle/Downloads/apache-couchdb-1.2.0/apache-couchdb-1.2.0/_build/../test/etap/150-invalid-view-seq.t . 6/? Just stops completely. This is on R15B which has been rebuilt to use the recommended older SSL version. I haven't looked into this crashing too closely but I'm suspicious that I only see it with couchdb and never with bigcouch and never using the 1.2.x branch from source or any branch for that matter In the command line tests, 2,7, 27, and 32 fail. but it differs from run to run. On Chrome attachment_ranges fails and it hangs on replicator_db With respect to performance I think comparisons with 1.1.x are important. I think almost any use case, contrived or otherwise should not be dismissed as a pathological or edge case. Bob's script is as simple as it gets and to me is a great smoke test. We need to figure out the reason 1.2 is clearly slower in this case. If there are specific scenarios that 1.2.x is optimized for then we should document that and provide reasons for the trade-offs Cheers, Bob On Feb 26, 2012, at 4:07 PM, Jan Lehnardt wrote: Bob, thanks for your reply I wasn't implying we should try to explain anything away. All of these are valid concerns, I just wanted to get a better understanding on where the bit flips from +0 to -1 and subsequently, how to address that boundary. Ideally we can just fix all of the things you mention, but I think it is important to understand them in detail, that's why I was going into them. Ultimately, I want to understand what we need to do to ship 1.2.0. On Feb 26, 2012, at 21:22 , Bob Dionne wrote: Jan, I'm -1 based on all of my evaluation. I've spent a few hours on this release now yesterday and today. It doesn't really pass what I would call the smoke test. Almost everything I've run into has an explanation: 1. crashes out of the box - that's R15B, you need to recompile SSL and Erlang (we'll note on release notes) Have we spent any time on figuring out what the trouble here is? 2. etaps hang running make check. Known issue. Our etap code is out of date, recent versions of etap don't even run their own unit tests I have seen the etap hang as well, and I wasn't diligent enough to report it in JIRA, I have done so now (COUCHDB-1424). 3. Futon tests fail. Some are known bugs (attachment ranges in Chrome) . Both Chrome and Safari also hang Do you have more details on where Chrome and Safari hang? Can you try their private browsing features, double/triple check that caches are empty? Can you get to a situation where you get all tests succeeding across all browsers, even if individual ones fail on one or two others? 4. standalone JS tests fail. Again most of these run when run by themselves Which ones? 5. performance. I used real production data *because* Stefan on user reported performance degradation on his data set. Any numbers are meaningless for a single test. I also ran scripts that BobN and Jason Smith posted that show a difference between 1.1.x and 1.2.x You are conflating an IRC discussion we've had into this thread. The performance regression reported is a good reason to look into other scenarios where we can show slowdowns. But we need to understand what's happening. Just from looking at dev@ all I see is some handwaving about some reports some people have done (Not to discourage any work that has been done on IRC and user@, but for the sake of a release vote thread, this related information needs to be on this mailing list). As I said on IRC, I'm happy to get my hands dirty to understand the regression at hand. But we need to know where we'd draw a line and say this isn't acceptable for a 1.2.0. 6. Reviewed patch pointed to by Jason that may be the cause but it's hard to say without knowing the code analysis that went into the changes. You can see obvious local optimizations that make good sense but those are often the ones that get you, without
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Feb 27, 2012, at 12:58 , Bob Dionne wrote: Thanks for the clarification. I hope I'm not conflating things by continuing the discussion here, I thought that's what you requested? The discussion we had on IRC was regarding collecting more data items for the performance regression before we start to draw conclusions. My intention here is to understand what needs doing before we can release 1.2.0. I'll reply inline for the other issues. I just downloaded the release candidate again to start fresh. make distcheck hangs on this step: /Users/bitdiddle/Downloads/apache-couchdb-1.2.0/apache-couchdb-1.2.0/_build/../test/etap/150-invalid-view-seq.t . 6/? Just stops completely. This is on R15B which has been rebuilt to use the recommended older SSL version. I haven't looked into this crashing too closely but I'm suspicious that I only see it with couchdb and never with bigcouch and never using the 1.2.x branch from source or any branch for that matter From the release you should run `make check`, not make distcheck. But I assume you see a hang there too, as I have and others (yet not everybody), too. I can't comment on BigCouch and what is different there. It is interesting that 1.2.x won't hang. For me, `make check` in 1.2.x on R15B hangs sometimes, in different places. I'm currently trying to gather more information about this. The question here is whether `make check` passing in R15B is a release requirement. In my vote I considered no, but I am happy to go with a community decision if it emerges. What is your take here? In addition, this just shouldn't be a question, so we should investigate why this happens at all and address the issue, hence COUCHDB-1424. Any insight here would be appreciated as well. In the command line tests, 2,7, 27, and 32 fail. but it differs from run to run. I assume you mean the JS tests. Again, this isn't supposed to work in 1.2.x. I'm happy to backport my changes from master to 1.2.x to make that work, but I refrained from that because I didn't want to bring too much change to a release branch. I'm happy to reconsider, but I don't think a release vote is a good place to discuss feature backports. On Chrome attachment_ranges fails and it hangs on replicator_db This one is an explaining away, but I think it is warranted. Chrome is broken for attachment_ranges. I don't know if we reported this upstream (Robert N?), but this isn't a release blocker. For the replicator_db test, can you try running that in other browsers. I understand it is not the best of situation (hence the move to the cli test suite for master), but if you get this test to pass in at least one other browsers, this isn't a problem that holds 1.2.x. With respect to performance I think comparisons with 1.1.x are important. I think almost any use case, contrived or otherwise should not be dismissed as a pathological or edge case. Bob's script is as simple as it gets and to me is a great smoke test. We need to figure out the reason 1.2 is clearly slower in this case. If there are specific scenarios that 1.2.x is optimized for then we should document that and provide reasons for the trade-offs I want to make absolutely clear that I take any report of performance regression very seriously. But I'm rather annoyed that no information about this ends up on dev@. I understand that on IRC there's some shared understanding of a few scenarios where performance regressions can be shown. I asked three times now that these be posted to this mailing list. I'm not asking for a comprehensive report, but anything really. I found Robert Newson's simple test script on IRC and ran that to test a suspicion of mine which I posted in an earlier mail (tiny docs - slower, bigger docs - faster). Nobody else bothered to post this here. I see no discussion about what is observed, what is expected, what would be acceptable for a release of 1.2.0 as is and what not. As far as this list is concerned, we know that a few people claimed that things are slower and it's very real and that we should hold the 1.2.0 release for it. I'm more than happy to hold the release until we figured out the things I asked for above and help out figuring it all out. But we need something to work with here. I also understand that this is a voluntary project and people don't have infinite time to spend, but at least a message of we're collecting things, will report when done, would be *great* to start. So far we only have a hold the horses, there might be a something going on. Please let me know if this request is unreasonable or whether I am overreacting. Sorry for the rant. To anyone who has been looking into performance regression, can you please send to this list any info you have? If you have a comprehensive analysis, awesome, if you just ran some script on a machine, just send us that, let's collect all the data to get this situation solved! We need your help.
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Bob D, can you give more details on the data set you're testing? Number of docs, size/complexity of docs, etc? Basically, enough info that I could write a script to automate building an equivalent database. I wrote a quick bash script to make a database and time a view build here: http://friendpaste.com/7kBiKJn3uX1KiGJAFPv4nK B. On 27 February 2012 13:15, Jan Lehnardt j...@apache.org wrote: On Feb 27, 2012, at 12:58 , Bob Dionne wrote: Thanks for the clarification. I hope I'm not conflating things by continuing the discussion here, I thought that's what you requested? The discussion we had on IRC was regarding collecting more data items for the performance regression before we start to draw conclusions. My intention here is to understand what needs doing before we can release 1.2.0. I'll reply inline for the other issues. I just downloaded the release candidate again to start fresh. make distcheck hangs on this step: /Users/bitdiddle/Downloads/apache-couchdb-1.2.0/apache-couchdb-1.2.0/_build/../test/etap/150-invalid-view-seq.t . 6/? Just stops completely. This is on R15B which has been rebuilt to use the recommended older SSL version. I haven't looked into this crashing too closely but I'm suspicious that I only see it with couchdb and never with bigcouch and never using the 1.2.x branch from source or any branch for that matter From the release you should run `make check`, not make distcheck. But I assume you see a hang there too, as I have and others (yet not everybody), too. I can't comment on BigCouch and what is different there. It is interesting that 1.2.x won't hang. For me, `make check` in 1.2.x on R15B hangs sometimes, in different places. I'm currently trying to gather more information about this. The question here is whether `make check` passing in R15B is a release requirement. In my vote I considered no, but I am happy to go with a community decision if it emerges. What is your take here? In addition, this just shouldn't be a question, so we should investigate why this happens at all and address the issue, hence COUCHDB-1424. Any insight here would be appreciated as well. In the command line tests, 2,7, 27, and 32 fail. but it differs from run to run. I assume you mean the JS tests. Again, this isn't supposed to work in 1.2.x. I'm happy to backport my changes from master to 1.2.x to make that work, but I refrained from that because I didn't want to bring too much change to a release branch. I'm happy to reconsider, but I don't think a release vote is a good place to discuss feature backports. On Chrome attachment_ranges fails and it hangs on replicator_db This one is an explaining away, but I think it is warranted. Chrome is broken for attachment_ranges. I don't know if we reported this upstream (Robert N?), but this isn't a release blocker. For the replicator_db test, can you try running that in other browsers. I understand it is not the best of situation (hence the move to the cli test suite for master), but if you get this test to pass in at least one other browsers, this isn't a problem that holds 1.2.x. With respect to performance I think comparisons with 1.1.x are important. I think almost any use case, contrived or otherwise should not be dismissed as a pathological or edge case. Bob's script is as simple as it gets and to me is a great smoke test. We need to figure out the reason 1.2 is clearly slower in this case. If there are specific scenarios that 1.2.x is optimized for then we should document that and provide reasons for the trade-offs I want to make absolutely clear that I take any report of performance regression very seriously. But I'm rather annoyed that no information about this ends up on dev@. I understand that on IRC there's some shared understanding of a few scenarios where performance regressions can be shown. I asked three times now that these be posted to this mailing list. I'm not asking for a comprehensive report, but anything really. I found Robert Newson's simple test script on IRC and ran that to test a suspicion of mine which I posted in an earlier mail (tiny docs - slower, bigger docs - faster). Nobody else bothered to post this here. I see no discussion about what is observed, what is expected, what would be acceptable for a release of 1.2.0 as is and what not. As far as this list is concerned, we know that a few people claimed that things are slower and it's very real and that we should hold the 1.2.0 release for it. I'm more than happy to hold the release until we figured out the things I asked for above and help out figuring it all out. But we need something to work with here. I also understand that this is a voluntary project and people don't have infinite time to spend, but at least a message of we're collecting things, will report when done, would be *great* to start. So far we only have a hold the
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
I just tried Jason's script (modified it to use 500 000 docs instead of 50 000) against 1.2.x and 1.1.1, using OTP R14B03. Here's my results: 1.2.x: $ port=5984 ./test.sh none Filling db. done HTTP/1.1 200 OK Server: CouchDB/1.2.0 (Erlang OTP/R14B03) Date: Mon, 27 Feb 2012 16:08:43 GMT Content-Type: text/plain; charset=utf-8 Content-Length: 252 Cache-Control: must-revalidate {db_name:db1,doc_count:51,doc_del_count:0,update_seq:51,purge_seq:0,compact_running:false,disk_size:130494577,data_size:130490673,instance_start_time:1330358830830086,disk_format_version:6,committed_update_seq:51} Building view. real1m5.725s user0m0.006s sys 0m0.005s done 1.1.1: $ port=5984 ./test.sh Filling db. done HTTP/1.1 200 OK Server: CouchDB/1.1.2a785d32f-git (Erlang OTP/R14B03) Date: Mon, 27 Feb 2012 16:15:33 GMT Content-Type: text/plain;charset=utf-8 Content-Length: 230 Cache-Control: must-revalidate {db_name:db1,doc_count:51,doc_del_count:0,update_seq:51,purge_seq:0,compact_running:false,disk_size:122142818,instance_start_time:1330359233327316,disk_format_version:5,committed_update_seq:51} Building view. real1m4.249s user0m0.006s sys 0m0.005s done I don't see any significant difference there. Regarding COUCHDB-1186, the only thing that might cause some non determinism and affect performance is the queing/dequeing. Depending on timings, it's possible the writer is dequeing less items per dequeue operation and therefore inserting smaller batches into the btree. The following small change ensures larger batches (while still respecting the queue max size/item count): http://friendpaste.com/178nPFgfyyeGf2vtNRpL0w Running the test with this change: $ port=5984 ./test.sh none Filling db. done HTTP/1.1 200 OK Server: CouchDB/1.2.0 (Erlang OTP/R14B03) Date: Mon, 27 Feb 2012 16:23:20 GMT Content-Type: text/plain; charset=utf-8 Content-Length: 252 Cache-Control: must-revalidate {db_name:db1,doc_count:51,doc_del_count:0,update_seq:51,purge_seq:0,compact_running:false,disk_size:130494577,data_size:130490673,instance_start_time:1330359706846104,disk_format_version:6,committed_update_seq:51} Building view. real0m49.762s user0m0.006s sys 0m0.005s done If there's no objection, I'll push that patch. Also, another note, I noticed sometime ago that with master, using OTP R15B I got a performance drop of 10% to 15% compared to using master with OTP R14B04. Maybe it applies to 1.2.x as well. On Mon, Feb 27, 2012 at 5:33 AM, Robert Newson rnew...@apache.org wrote: Bob D, can you give more details on the data set you're testing? Number of docs, size/complexity of docs, etc? Basically, enough info that I could write a script to automate building an equivalent database. I wrote a quick bash script to make a database and time a view build here: http://friendpaste.com/7kBiKJn3uX1KiGJAFPv4nK B. On 27 February 2012 13:15, Jan Lehnardt j...@apache.org wrote: On Feb 27, 2012, at 12:58 , Bob Dionne wrote: Thanks for the clarification. I hope I'm not conflating things by continuing the discussion here, I thought that's what you requested? The discussion we had on IRC was regarding collecting more data items for the performance regression before we start to draw conclusions. My intention here is to understand what needs doing before we can release 1.2.0. I'll reply inline for the other issues. I just downloaded the release candidate again to start fresh. make distcheck hangs on this step: /Users/bitdiddle/Downloads/apache-couchdb-1.2.0/apache-couchdb-1.2.0/_build/../test/etap/150-invalid-view-seq.t . 6/? Just stops completely. This is on R15B which has been rebuilt to use the recommended older SSL version. I haven't looked into this crashing too closely but I'm suspicious that I only see it with couchdb and never with bigcouch and never using the 1.2.x branch from source or any branch for that matter From the release you should run `make check`, not make distcheck. But I assume you see a hang there too, as I have and others (yet not everybody), too. I can't comment on BigCouch and what is different there. It is interesting that 1.2.x won't hang. For me, `make check` in 1.2.x on R15B hangs sometimes, in different places. I'm currently trying to gather more information about this. The question here is whether `make check` passing in R15B is a release requirement. In my vote I considered no, but I am happy to go with a community decision if it emerges. What is your take here? In addition, this just shouldn't be a question, so we should investigate why this happens at all and address the issue, hence COUCHDB-1424. Any insight here would be appreciated as well. In the command line tests, 2,7, 27, and 32 fail. but it differs from run to run. I assume you mean the JS tests. Again, this isn't supposed to work in 1.2.x. I'm happy to backport my changes from master to 1.2.x to make that
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Mon, Feb 27, 2012 at 1:15 PM, Jan Lehnardt j...@apache.org wrote: It is interesting that 1.2.x won't hang. There is no difference between the files you have on your branch, and the files in the release tarball. You can verify this yourself by following the steps here: http://wiki.apache.org/couchdb/Test_procedure How may times have you tested this? If you run make check on the branch 5 times, how many fail? If you run make check from the tarball 5 times, how many file? The question here is whether `make check` passing in R15B is a release requirement. In my vote I considered no, but I am happy to go with a community decision if it emerges. What is your take here? Yes, this is a release blocker. In addition, this just shouldn't be a question, so we should investigate why this happens at all and address the issue, hence COUCHDB-1424. Any insight here would be appreciated as well. Agreed. In the command line tests, 2,7, 27, and 32 fail. but it differs from run to run. I assume you mean the JS tests. Again, this isn't supposed to work in 1.2.x. I'm happy to backport my changes from master to 1.2.x to make that work, but I refrained from that because I didn't want to bring too much change to a release branch. I'm happy to reconsider, but I don't think a release vote is a good place to discuss feature backports. Jan, I am starting to think of our release vote rounds as release candidates. In so much as, the activity they seem to kick-off seems to be the sort of activity you hope to kick off with a regular release candidate. Does that make sense? Within that context, I think it's fine to talk about stuff like this. A release voting round is a prompt for people to get their shit together. On Chrome attachment_ranges fails and it hangs on replicator_db This one is an explaining away, but I think it is warranted. Chrome is broken for attachment_ranges. I don't know if we reported this upstream (Robert N?), but this isn't a release blocker. For the replicator_db test, can you try running that in other browsers. I understand it is not the best of situation (hence the move to the cli test suite for master), but if you get this test to pass in at least one other browsers, this isn't a problem that holds 1.2.x. We only support Firefox with the test suite. What am I missing?
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Feb 27, 2012, at 18:19 , Noah Slater wrote: On Mon, Feb 27, 2012 at 1:15 PM, Jan Lehnardt j...@apache.org wrote: It is interesting that 1.2.x won't hang. There is no difference between the files you have on your branch, and the files in the release tarball. For me both hang. You can verify this yourself by following the steps here: http://wiki.apache.org/couchdb/Test_procedure How may times have you tested this? If you run make check on the branch 5 times, how many fail? If you run make check from the tarball 5 times, how many file? They all fail. Only occasionally it doesn't. I can't give it a number, but maybe it is 1 in 10 runs. The question here is whether `make check` passing in R15B is a release requirement. In my vote I considered no, but I am happy to go with a community decision if it emerges. What is your take here? Yes, this is a release blocker. See https://issues.apache.org/jira/browse/COUCHDB-1424 for details on this. So far we haven't been able to show that this happens on systems other than Mac OS X 10.7.3. If true, I wouldn't consider this a release blocker. In the command line tests, 2,7, 27, and 32 fail. but it differs from run to run. I assume you mean the JS tests. Again, this isn't supposed to work in 1.2.x. I'm happy to backport my changes from master to 1.2.x to make that work, but I refrained from that because I didn't want to bring too much change to a release branch. I'm happy to reconsider, but I don't think a release vote is a good place to discuss feature backports. Jan, I am starting to think of our release vote rounds as release candidates. In so much as, the activity they seem to kick-off seems to be the sort of activity you hope to kick off with a regular release candidate. Does that make sense? Within that context, I think it's fine to talk about stuff like this. A release voting round is a prompt for people to get their shit together. That's a fair point of view, but I'm looking at this differently. A release branch should be in good shape most of the time and signing off on a release by a vote should be a formality, not start discussions what the scope of the release is. I'm actually thinking about introducing proper RC release to bridge the attention / release gap, but that's a separate discussion. On Chrome attachment_ranges fails and it hangs on replicator_db This one is an explaining away, but I think it is warranted. Chrome is broken for attachment_ranges. I don't know if we reported this upstream (Robert N?), but this isn't a release blocker. For the replicator_db test, can you try running that in other browsers. I understand it is not the best of situation (hence the move to the cli test suite for master), but if you get this test to pass in at least one other browsers, this isn't a problem that holds 1.2.x. We only support Firefox with the test suite. What am I missing? Many people don't use Firefox :) Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Mon, Feb 27, 2012 at 5:41 PM, Jan Lehnardt j...@apache.org wrote: See https://issues.apache.org/jira/browse/COUCHDB-1424 for details on this. So far we haven't been able to show that this happens on systems other than Mac OS X 10.7.3. If true, I wouldn't consider this a release blocker. This doesn't happen for me, or I would not have called the vote, and I am using that exact platform. That's a fair point of view, but I'm looking at this differently. A release branch should be in good shape most of the time and signing off on a release by a vote should be a formality, not start discussions what the scope of the release is. I'm actually thinking about introducing proper RC release to bridge the attention / release gap, but that's a separate discussion. Tabled. Many people don't use Firefox :) http://wiki.apache.org/couchdb/Test_procedure Please note that the unit tests do not run properly under Safari. If you report problems with Safari, you will be asked to test again using Firefox.
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Feb 27, 2012, at 19:11 , Noah Slater wrote: On Mon, Feb 27, 2012 at 5:41 PM, Jan Lehnardt j...@apache.org wrote: See https://issues.apache.org/jira/browse/COUCHDB-1424 for details on this. So far we haven't been able to show that this happens on systems other than Mac OS X 10.7.3. If true, I wouldn't consider this a release blocker. This doesn't happen for me, or I would not have called the vote, and I am using that exact platform. Yeah, sorry, I should have been more clear, it only happens for some Mac users. Many people don't use Firefox :) http://wiki.apache.org/couchdb/Test_procedure Please note that the unit tests do not run properly under Safari. If you report problems with Safari, you will be asked to test again using Firefox. I don't know why Bob D isn't testing in Firefox. Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Feb 23, 2012, at 18:09 , Jan Lehnardt wrote: I'm happy to give this a +1 if we put a warning about R15B on the download page. http://git-wip-us.apache.org/repos/asf?p=couchdb.git;a=commit;h=b1af764b refers to an issue that makes operation under R15B potentially dangerous. We haven't seen a report about any issues, but the Erlang manual quote severe potential issues: These changes are essential to not crash the emulator or worse cause malfunction. Without them a driver may return garbage in the high 32 bits to the emulator causing it to build a huge result from random bytes either crashing on memory allocation or succeeding with a random result from the driver call. and The argument type change is from signed to unsigned which may cause problems for e.g. loop termination conditions or error conditions if you just change the types all over the place. — http://www.erlang.org/doc/man/erl_driver.html#rewrites_for_64_bits The fix is simple enough (see the link above), but this makes me not feel comfortable recommending this release for operation on R15B. I'd be fine with releasing 1.2.0 as is* and make it clear in the release notes that R15B support isn't there yet and ship a 1.2.1 soonish with the fix. * but if we are holding 1.2.0 for any of the other issues we are discussing, I'd say we include this one. What do you think? Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
If the R15/64 bit fix is the only issue with round 2, then I vote with Jan to defer it to 1.2.1. If we're opening up round 3 for any reason, I'd like to see it go in. b. On 27 February 2012 19:09, Jan Lehnardt j...@apache.org wrote: On Feb 23, 2012, at 18:09 , Jan Lehnardt wrote: I'm happy to give this a +1 if we put a warning about R15B on the download page. http://git-wip-us.apache.org/repos/asf?p=couchdb.git;a=commit;h=b1af764b refers to an issue that makes operation under R15B potentially dangerous. We haven't seen a report about any issues, but the Erlang manual quote severe potential issues: These changes are essential to not crash the emulator or worse cause malfunction. Without them a driver may return garbage in the high 32 bits to the emulator causing it to build a huge result from random bytes either crashing on memory allocation or succeeding with a random result from the driver call. and The argument type change is from signed to unsigned which may cause problems for e.g. loop termination conditions or error conditions if you just change the types all over the place. — http://www.erlang.org/doc/man/erl_driver.html#rewrites_for_64_bits The fix is simple enough (see the link above), but this makes me not feel comfortable recommending this release for operation on R15B. I'd be fine with releasing 1.2.0 as is* and make it clear in the release notes that R15B support isn't there yet and ship a 1.2.1 soonish with the fix. * but if we are holding 1.2.0 for any of the other issues we are discussing, I'd say we include this one. What do you think? Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
If the fix is simple enough, I would prefer to see it in. The effects sound moderately serious, and this effects the most recent Erlang, not some legacy shit. On Mon, Feb 27, 2012 at 8:18 PM, Robert Newson rnew...@apache.org wrote: If the R15/64 bit fix is the only issue with round 2, then I vote with Jan to defer it to 1.2.1. If we're opening up round 3 for any reason, I'd like to see it go in. b. On 27 February 2012 19:09, Jan Lehnardt j...@apache.org wrote: On Feb 23, 2012, at 18:09 , Jan Lehnardt wrote: I'm happy to give this a +1 if we put a warning about R15B on the download page. http://git-wip-us.apache.org/repos/asf?p=couchdb.git;a=commit;h=b1af764brefers to an issue that makes operation under R15B potentially dangerous. We haven't seen a report about any issues, but the Erlang manual quote severe potential issues: These changes are essential to not crash the emulator or worse cause malfunction. Without them a driver may return garbage in the high 32 bits to the emulator causing it to build a huge result from random bytes either crashing on memory allocation or succeeding with a random result from the driver call. and The argument type change is from signed to unsigned which may cause problems for e.g. loop termination conditions or error conditions if you just change the types all over the place. — http://www.erlang.org/doc/man/erl_driver.html#rewrites_for_64_bits The fix is simple enough (see the link above), but this makes me not feel comfortable recommending this release for operation on R15B. I'd be fine with releasing 1.2.0 as is* and make it clear in the release notes that R15B support isn't there yet and ship a 1.2.1 soonish with the fix. * but if we are holding 1.2.0 for any of the other issues we are discussing, I'd say we include this one. What do you think? Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Hi, Filipe. Most people seem to be holding their OTP build constant for these tests. If you have the time, would you please check out https://github.com/jhs/slow_couchdb It uses seatoncouch mixed with Bob's script to run a basic benchmark. I expect more template types to grow to help create different data profiles. Anyway, here are my results with 500k documents. Note that I built from your optimization commit, then its parent. https://gist.github.com/1928169 tl;dr = 2:50 before your commit; 4:13 after. On Mon, Feb 27, 2012 at 11:33 PM, Filipe David Manana fdman...@apache.org wrote: I just tried Jason's script (modified it to use 500 000 docs instead of 50 000) against 1.2.x and 1.1.1, using OTP R14B03. Here's my results: 1.2.x: $ port=5984 ./test.sh none Filling db. done HTTP/1.1 200 OK Server: CouchDB/1.2.0 (Erlang OTP/R14B03) Date: Mon, 27 Feb 2012 16:08:43 GMT Content-Type: text/plain; charset=utf-8 Content-Length: 252 Cache-Control: must-revalidate {db_name:db1,doc_count:51,doc_del_count:0,update_seq:51,purge_seq:0,compact_running:false,disk_size:130494577,data_size:130490673,instance_start_time:1330358830830086,disk_format_version:6,committed_update_seq:51} Building view. real 1m5.725s user 0m0.006s sys 0m0.005s done 1.1.1: $ port=5984 ./test.sh Filling db. done HTTP/1.1 200 OK Server: CouchDB/1.1.2a785d32f-git (Erlang OTP/R14B03) Date: Mon, 27 Feb 2012 16:15:33 GMT Content-Type: text/plain;charset=utf-8 Content-Length: 230 Cache-Control: must-revalidate {db_name:db1,doc_count:51,doc_del_count:0,update_seq:51,purge_seq:0,compact_running:false,disk_size:122142818,instance_start_time:1330359233327316,disk_format_version:5,committed_update_seq:51} Building view. real 1m4.249s user 0m0.006s sys 0m0.005s done I don't see any significant difference there. Regarding COUCHDB-1186, the only thing that might cause some non determinism and affect performance is the queing/dequeing. Depending on timings, it's possible the writer is dequeing less items per dequeue operation and therefore inserting smaller batches into the btree. The following small change ensures larger batches (while still respecting the queue max size/item count): http://friendpaste.com/178nPFgfyyeGf2vtNRpL0w Running the test with this change: $ port=5984 ./test.sh none Filling db. done HTTP/1.1 200 OK Server: CouchDB/1.2.0 (Erlang OTP/R14B03) Date: Mon, 27 Feb 2012 16:23:20 GMT Content-Type: text/plain; charset=utf-8 Content-Length: 252 Cache-Control: must-revalidate {db_name:db1,doc_count:51,doc_del_count:0,update_seq:51,purge_seq:0,compact_running:false,disk_size:130494577,data_size:130490673,instance_start_time:1330359706846104,disk_format_version:6,committed_update_seq:51} Building view. real 0m49.762s user 0m0.006s sys 0m0.005s done If there's no objection, I'll push that patch. Also, another note, I noticed sometime ago that with master, using OTP R15B I got a performance drop of 10% to 15% compared to using master with OTP R14B04. Maybe it applies to 1.2.x as well. On Mon, Feb 27, 2012 at 5:33 AM, Robert Newson rnew...@apache.org wrote: Bob D, can you give more details on the data set you're testing? Number of docs, size/complexity of docs, etc? Basically, enough info that I could write a script to automate building an equivalent database. I wrote a quick bash script to make a database and time a view build here: http://friendpaste.com/7kBiKJn3uX1KiGJAFPv4nK B. On 27 February 2012 13:15, Jan Lehnardt j...@apache.org wrote: On Feb 27, 2012, at 12:58 , Bob Dionne wrote: Thanks for the clarification. I hope I'm not conflating things by continuing the discussion here, I thought that's what you requested? The discussion we had on IRC was regarding collecting more data items for the performance regression before we start to draw conclusions. My intention here is to understand what needs doing before we can release 1.2.0. I'll reply inline for the other issues. I just downloaded the release candidate again to start fresh. make distcheck hangs on this step: /Users/bitdiddle/Downloads/apache-couchdb-1.2.0/apache-couchdb-1.2.0/_build/../test/etap/150-invalid-view-seq.t . 6/? Just stops completely. This is on R15B which has been rebuilt to use the recommended older SSL version. I haven't looked into this crashing too closely but I'm suspicious that I only see it with couchdb and never with bigcouch and never using the 1.2.x branch from source or any branch for that matter From the release you should run `make check`, not make distcheck. But I assume you see a hang there too, as I have and others (yet not everybody), too. I can't comment on BigCouch and what is different there. It is interesting that 1.2.x won't hang. For me, `make check` in 1.2.x on R15B hangs sometimes, in different places. I'm currently trying to gather
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Jason, can't reproduce those results, not even close: http://friendpaste.com/1L4pHH8WQchaLIMCWhKX9Z Before COUCHDB-1186 fdmanana 16:58:02 ~/git/hub/slow_couchdb (master) docs=50 batch=5 ./bench.sh small_doc.tpl Server: CouchDB/1.2.0a-a68a792-git (Erlang OTP/R14B03) {couchdb:Welcome,version:1.2.0a-a68a792-git} [INFO] Created DB named `db1' [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs Building view. {total_rows:50,offset:0,rows:[ {id:doc1,key:1,value:1} ]} real0m56.241s user0m0.006s sys 0m0.005s After COUCHDB-1186 fdmanana 17:02:02 ~/git/hub/slow_couchdb (master) docs=50 batch=5 ./bench.sh small_doc.tpl Server: CouchDB/1.2.0a-f023052-git (Erlang OTP/R14B03) {couchdb:Welcome,version:1.2.0a-f023052-git} [INFO] Created DB named `db1' [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs Building view. {total_rows:50,offset:0,rows:[ {id:doc1,key:1,value:1} ]} real1m11.694s user0m0.006s sys 0m0.005s fdmanana 17:06:01 ~/git/hub/slow_couchdb (master) 1.2.0a-f023052-git with patch http://friendpaste.com/178nPFgfyyeGf2vtNRpL0w applied on top fdmanana 17:06:53 ~/git/hub/slow_couchdb (master) docs=50 batch=5 ./bench.sh small_doc.tpl Server: CouchDB/1.2.0a-f023052-git (Erlang OTP/R14B03) {couchdb:Welcome,version:1.2.0a-f023052-git} [INFO] Created DB named `db1' [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs Building view. {total_rows:50,offset:0,rows:[ {id:doc1,key:1,value:1} ]} real0m51.089s user0m0.006s sys 0m0.004s fdmanana 17:10:29 ~/git/hub/slow_couchdb (master) Can you try with R14B0x and also with the patch http://friendpaste.com/178nPFgfyyeGf2vtNRpL0w ? Back then I made all testing on a machine with a spinning disk, so the writer process was slower and likely dequeing more KV pairs from the work queue on each dequeue operation. The tests I did just now are on a machine with a ssd disk. On Mon, Feb 27, 2012 at 4:40 PM, Jason Smith j...@apache.org wrote: Hi, Filipe. Most people seem to be holding their OTP build constant for these tests. If you have the time, would you please check out https://github.com/jhs/slow_couchdb It uses seatoncouch mixed with Bob's script to run a basic benchmark. I expect more template types to grow to help create different data profiles. Anyway, here are my results with 500k documents. Note that I built from your optimization commit, then its parent. https://gist.github.com/1928169 tl;dr = 2:50 before your commit; 4:13 after. On Mon, Feb 27, 2012 at 11:33 PM, Filipe David Manana fdman...@apache.org wrote: I just tried Jason's script (modified it to use 500 000 docs instead of 50 000) against 1.2.x and 1.1.1, using OTP R14B03. Here's my results: 1.2.x: $ port=5984 ./test.sh none Filling db. done HTTP/1.1 200 OK Server: CouchDB/1.2.0 (Erlang OTP/R14B03) Date: Mon, 27 Feb 2012 16:08:43 GMT Content-Type: text/plain; charset=utf-8 Content-Length: 252 Cache-Control: must-revalidate {db_name:db1,doc_count:51,doc_del_count:0,update_seq:51,purge_seq:0,compact_running:false,disk_size:130494577,data_size:130490673,instance_start_time:1330358830830086,disk_format_version:6,committed_update_seq:51} Building view. real 1m5.725s user 0m0.006s sys 0m0.005s done 1.1.1: $ port=5984 ./test.sh Filling db. done HTTP/1.1 200 OK Server: CouchDB/1.1.2a785d32f-git (Erlang OTP/R14B03) Date: Mon, 27 Feb 2012 16:15:33 GMT Content-Type: text/plain;charset=utf-8 Content-Length: 230 Cache-Control: must-revalidate
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Mon, Feb 27, 2012 at 7:18 PM, Filipe David Manana fdman...@apache.org wrote: Jason, can't reproduce those results, not even close: http://friendpaste.com/1L4pHH8WQchaLIMCWhKX9Z Before COUCHDB-1186 fdmanana 16:58:02 ~/git/hub/slow_couchdb (master) docs=50 batch=5 ./bench.sh small_doc.tpl Server: CouchDB/1.2.0a-a68a792-git (Erlang OTP/R14B03) {couchdb:Welcome,version:1.2.0a-a68a792-git} [INFO] Created DB named `db1' [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs Building view. {total_rows:50,offset:0,rows:[ {id:doc1,key:1,value:1} ]} real 0m56.241s user 0m0.006s sys 0m0.005s After COUCHDB-1186 fdmanana 17:02:02 ~/git/hub/slow_couchdb (master) docs=50 batch=5 ./bench.sh small_doc.tpl Server: CouchDB/1.2.0a-f023052-git (Erlang OTP/R14B03) {couchdb:Welcome,version:1.2.0a-f023052-git} [INFO] Created DB named `db1' [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs Building view. {total_rows:50,offset:0,rows:[ {id:doc1,key:1,value:1} ]} real 1m11.694s user 0m0.006s sys 0m0.005s fdmanana 17:06:01 ~/git/hub/slow_couchdb (master) 1.2.0a-f023052-git with patch http://friendpaste.com/178nPFgfyyeGf2vtNRpL0w applied on top fdmanana 17:06:53 ~/git/hub/slow_couchdb (master) docs=50 batch=5 ./bench.sh small_doc.tpl Server: CouchDB/1.2.0a-f023052-git (Erlang OTP/R14B03) {couchdb:Welcome,version:1.2.0a-f023052-git} [INFO] Created DB named `db1' [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs [INFO] Uploaded 5 documents via _bulk_docs Building view. {total_rows:50,offset:0,rows:[ {id:doc1,key:1,value:1} ]} real 0m51.089s user 0m0.006s sys 0m0.004s fdmanana 17:10:29 ~/git/hub/slow_couchdb (master) Can you try with R14B0x and also with the patch http://friendpaste.com/178nPFgfyyeGf2vtNRpL0w ? Back then I made all testing on a machine with a spinning disk, so the writer process was slower and likely dequeing more KV pairs from the work queue on each dequeue operation. The tests I did just now are on a machine with a ssd disk. Yeah, I've seen the btree behave quite differently on SSD's vs HDD's (same code had drastically different runtime characteristics). In other words, can we get a report of what type of disk everyone is running on? On Mon, Feb 27, 2012 at 4:40 PM, Jason Smith j...@apache.org wrote: Hi, Filipe. Most people seem to be holding their OTP build constant for these tests. If you have the time, would you please check out https://github.com/jhs/slow_couchdb It uses seatoncouch mixed with Bob's script to run a basic benchmark. I expect more template types to grow to help create different data profiles. Anyway, here are my results with 500k documents. Note that I built from your optimization commit, then its parent. https://gist.github.com/1928169 tl;dr = 2:50 before your commit; 4:13 after. On Mon, Feb 27, 2012 at 11:33 PM, Filipe David Manana fdman...@apache.org wrote: I just tried Jason's script (modified it to use 500 000 docs instead of 50 000) against 1.2.x and 1.1.1, using OTP R14B03. Here's my results: 1.2.x: $ port=5984 ./test.sh none Filling db. done HTTP/1.1 200 OK Server: CouchDB/1.2.0 (Erlang OTP/R14B03) Date: Mon, 27 Feb 2012 16:08:43 GMT Content-Type: text/plain; charset=utf-8 Content-Length: 252 Cache-Control: must-revalidate {db_name:db1,doc_count:51,doc_del_count:0,update_seq:51,purge_seq:0,compact_running:false,disk_size:130494577,data_size:130490673,instance_start_time:1330358830830086,disk_format_version:6,committed_update_seq:51} Building view. real 1m5.725s user 0m0.006s sys 0m0.005s done 1.1.1: $ port=5984 ./test.sh Filling db. done HTTP/1.1 200 OK
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
-1 R15B on OS X Lion I rebuilt OTP with an older SSL and that gets past all the crashes (thanks Filipe). I still see hangs when running make check, though any particular etap that hangs will run ok by itself. The Futon tests never run to completion in Chrome without hanging and the standalone JS tests also have fails. I tested the performance of view indexing, using a modest 200K doc db with a large complex view and there's a clear regression between 1.1.x and 1.2.x Others report similar results On Feb 23, 2012, at 5:25 PM, Bob Dionne wrote: sorry Noah, I'm in debug mode today so I don't care to start mucking with my stack, recompiling erlang, etc... I did try using that build repeatedly and it crashes all the time. I find it very odd and I had seen those before as I said on my older macbook. I do see the hangs Jan describes in the etaps, they have been there right along, so I'm confident this just the SSL issue. Why it only happens on the build is puzzling, any source build of any branch works just peachy. So I'd say I'm +1 based on my use of the 1.2.x branch but I'd like to hear from Stefan, who reported the severe performance regression. BobN seems to think we can ignore that, it's something flaky in that fellow's environment. I tend to agree but I'm conservative On Feb 23, 2012, at 1:23 PM, Noah Slater wrote: Can someone convince me this bus error stuff and segfaults is not a blocking issue. Bob tells me that he's followed the steps above and he's still experiencing the issues. Bob, you did follow the steps to install your own SSL right? On Thu, Feb 23, 2012 at 5:09 PM, Jan Lehnardt j...@apache.org wrote: On Feb 23, 2012, at 00:28 , Noah Slater wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting, Signature and hashes check out. Mac OS X 10.7.3, 64bit, SpiderMonkey 1.8.0, Erlang R14B04: make check works fine, browser tests in Safari work fine. Mac OS X 10.7.3, 64bit, SpiderMonkey 1.8.5, Erlang R14B04: make check works fine, browser tests in Safari work fine. FreeBSD 9.0, 64bit, SpiderMonkey 1.7.0, Erlang R14B04: make check works fine, browser tests in Safari work fine. CentOS 6.2, 64bit, SpiderMonkey 1.8.5, Erlang R14B04: make check works fine, browser tests in Firefox work fine. Ubuntu 11.4, 64bit, SpiderMonkey 1.8.5, Erlang R14B02: make check works fine, browser tests in Firefox work fine. Ubuntu 10.4, 32bit, SpiderMonkey 1.8.0, Erlang R13B03: make check fails in - 076-file-compression.t: https://gist.github.com/1893373 - 220-compaction-daemon.t: https://gist.github.com/1893387 This on runs in a VM and is 32bit, so I don't know if there's anything in the tests that rely on 64bittyness or the R14B03. Filipe, I think you worked on both features, do you have an idea? I tried running it all through Erlang R15B on Mac OS X 1.7.3, but a good way into `make check` the tests would just stop and hang. The last time, repeatedly in 160-vhosts.t, but when run alone, that test finished in under five seconds. I'm not sure what the issue is here. Despite the things above, I'm happy to give this a +1 if we put a warning about R15B on the download page. Great work all! Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Feb 26, 2012, at 13:43 , Robert Newson wrote: We're investigating a view indexing performance regression right now. I think it's still valuable to get community responses on whether the build artifact passes all the tests, though. If folks also want to try out building larger views, especially if they can compare the build time to CouchDB 1.1.1 that would be awesome. I ran Robert's test with a tiny doc and minimal view and saw the 25% performance drop, but using more realistic doc sizes and views show and improvement for 1.2.x (in one case 1k docs with a view that emits a complex key and an integer for the value). Filipe's https://github.com/fdmanana/seatoncouch makes it really easy to populate different sized docs and databases. If anyone could hack up a version of Robert's script combined with seatoncouch to make different sizes for documents, we can find out what to expect in performance differences. I'd say if we find one pathological case to be worse but prove across the board, things are faster, that this isn't a blocker. Cheers Jan -- B. On 24 February 2012 15:40, Benoit Chesneau bchesn...@gmail.com wrote: On Fri, Feb 24, 2012 at 4:39 PM, Benoit Chesneau bchesn...@gmail.com wrote: On Thu, Feb 23, 2012 at 12:28 AM, Noah Slater nsla...@tumbolia.org wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting, N +1 make check , js tests signatures ok erlang R14B04 - osx lion snow leopard, ubuntu 11.04, 11.10, debian 6, fbsd 8.2 - benoît q note: i had to buy my openssl to have it working with it on osx lion. Also I used spidermonkey 1.8.5.
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Feb 26, 2012, at 13:58 , Bob Dionne wrote: -1 R15B on OS X Lion I rebuilt OTP with an older SSL and that gets past all the crashes (thanks Filipe). I still see hangs when running make check, though any particular etap that hangs will run ok by itself. The Futon tests never run to completion in Chrome without hanging and the standalone JS tests also have fails. What part of this do you consider the -1? Can you try running the JS tests in Firefox and or Safari? Can you get all tests pass at least once across all browsers? The cli JS suite isn't supposed to work, so that isn't a criterion. I've seen the hang in make check for R15B while individual tests run as well, but I don't consider this blocking. While I understand and support the notion that tests shouldn't fail, period, we gotta work with what we have and master already has significant improvements. What would you like to see changed to not -1 this release? I tested the performance of view indexing, using a modest 200K doc db with a large complex view and there's a clear regression between 1.1.x and 1.2.x Others report similar results What is a large complex view? The complexity of the map/reduce functions is rarely an indicator of performance, it's usually input doc size and output/emit()/reduce data size. How big are the docs in your test and how big is the returned data? I understand the changes for 1.2.x will improve larger-data scenarios more significantly. Cheers Jan -- On Feb 23, 2012, at 5:25 PM, Bob Dionne wrote: sorry Noah, I'm in debug mode today so I don't care to start mucking with my stack, recompiling erlang, etc... I did try using that build repeatedly and it crashes all the time. I find it very odd and I had seen those before as I said on my older macbook. I do see the hangs Jan describes in the etaps, they have been there right along, so I'm confident this just the SSL issue. Why it only happens on the build is puzzling, any source build of any branch works just peachy. So I'd say I'm +1 based on my use of the 1.2.x branch but I'd like to hear from Stefan, who reported the severe performance regression. BobN seems to think we can ignore that, it's something flaky in that fellow's environment. I tend to agree but I'm conservative On Feb 23, 2012, at 1:23 PM, Noah Slater wrote: Can someone convince me this bus error stuff and segfaults is not a blocking issue. Bob tells me that he's followed the steps above and he's still experiencing the issues. Bob, you did follow the steps to install your own SSL right? On Thu, Feb 23, 2012 at 5:09 PM, Jan Lehnardt j...@apache.org wrote: On Feb 23, 2012, at 00:28 , Noah Slater wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting, Signature and hashes check out. Mac OS X 10.7.3, 64bit, SpiderMonkey 1.8.0, Erlang R14B04: make check works fine, browser tests in Safari work fine. Mac OS X 10.7.3, 64bit, SpiderMonkey 1.8.5, Erlang R14B04: make check works fine, browser tests in Safari work fine. FreeBSD 9.0, 64bit, SpiderMonkey 1.7.0, Erlang R14B04: make check works fine, browser tests in Safari work fine. CentOS 6.2, 64bit, SpiderMonkey 1.8.5, Erlang R14B04: make check works fine, browser tests in Firefox work fine. Ubuntu 11.4, 64bit, SpiderMonkey 1.8.5, Erlang R14B02: make check works fine, browser tests in Firefox work fine. Ubuntu 10.4, 32bit, SpiderMonkey 1.8.0, Erlang R13B03: make check fails in - 076-file-compression.t: https://gist.github.com/1893373 - 220-compaction-daemon.t: https://gist.github.com/1893387 This on runs in a VM and is 32bit, so I don't know if there's anything in the tests that rely on 64bittyness or the R14B03. Filipe, I think you worked on both features, do you have an idea? I tried running it all through Erlang R15B on Mac OS X 1.7.3, but a good way into `make check` the tests would just stop and hang. The last time, repeatedly in 160-vhosts.t, but when run alone, that test finished in under five seconds. I'm not sure what the issue is here. Despite the things above, I'm happy to give this a +1 if we put a warning about R15B on the download page. Great work all! Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Jan, I'm -1 based on all of my evaluation. I've spent a few hours on this release now yesterday and today. It doesn't really pass what I would call the smoke test. Almost everything I've run into has an explanation: 1. crashes out of the box - that's R15B, you need to recompile SSL and Erlang (we'll note on release notes) 2. etaps hang running make check. Known issue. Our etap code is out of date, recent versions of etap don't even run their own unit tests 3. Futon tests fail. Some are known bugs (attachment ranges in Chrome) . Both Chrome and Safari also hang 4. standalone JS tests fail. Again most of these run when run by themselves 5. performance. I used real production data *because* Stefan on user reported performance degradation on his data set. Any numbers are meaningless for a single test. I also ran scripts that BobN and Jason Smith posted that show a difference between 1.1.x and 1.2.x 6. Reviewed patch pointed to by Jason that may be the cause but it's hard to say without knowing the code analysis that went into the changes. You can see obvious local optimizations that make good sense but those are often the ones that get you, without knowing the call counts. Many of these issues can be explained away, but I think end users will be less forgiving. I think we already struggle with view performance. I'm interested to see how others evaluate this regression. I'll try this seatoncouch tool you mention later to see if I can construct some more definitive tests. Best, Bob On Feb 26, 2012, at 2:29 PM, Jan Lehnardt wrote: On Feb 26, 2012, at 13:58 , Bob Dionne wrote: -1 R15B on OS X Lion I rebuilt OTP with an older SSL and that gets past all the crashes (thanks Filipe). I still see hangs when running make check, though any particular etap that hangs will run ok by itself. The Futon tests never run to completion in Chrome without hanging and the standalone JS tests also have fails. What part of this do you consider the -1? Can you try running the JS tests in Firefox and or Safari? Can you get all tests pass at least once across all browsers? The cli JS suite isn't supposed to work, so that isn't a criterion. I've seen the hang in make check for R15B while individual tests run as well, but I don't consider this blocking. While I understand and support the notion that tests shouldn't fail, period, we gotta work with what we have and master already has significant improvements. What would you like to see changed to not -1 this release? I tested the performance of view indexing, using a modest 200K doc db with a large complex view and there's a clear regression between 1.1.x and 1.2.x Others report similar results What is a large complex view? The complexity of the map/reduce functions is rarely an indicator of performance, it's usually input doc size and output/emit()/reduce data size. How big are the docs in your test and how big is the returned data? I understand the changes for 1.2.x will improve larger-data scenarios more significantly. Cheers Jan -- On Feb 23, 2012, at 5:25 PM, Bob Dionne wrote: sorry Noah, I'm in debug mode today so I don't care to start mucking with my stack, recompiling erlang, etc... I did try using that build repeatedly and it crashes all the time. I find it very odd and I had seen those before as I said on my older macbook. I do see the hangs Jan describes in the etaps, they have been there right along, so I'm confident this just the SSL issue. Why it only happens on the build is puzzling, any source build of any branch works just peachy. So I'd say I'm +1 based on my use of the 1.2.x branch but I'd like to hear from Stefan, who reported the severe performance regression. BobN seems to think we can ignore that, it's something flaky in that fellow's environment. I tend to agree but I'm conservative On Feb 23, 2012, at 1:23 PM, Noah Slater wrote: Can someone convince me this bus error stuff and segfaults is not a blocking issue. Bob tells me that he's followed the steps above and he's still experiencing the issues. Bob, you did follow the steps to install your own SSL right? On Thu, Feb 23, 2012 at 5:09 PM, Jan Lehnardt j...@apache.org wrote: On Feb 23, 2012, at 00:28 , Noah Slater wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting,
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Bob, thanks for your reply I wasn't implying we should try to explain anything away. All of these are valid concerns, I just wanted to get a better understanding on where the bit flips from +0 to -1 and subsequently, how to address that boundary. Ideally we can just fix all of the things you mention, but I think it is important to understand them in detail, that's why I was going into them. Ultimately, I want to understand what we need to do to ship 1.2.0. On Feb 26, 2012, at 21:22 , Bob Dionne wrote: Jan, I'm -1 based on all of my evaluation. I've spent a few hours on this release now yesterday and today. It doesn't really pass what I would call the smoke test. Almost everything I've run into has an explanation: 1. crashes out of the box - that's R15B, you need to recompile SSL and Erlang (we'll note on release notes) Have we spent any time on figuring out what the trouble here is? 2. etaps hang running make check. Known issue. Our etap code is out of date, recent versions of etap don't even run their own unit tests I have seen the etap hang as well, and I wasn't diligent enough to report it in JIRA, I have done so now (COUCHDB-1424). 3. Futon tests fail. Some are known bugs (attachment ranges in Chrome) . Both Chrome and Safari also hang Do you have more details on where Chrome and Safari hang? Can you try their private browsing features, double/triple check that caches are empty? Can you get to a situation where you get all tests succeeding across all browsers, even if individual ones fail on one or two others? 4. standalone JS tests fail. Again most of these run when run by themselves Which ones? 5. performance. I used real production data *because* Stefan on user reported performance degradation on his data set. Any numbers are meaningless for a single test. I also ran scripts that BobN and Jason Smith posted that show a difference between 1.1.x and 1.2.x You are conflating an IRC discussion we've had into this thread. The performance regression reported is a good reason to look into other scenarios where we can show slowdowns. But we need to understand what's happening. Just from looking at dev@ all I see is some handwaving about some reports some people have done (Not to discourage any work that has been done on IRC and user@, but for the sake of a release vote thread, this related information needs to be on this mailing list). As I said on IRC, I'm happy to get my hands dirty to understand the regression at hand. But we need to know where we'd draw a line and say this isn't acceptable for a 1.2.0. 6. Reviewed patch pointed to by Jason that may be the cause but it's hard to say without knowing the code analysis that went into the changes. You can see obvious local optimizations that make good sense but those are often the ones that get you, without knowing the call counts. That is a point that wasn't included in your previous mail. It's great that there is progress, thanks for looking into this! Many of these issues can be explained away, but I think end users will be less forgiving. I think we already struggle with view performance. I'm interested to see how others evaluate this regression. I'll try this seatoncouch tool you mention later to see if I can construct some more definitive tests. Again, I'm not trying to explain anything away. I want to get a shared understanding of the issues you raised and where we stand on solving them squared against the ongoing 1.2.0 release. And again: Thanks for doing this thorough review and looking into performance issue. I hope with your help we can understand all these things a lot better very soon :) Cheers Jan -- Best, Bob On Feb 26, 2012, at 2:29 PM, Jan Lehnardt wrote: On Feb 26, 2012, at 13:58 , Bob Dionne wrote: -1 R15B on OS X Lion I rebuilt OTP with an older SSL and that gets past all the crashes (thanks Filipe). I still see hangs when running make check, though any particular etap that hangs will run ok by itself. The Futon tests never run to completion in Chrome without hanging and the standalone JS tests also have fails. What part of this do you consider the -1? Can you try running the JS tests in Firefox and or Safari? Can you get all tests pass at least once across all browsers? The cli JS suite isn't supposed to work, so that isn't a criterion. I've seen the hang in make check for R15B while individual tests run as well, but I don't consider this blocking. While I understand and support the notion that tests shouldn't fail, period, we gotta work with what we have and master already has significant improvements. What would you like to see changed to not -1 this release? I tested the performance of view indexing, using a modest 200K doc db with a large complex view and there's a clear regression between 1.1.x and 1.2.x Others report similar results What is a large complex view? The complexity of the
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
OS X 10.7.3, Erlang R15B, Spidermonkey 1.8.5, tests run in Chrome 19.0.1049.3 dev * signatures okay * make check okay * test suite okay (the attachment_ranges test constantly fails [1] in Chrome, but works fine in FF 10.0) So I'm +1 [1]: expected 'bytes 0-28/29', got 'bytes 0-29/29' expected '29', got '30' On 23.02.2012, at 00:28, Noah Slater wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting, N
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
attachment_ranges is expected to fail on Chrome because Chrome is wrong. to prove this more clearly, change line 29 of the test to Range: bytes=0-1000 and run it again, Chrome will tell you it fetched 1001 bytes of the 28 byte attachment. A neat trick. B. On 24 February 2012 09:03, Sebastian Cohnen sebastiancoh...@googlemail.com wrote: OS X 10.7.3, Erlang R15B, Spidermonkey 1.8.5, tests run in Chrome 19.0.1049.3 dev * signatures okay * make check okay * test suite okay (the attachment_ranges test constantly fails [1] in Chrome, but works fine in FF 10.0) So I'm +1 [1]: expected 'bytes 0-28/29', got 'bytes 0-29/29' expected '29', got '30' On 23.02.2012, at 00:28, Noah Slater wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting, N
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Thu, Feb 23, 2012 at 00:28, Noah Slater nsla...@tumbolia.org wrote: We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ Gentoo Linux 64-bits, Erlang 13B4, Spidermonkey 1.8.5. Signatures check out, make check passes. Browser tests pass in Firefox 12.0a2, although replicator_db_security failed once before it succeeded. +1 on release. Cheers, Dirkjan
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Fedora 15 64bit, Erlang 14B04, Spidermonkey 1.8.5 Everything checks out! Jonathan Porta On Fri, Feb 24, 2012 at 3:44 AM, Dirkjan Ochtman dirk...@ochtman.nl wrote: On Thu, Feb 23, 2012 at 00:28, Noah Slater nsla...@tumbolia.org wrote: We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ Gentoo Linux 64-bits, Erlang 13B4, Spidermonkey 1.8.5. Signatures check out, make check passes. Browser tests pass in Firefox 12.0a2, although replicator_db_security failed once before it succeeded. +1 on release. Cheers, Dirkjan
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Jonathan, you forgot to cast your vote! (Which needs to be either +1 or −1) On 24.02.2012, at 14:50, Jonathan Porta wrote: Fedora 15 64bit, Erlang 14B04, Spidermonkey 1.8.5 Everything checks out! Jonathan Porta On Fri, Feb 24, 2012 at 3:44 AM, Dirkjan Ochtman dirk...@ochtman.nl wrote: On Thu, Feb 23, 2012 at 00:28, Noah Slater nsla...@tumbolia.org wrote: We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ Gentoo Linux 64-bits, Erlang 13B4, Spidermonkey 1.8.5. Signatures check out, make check passes. Browser tests pass in Firefox 12.0a2, although replicator_db_security failed once before it succeeded. +1 on release. Cheers, Dirkjan
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Wed, Feb 22, 2012 at 11:28 PM, Noah Slater nsla...@tumbolia.org wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure +1 Good build, make check, browser tests pass: * Fedora 16, R15B * Ubuntu 11.10, R15B * Ubuntu 10.04 LTS, R15B * OS X Lion, R15B, no crashes or hangs either -- Iris Couch
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
+1! Jonathan Porta On Fri, Feb 24, 2012 at 7:07 AM, Sebastian Cohnen sebastiancoh...@googlemail.com wrote: Jonathan, you forgot to cast your vote! (Which needs to be either +1 or -1) On 24.02.2012, at 14:50, Jonathan Porta wrote: Fedora 15 64bit, Erlang 14B04, Spidermonkey 1.8.5 Everything checks out! Jonathan Porta On Fri, Feb 24, 2012 at 3:44 AM, Dirkjan Ochtman dirk...@ochtman.nl wrote: On Thu, Feb 23, 2012 at 00:28, Noah Slater nsla...@tumbolia.org wrote: We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ Gentoo Linux 64-bits, Erlang 13B4, Spidermonkey 1.8.5. Signatures check out, make check passes. Browser tests pass in Firefox 12.0a2, although replicator_db_security failed once before it succeeded. +1 on release. Cheers, Dirkjan
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Thu, Feb 23, 2012 at 12:28 AM, Noah Slater nsla...@tumbolia.org wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting, N +1 make check , js tests signatures ok erlang R14B04 - osx lion snow leopard, ubuntu 11.04, 11.10, debian 6, fbsd 8.2 - benoît
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Fri, Feb 24, 2012 at 4:39 PM, Benoit Chesneau bchesn...@gmail.com wrote: On Thu, Feb 23, 2012 at 12:28 AM, Noah Slater nsla...@tumbolia.org wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting, N +1 make check , js tests signatures ok erlang R14B04 - osx lion snow leopard, ubuntu 11.04, 11.10, debian 6, fbsd 8.2 - benoît q note: i had to buy my openssl to have it working with it on osx lion. Also I used spidermonkey 1.8.5.
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
+1 md5, sha, sig verified. make check passes. futon fully passes in FF 10.0. fully passes in Chrome 17 (except attachment_ranges). OS X 10.7.3, spidermonkey 1.8.5, icu 4.8.1.1. B. On 23 February 2012 04:39, Brian Mitchell binar...@gmail.com wrote: Same result as last vote on OS X 10.7.3 using Erlang OTP R15B and spidermonkey 1.8.5 . `make check` hangs on 076-file-compression.t and passes in the browser as long as I'm in private browsing mode or have the cache disabled. I'll do a test on Linux tomorrow with R14B04 before I +1. Brian. On Wednesday, February 22, 2012 at 18:28 , Noah Slater wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting, N
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
omitted to say this was R14B04, sorry. On 23 February 2012 11:33, Robert Newson rnew...@apache.org wrote: +1 md5, sha, sig verified. make check passes. futon fully passes in FF 10.0. fully passes in Chrome 17 (except attachment_ranges). OS X 10.7.3, spidermonkey 1.8.5, icu 4.8.1.1. B. On 23 February 2012 04:39, Brian Mitchell binar...@gmail.com wrote: Same result as last vote on OS X 10.7.3 using Erlang OTP R15B and spidermonkey 1.8.5 . `make check` hangs on 076-file-compression.t and passes in the browser as long as I'm in private browsing mode or have the cache disabled. I'll do a test on Linux tomorrow with R14B04 before I +1. Brian. On Wednesday, February 22, 2012 at 18:28 , Noah Slater wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting, N
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Thu, Feb 23, 2012 at 3:51 AM, Paul Davis paul.joseph.da...@gmail.com wrote: More than likely this is the SSL issue that Erlang has had issues with. Perhaps its time that I go in and see if I can't patch Erlang to use the newer OS X functions. Filipe had more details but IIRC the fix was basically to compile your own SSL and get Erlang to link against that. Yes smth like: https://github.com/refuge/refuge/wiki/Refuge-Build-on-MacOSX-Lion - benoît
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Bob, When I was suffering this, I would get bus errors or segmentation faults at random. They are non-deterministic, as far as I can tell. As I said at the start of this thread, I got a segfault on my first run through, and not my second or third. On Thu, Feb 23, 2012 at 3:06 AM, Bob Dionne dio...@dionne-associates.comwrote: perhaps, oddly if I try it without running make distcheck first I get a different error: Segmentation fault: 11 dependencies are fun On Feb 22, 2012, at 9:51 PM, Paul Davis wrote: More than likely this is the SSL issue that Erlang has had issues with. Perhaps its time that I go in and see if I can't patch Erlang to use the newer OS X functions. Filipe had more details but IIRC the fix was basically to compile your own SSL and get Erlang to link against that. On Wed, Feb 22, 2012 at 8:07 PM, Bob Dionne dio...@dionne-associates.com wrote: I've been doing that all along, master branch, 1.0.x, any source build is fine. I saw these occasionally before on the Mackbook Pro, this is the first one I've seen on the MBA It is odd that I'm only seeing it on this build, the only diff being I hardly ever run make distcheck. I'll try it without running that. On Feb 22, 2012, at 8:02 PM, Noah Slater wrote: Could you try again a few times maybe, try testing the source directly from your checkout? On Thu, Feb 23, 2012 at 12:26 AM, Bob Dionne dio...@dionne-associates.comwrote: I've seen this page before Noah, and something like these Bus errors. I think it might be the use of R15B What's odd is that I don't have any problems withe the 1.2 branch, or any branch for that matter On Feb 22, 2012, at 6:55 PM, Noah Slater wrote: Can you try this and report back please: http://wiki.apache.org/couchdb/Troubleshooting#Segmentation_faults_and_bus_errors On Wed, Feb 22, 2012 at 11:51 PM, Bob Dionne dio...@dionne-associates.comwrote: make distcheck ran ok server keeps crashing with: Bus error: 10 On Feb 22, 2012, at 6:29 PM, Noah Slater wrote: I might add that I got a segfault on the first try with the test suite, but was unable to reproduce it with two further runs. I would appreciate some feedback on this to assure me that it's nothing to worry about.
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Can someone convince me this bus error stuff and segfaults is not a blocking issue. Bob tells me that he's followed the steps above and he's still experiencing the issues. Bob, you did follow the steps to install your own SSL right? On Thu, Feb 23, 2012 at 5:09 PM, Jan Lehnardt j...@apache.org wrote: On Feb 23, 2012, at 00:28 , Noah Slater wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting, Signature and hashes check out. Mac OS X 10.7.3, 64bit, SpiderMonkey 1.8.0, Erlang R14B04: make check works fine, browser tests in Safari work fine. Mac OS X 10.7.3, 64bit, SpiderMonkey 1.8.5, Erlang R14B04: make check works fine, browser tests in Safari work fine. FreeBSD 9.0, 64bit, SpiderMonkey 1.7.0, Erlang R14B04: make check works fine, browser tests in Safari work fine. CentOS 6.2, 64bit, SpiderMonkey 1.8.5, Erlang R14B04: make check works fine, browser tests in Firefox work fine. Ubuntu 11.4, 64bit, SpiderMonkey 1.8.5, Erlang R14B02: make check works fine, browser tests in Firefox work fine. Ubuntu 10.4, 32bit, SpiderMonkey 1.8.0, Erlang R13B03: make check fails in - 076-file-compression.t: https://gist.github.com/1893373 - 220-compaction-daemon.t: https://gist.github.com/1893387 This on runs in a VM and is 32bit, so I don't know if there's anything in the tests that rely on 64bittyness or the R14B03. Filipe, I think you worked on both features, do you have an idea? I tried running it all through Erlang R15B on Mac OS X 1.7.3, but a good way into `make check` the tests would just stop and hang. The last time, repeatedly in 160-vhosts.t, but when run alone, that test finished in under five seconds. I'm not sure what the issue is here. Despite the things above, I'm happy to give this a +1 if we put a warning about R15B on the download page. Great work all! Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Thu, Feb 23, 2012 at 9:09 AM, Jan Lehnardt j...@apache.org wrote: On Feb 23, 2012, at 00:28 , Noah Slater wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting, Signature and hashes check out. Mac OS X 10.7.3, 64bit, SpiderMonkey 1.8.0, Erlang R14B04: make check works fine, browser tests in Safari work fine. Mac OS X 10.7.3, 64bit, SpiderMonkey 1.8.5, Erlang R14B04: make check works fine, browser tests in Safari work fine. FreeBSD 9.0, 64bit, SpiderMonkey 1.7.0, Erlang R14B04: make check works fine, browser tests in Safari work fine. CentOS 6.2, 64bit, SpiderMonkey 1.8.5, Erlang R14B04: make check works fine, browser tests in Firefox work fine. Ubuntu 11.4, 64bit, SpiderMonkey 1.8.5, Erlang R14B02: make check works fine, browser tests in Firefox work fine. Ubuntu 10.4, 32bit, SpiderMonkey 1.8.0, Erlang R13B03: make check fails in - 076-file-compression.t: https://gist.github.com/1893373 - 220-compaction-daemon.t: https://gist.github.com/1893387 This on runs in a VM and is 32bit, so I don't know if there's anything in the tests that rely on 64bittyness or the R14B03. Filipe, I think you worked on both features, do you have an idea? The compression fails because snappy NIF needs R13B04+ (forgot to update the test to work with R13B03 and below). For the compaction one, on ubuntu/debian you need to install the package 'erlang-os-mon' I tried running it all through Erlang R15B on Mac OS X 1.7.3, but a good way into `make check` the tests would just stop and hang. The last time, repeatedly in 160-vhosts.t, but when run alone, that test finished in under five seconds. I'm not sure what the issue is here. Despite the things above, I'm happy to give this a +1 if we put a warning about R15B on the download page. Great work all! Cheers Jan -- -- Filipe David Manana, Reasonable men adapt themselves to the world. Unreasonable men adapt the world to themselves. That's why all progress depends on unreasonable men.
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Feb 23, 2012, at 19:47 , Filipe David Manana wrote: On Thu, Feb 23, 2012 at 9:09 AM, Jan Lehnardt j...@apache.org wrote: On Feb 23, 2012, at 00:28 , Noah Slater wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting, Signature and hashes check out. Mac OS X 10.7.3, 64bit, SpiderMonkey 1.8.0, Erlang R14B04: make check works fine, browser tests in Safari work fine. Mac OS X 10.7.3, 64bit, SpiderMonkey 1.8.5, Erlang R14B04: make check works fine, browser tests in Safari work fine. FreeBSD 9.0, 64bit, SpiderMonkey 1.7.0, Erlang R14B04: make check works fine, browser tests in Safari work fine. CentOS 6.2, 64bit, SpiderMonkey 1.8.5, Erlang R14B04: make check works fine, browser tests in Firefox work fine. Ubuntu 11.4, 64bit, SpiderMonkey 1.8.5, Erlang R14B02: make check works fine, browser tests in Firefox work fine. Ubuntu 10.4, 32bit, SpiderMonkey 1.8.0, Erlang R13B03: make check fails in - 076-file-compression.t: https://gist.github.com/1893373 - 220-compaction-daemon.t: https://gist.github.com/1893387 This on runs in a VM and is 32bit, so I don't know if there's anything in the tests that rely on 64bittyness or the R14B03. Filipe, I think you worked on both features, do you have an idea? The compression fails because snappy NIF needs R13B04+ (forgot to update the test to work with R13B03 and below). For the compaction one, on ubuntu/debian you need to install the package 'erlang-os-mon' Thanks for the update Filipe! My +1 stands :) Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
+1 (checks passed on Arch Linux 3.2.6 64bit, R14B04, Spidermonkey 1.8.5) On Wednesday, February 22, 2012 at 23:39 , Brian Mitchell wrote: Same result as last vote on OS X 10.7.3 using Erlang OTP R15B and spidermonkey 1.8.5 . `make check` hangs on 076-file-compression.t and passes in the browser as long as I'm in private browsing mode or have the cache disabled. I'll do a test on Linux tomorrow with R14B04 before I +1. Brian. On Wednesday, February 22, 2012 at 18:28 , Noah Slater wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting, N
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
sorry Noah, I'm in debug mode today so I don't care to start mucking with my stack, recompiling erlang, etc... I did try using that build repeatedly and it crashes all the time. I find it very odd and I had seen those before as I said on my older macbook. I do see the hangs Jan describes in the etaps, they have been there right along, so I'm confident this just the SSL issue. Why it only happens on the build is puzzling, any source build of any branch works just peachy. So I'd say I'm +1 based on my use of the 1.2.x branch but I'd like to hear from Stefan, who reported the severe performance regression. BobN seems to think we can ignore that, it's something flaky in that fellow's environment. I tend to agree but I'm conservative On Feb 23, 2012, at 1:23 PM, Noah Slater wrote: Can someone convince me this bus error stuff and segfaults is not a blocking issue. Bob tells me that he's followed the steps above and he's still experiencing the issues. Bob, you did follow the steps to install your own SSL right? On Thu, Feb 23, 2012 at 5:09 PM, Jan Lehnardt j...@apache.org wrote: On Feb 23, 2012, at 00:28 , Noah Slater wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting, Signature and hashes check out. Mac OS X 10.7.3, 64bit, SpiderMonkey 1.8.0, Erlang R14B04: make check works fine, browser tests in Safari work fine. Mac OS X 10.7.3, 64bit, SpiderMonkey 1.8.5, Erlang R14B04: make check works fine, browser tests in Safari work fine. FreeBSD 9.0, 64bit, SpiderMonkey 1.7.0, Erlang R14B04: make check works fine, browser tests in Safari work fine. CentOS 6.2, 64bit, SpiderMonkey 1.8.5, Erlang R14B04: make check works fine, browser tests in Firefox work fine. Ubuntu 11.4, 64bit, SpiderMonkey 1.8.5, Erlang R14B02: make check works fine, browser tests in Firefox work fine. Ubuntu 10.4, 32bit, SpiderMonkey 1.8.0, Erlang R13B03: make check fails in - 076-file-compression.t: https://gist.github.com/1893373 - 220-compaction-daemon.t: https://gist.github.com/1893387 This on runs in a VM and is 32bit, so I don't know if there's anything in the tests that rely on 64bittyness or the R14B03. Filipe, I think you worked on both features, do you have an idea? I tried running it all through Erlang R15B on Mac OS X 1.7.3, but a good way into `make check` the tests would just stop and hang. The last time, repeatedly in 160-vhosts.t, but when run alone, that test finished in under five seconds. I'm not sure what the issue is here. Despite the things above, I'm happy to give this a +1 if we put a warning about R15B on the download page. Great work all! Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
I should clarify that I didn't mean ignore, but a test on virtual machines, of unknown provenance, that are merely similar is enough to make me very suspicious of any benchmark. If anyone wanted to perform a more rigorous and diligent test (like, say, *only* changing the couchdb source code between identical test runs), then I'd certainly want to halt the release if the 1/10th result were confirmed. b. On 23 February 2012 22:25, Bob Dionne dio...@dionne-associates.com wrote: sorry Noah, I'm in debug mode today so I don't care to start mucking with my stack, recompiling erlang, etc... I did try using that build repeatedly and it crashes all the time. I find it very odd and I had seen those before as I said on my older macbook. I do see the hangs Jan describes in the etaps, they have been there right along, so I'm confident this just the SSL issue. Why it only happens on the build is puzzling, any source build of any branch works just peachy. So I'd say I'm +1 based on my use of the 1.2.x branch but I'd like to hear from Stefan, who reported the severe performance regression. BobN seems to think we can ignore that, it's something flaky in that fellow's environment. I tend to agree but I'm conservative On Feb 23, 2012, at 1:23 PM, Noah Slater wrote: Can someone convince me this bus error stuff and segfaults is not a blocking issue. Bob tells me that he's followed the steps above and he's still experiencing the issues. Bob, you did follow the steps to install your own SSL right? On Thu, Feb 23, 2012 at 5:09 PM, Jan Lehnardt j...@apache.org wrote: On Feb 23, 2012, at 00:28 , Noah Slater wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting, Signature and hashes check out. Mac OS X 10.7.3, 64bit, SpiderMonkey 1.8.0, Erlang R14B04: make check works fine, browser tests in Safari work fine. Mac OS X 10.7.3, 64bit, SpiderMonkey 1.8.5, Erlang R14B04: make check works fine, browser tests in Safari work fine. FreeBSD 9.0, 64bit, SpiderMonkey 1.7.0, Erlang R14B04: make check works fine, browser tests in Safari work fine. CentOS 6.2, 64bit, SpiderMonkey 1.8.5, Erlang R14B04: make check works fine, browser tests in Firefox work fine. Ubuntu 11.4, 64bit, SpiderMonkey 1.8.5, Erlang R14B02: make check works fine, browser tests in Firefox work fine. Ubuntu 10.4, 32bit, SpiderMonkey 1.8.0, Erlang R13B03: make check fails in - 076-file-compression.t: https://gist.github.com/1893373 - 220-compaction-daemon.t: https://gist.github.com/1893387 This on runs in a VM and is 32bit, so I don't know if there's anything in the tests that rely on 64bittyness or the R14B03. Filipe, I think you worked on both features, do you have an idea? I tried running it all through Erlang R15B on Mac OS X 1.7.3, but a good way into `make check` the tests would just stop and hang. The last time, repeatedly in 160-vhosts.t, but when run alone, that test finished in under five seconds. I'm not sure what the issue is here. Despite the things above, I'm happy to give this a +1 if we put a warning about R15B on the download page. Great work all! Cheers Jan --
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
On Feb 23, 2012, at 00:28 , Noah Slater wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure +1 md5, sha, sig OK diff OK Windows 7 Enterprise x64 firefox 6.01 both R14B04 and R15B OpenSSL 0.9.8r SpiderMonkey 1.8.5 ICU 4.6.1 curl 7.23.1 - make check n/a binaries from https://people.apache.org/~dch/dist/1.2.0/ - both sigs OK - no malware found - user verification OK - futon OK Linux Mint LMDE x64 R14B03 ICU 4.4.2 OpenSSL 1.0.0d Spidermonkey libmozjs5d (JavaScript-C 1.8.0 pre-release 1 2007-10-03) - make distcheck OK - make check OK Firefox 9.0.1 - user verification OK - futon OK A+ Dave
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Are you going to call a separate vote on these Dave? On Thu, Feb 23, 2012 at 11:07 PM, Dave Cottlehuber d...@muse.net.nz wrote: On Feb 23, 2012, at 00:28 , Noah Slater wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure +1 md5, sha, sig OK diff OK Windows 7 Enterprise x64 firefox 6.01 both R14B04 and R15B OpenSSL 0.9.8r SpiderMonkey 1.8.5 ICU 4.6.1 curl 7.23.1 - make check n/a binaries from https://people.apache.org/~dch/dist/1.2.0/ - both sigs OK - no malware found - user verification OK - futon OK Linux Mint LMDE x64 R14B03 ICU 4.4.2 OpenSSL 1.0.0d Spidermonkey libmozjs5d (JavaScript-C 1.8.0 pre-release 1 2007-10-03) - make distcheck OK - make check OK Firefox 9.0.1 - user verification OK - futon OK A+ Dave
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
I might add that I got a segfault on the first try with the test suite, but was unable to reproduce it with two further runs. I would appreciate some feedback on this to assure me that it's nothing to worry about.
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
make distcheck ran ok server keeps crashing with: Bus error: 10 On Feb 22, 2012, at 6:29 PM, Noah Slater wrote: I might add that I got a segfault on the first try with the test suite, but was unable to reproduce it with two further runs. I would appreciate some feedback on this to assure me that it's nothing to worry about.
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Can you try this and report back please: http://wiki.apache.org/couchdb/Troubleshooting#Segmentation_faults_and_bus_errors On Wed, Feb 22, 2012 at 11:51 PM, Bob Dionne dio...@dionne-associates.comwrote: make distcheck ran ok server keeps crashing with: Bus error: 10 On Feb 22, 2012, at 6:29 PM, Noah Slater wrote: I might add that I got a segfault on the first try with the test suite, but was unable to reproduce it with two further runs. I would appreciate some feedback on this to assure me that it's nothing to worry about.
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
I've seen this page before Noah, and something like these Bus errors. I think it might be the use of R15B What's odd is that I don't have any problems withe the 1.2 branch, or any branch for that matter On Feb 22, 2012, at 6:55 PM, Noah Slater wrote: Can you try this and report back please: http://wiki.apache.org/couchdb/Troubleshooting#Segmentation_faults_and_bus_errors On Wed, Feb 22, 2012 at 11:51 PM, Bob Dionne dio...@dionne-associates.comwrote: make distcheck ran ok server keeps crashing with: Bus error: 10 On Feb 22, 2012, at 6:29 PM, Noah Slater wrote: I might add that I got a segfault on the first try with the test suite, but was unable to reproduce it with two further runs. I would appreciate some feedback on this to assure me that it's nothing to worry about.
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Could you try again a few times maybe, try testing the source directly from your checkout? On Thu, Feb 23, 2012 at 12:26 AM, Bob Dionne dio...@dionne-associates.comwrote: I've seen this page before Noah, and something like these Bus errors. I think it might be the use of R15B What's odd is that I don't have any problems withe the 1.2 branch, or any branch for that matter On Feb 22, 2012, at 6:55 PM, Noah Slater wrote: Can you try this and report back please: http://wiki.apache.org/couchdb/Troubleshooting#Segmentation_faults_and_bus_errors On Wed, Feb 22, 2012 at 11:51 PM, Bob Dionne dio...@dionne-associates.comwrote: make distcheck ran ok server keeps crashing with: Bus error: 10 On Feb 22, 2012, at 6:29 PM, Noah Slater wrote: I might add that I got a segfault on the first try with the test suite, but was unable to reproduce it with two further runs. I would appreciate some feedback on this to assure me that it's nothing to worry about.
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
I've been doing that all along, master branch, 1.0.x, any source build is fine. I saw these occasionally before on the Mackbook Pro, this is the first one I've seen on the MBA It is odd that I'm only seeing it on this build, the only diff being I hardly ever run make distcheck. I'll try it without running that. On Feb 22, 2012, at 8:02 PM, Noah Slater wrote: Could you try again a few times maybe, try testing the source directly from your checkout? On Thu, Feb 23, 2012 at 12:26 AM, Bob Dionne dio...@dionne-associates.comwrote: I've seen this page before Noah, and something like these Bus errors. I think it might be the use of R15B What's odd is that I don't have any problems withe the 1.2 branch, or any branch for that matter On Feb 22, 2012, at 6:55 PM, Noah Slater wrote: Can you try this and report back please: http://wiki.apache.org/couchdb/Troubleshooting#Segmentation_faults_and_bus_errors On Wed, Feb 22, 2012 at 11:51 PM, Bob Dionne dio...@dionne-associates.comwrote: make distcheck ran ok server keeps crashing with: Bus error: 10 On Feb 22, 2012, at 6:29 PM, Noah Slater wrote: I might add that I got a segfault on the first try with the test suite, but was unable to reproduce it with two further runs. I would appreciate some feedback on this to assure me that it's nothing to worry about.
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
More than likely this is the SSL issue that Erlang has had issues with. Perhaps its time that I go in and see if I can't patch Erlang to use the newer OS X functions. Filipe had more details but IIRC the fix was basically to compile your own SSL and get Erlang to link against that. On Wed, Feb 22, 2012 at 8:07 PM, Bob Dionne dio...@dionne-associates.com wrote: I've been doing that all along, master branch, 1.0.x, any source build is fine. I saw these occasionally before on the Mackbook Pro, this is the first one I've seen on the MBA It is odd that I'm only seeing it on this build, the only diff being I hardly ever run make distcheck. I'll try it without running that. On Feb 22, 2012, at 8:02 PM, Noah Slater wrote: Could you try again a few times maybe, try testing the source directly from your checkout? On Thu, Feb 23, 2012 at 12:26 AM, Bob Dionne dio...@dionne-associates.comwrote: I've seen this page before Noah, and something like these Bus errors. I think it might be the use of R15B What's odd is that I don't have any problems withe the 1.2 branch, or any branch for that matter On Feb 22, 2012, at 6:55 PM, Noah Slater wrote: Can you try this and report back please: http://wiki.apache.org/couchdb/Troubleshooting#Segmentation_faults_and_bus_errors On Wed, Feb 22, 2012 at 11:51 PM, Bob Dionne dio...@dionne-associates.comwrote: make distcheck ran ok server keeps crashing with: Bus error: 10 On Feb 22, 2012, at 6:29 PM, Noah Slater wrote: I might add that I got a segfault on the first try with the test suite, but was unable to reproduce it with two further runs. I would appreciate some feedback on this to assure me that it's nothing to worry about.
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
perhaps, oddly if I try it without running make distcheck first I get a different error: Segmentation fault: 11 dependencies are fun On Feb 22, 2012, at 9:51 PM, Paul Davis wrote: More than likely this is the SSL issue that Erlang has had issues with. Perhaps its time that I go in and see if I can't patch Erlang to use the newer OS X functions. Filipe had more details but IIRC the fix was basically to compile your own SSL and get Erlang to link against that. On Wed, Feb 22, 2012 at 8:07 PM, Bob Dionne dio...@dionne-associates.com wrote: I've been doing that all along, master branch, 1.0.x, any source build is fine. I saw these occasionally before on the Mackbook Pro, this is the first one I've seen on the MBA It is odd that I'm only seeing it on this build, the only diff being I hardly ever run make distcheck. I'll try it without running that. On Feb 22, 2012, at 8:02 PM, Noah Slater wrote: Could you try again a few times maybe, try testing the source directly from your checkout? On Thu, Feb 23, 2012 at 12:26 AM, Bob Dionne dio...@dionne-associates.comwrote: I've seen this page before Noah, and something like these Bus errors. I think it might be the use of R15B What's odd is that I don't have any problems withe the 1.2 branch, or any branch for that matter On Feb 22, 2012, at 6:55 PM, Noah Slater wrote: Can you try this and report back please: http://wiki.apache.org/couchdb/Troubleshooting#Segmentation_faults_and_bus_errors On Wed, Feb 22, 2012 at 11:51 PM, Bob Dionne dio...@dionne-associates.comwrote: make distcheck ran ok server keeps crashing with: Bus error: 10 On Feb 22, 2012, at 6:29 PM, Noah Slater wrote: I might add that I got a segfault on the first try with the test suite, but was unable to reproduce it with two further runs. I would appreciate some feedback on this to assure me that it's nothing to worry about.
Re: [VOTE] Apache CouchDB 1.2.0 release, second round
Same result as last vote on OS X 10.7.3 using Erlang OTP R15B and spidermonkey 1.8.5 . `make check` hangs on 076-file-compression.t and passes in the browser as long as I'm in private browsing mode or have the cache disabled. I'll do a test on Linux tomorrow with R14B04 before I +1. Brian. On Wednesday, February 22, 2012 at 18:28 , Noah Slater wrote: Hello, I would like call a vote for the Apache CouchDB 1.2.0 release, second round. We encourage the whole community to download and test these release artifacts so that any critical issues can be resolved before the release is made. Everyone is free to vote on this release, so get stuck in! We are voting on the following release artifacts: http://people.apache.org/~nslater/dist/1.2.0/ These artifacts have been built from the following tree-ish in Git: 4cd60f3d1683a3445c3248f48ae064fb573db2a1 Please follow the test procedure before voting: http://wiki.apache.org/couchdb/Test_procedure Thank you. Happy voting, N