Re: ******* Re: Time for 2.4.0 GA??
Thanks for the understanding. Experience is that win issues hardly land in the status. Mostly the answer is make a bug report, and it is of the table and sitting as reported bug for a very long time. Take for example the serious bug from Mario about the balancer in 2.3. An other are the mod_fcgid bugs. Also I cannot believe that eg the rewrite p issue was/is(?) only win. I propose that the ASF is also going to run windows on one of the heavier/complex sites !! Maybe a vote on that can started. Please consider. Steffen Op 16 jan. 2012 om 06:15 heeft Sander Temme scte...@apache.org het volgende geschreven: On Jan 12, 2012, at 2:24 AM, Steffen wrote: We have at least 4 hard bugs in 2.3.16. Known for a long time, and no need to exposure more for these. Are they listed in STATUS under the Release Showstoppers section? That currently only lists the need to remove undocumented modules as showstopper, and I believe we have even booked some progress on that one. Going by STATUS, things do indeed look stable, good and happy. This is not so much for folks like yourself who are intimately familiar with these bugs, but more for the rest of us who don't use Windows and may never encounter the problems. Fine a GA, with a big note that it is not ready for Windows and advising to run 2.2.21 as proven stable. I would hate that: we have lots of users on Windows and I'd like to enable them to use the latest, greatest. S. So not happy with 2.4. Op 11 jan. 2012 om 16:37 heeft Jim Jagielski j...@jagunet.com het volgende geschreven: Things look stable. Things look good. Things look happy. So are we ready to finally make this happen?? I offer to RM. -- scte...@apache.orghttp://www.temme.net/sander/ PGP FP: FC5A 6FC6 2E25 2DFD 8007 EE23 9BB8 63B0 F51B B88A View my availability: http://tungle.me/sctemme
Re: ******* Re: Time for 2.4.0 GA??
On 16 Jan 2012, at 2:31 PM, Steffen wrote: Thanks for the understanding. Experience is that win issues hardly land in the status. Mostly the answer is make a bug report, and it is of the table and sitting as reported bug for a very long time. Take for example the serious bug from Mario about the balancer in 2.3. An other are the mod_fcgid bugs. Also I cannot believe that eg the rewrite p issue was/is(?) only win. I propose that the ASF is also going to run windows on one of the heavier/complex sites !! Maybe a vote on that can started. Please consider. The best way to find and fix the Windows bugs is to get the code into the hands of the Windows users, and to do that we must release v2.4.0 as soon as possible. Trying to delay the release indefinitely until an arbitrary list of bugs is fixed is counterproductive - you're just keeping the code out of the hands of those who are able to either describe the problem in enough detail that it can be reproduced, or those who could potentially fix the problem and submit the fix to us. The dot zero in 2.4.0 is well understood. Regards, Graham -- smime.p7s Description: S/MIME cryptographic signature
Re: ******* Re: Time for 2.4.0 GA??
Also I cannot believe that eg the rewrite p issue was/is(?) only win. It wasn't windows-only.
Re: ******* Re: Time for 2.4.0 GA??
Unbelievable this answer, I am stimulating users to use it and with success. I feel, you are accusing me here and the windows community. No problem, we are continuing to let you know what issues we discover. In principal up to you how you handle issues Keep in mind we are just users, no dev's. When you and other dev's should participate at AL, should be good for the two worlds. So no way that the ASF would run a windows site, pity. For example I think Mario described the balancer issue quite good, but see no attention for a long time. More examples if you want. The former ipv6 issues is the winner, years before some give attention, even we (sob) supplied a fix. Also mod_fcgid has a fix provided by Mario, also not picked up. AL is providing a mod_fcgid binary with three fixes, so at some point we can solve things by ourselves. I already agreed to 2.4 GA, why wait so long. Not me is delaying. 2.3 is already in hands/running of at least a few thousands users for quite some time. Based on the number of downloads and the good/bad reports we get. Special the PHP community is a big downloader and use. Enough to see how 2.4 should go. I think 2.3 is great, some has bad luck with eg. no ssl usable. To run 2.3 without SSL issues, 2.2 in front(proxy) overcomes it. So no hurry to solve the issue. Steffen Op 16 jan. 2012 om 13:07 heeft Graham Leggett minf...@sharp.fm het volgende geschreven: On 16 Jan 2012, at 2:31 PM, Steffen wrote: Thanks for the understanding. Experience is that win issues hardly land in the status. Mostly the answer is make a bug report, and it is of the table and sitting as reported bug for a very long time. Take for example the serious bug from Mario about the balancer in 2.3. An other are the mod_fcgid bugs. Also I cannot believe that eg the rewrite p issue was/is(?) only win. I propose that the ASF is also going to run windows on one of the heavier/complex sites !! Maybe a vote on that can started. Please consider. The best way to find and fix the Windows bugs is to get the code into the hands of the Windows users, and to do that we must release v2.4.0 as soon as possible. Trying to delay the release indefinitely until an arbitrary list of bugs is fixed is counterproductive - you're just keeping the code out of the hands of those who are able to either describe the problem in enough detail that it can be reproduced, or those who could potentially fix the problem and submit the fix to us. The dot zero in 2.4.0 is well understood. Regards, Graham --
Re: Time for 2.4.0 GA??
On Thursday 12 January 2012, Jim Jagielski wrote: On monday (Jan 16th), I plan to TR 2.4.0... +1
Re: Time for 2.4.0 GA??
On Jan 12, 2012, at 2:24 AM, Steffen wrote: We have at least 4 hard bugs in 2.3.16. Known for a long time, and no need to exposure more for these. Are they listed in STATUS under the Release Showstoppers section? That currently only lists the need to remove undocumented modules as showstopper, and I believe we have even booked some progress on that one. Going by STATUS, things do indeed look stable, good and happy. This is not so much for folks like yourself who are intimately familiar with these bugs, but more for the rest of us who don't use Windows and may never encounter the problems. Fine a GA, with a big note that it is not ready for Windows and advising to run 2.2.21 as proven stable. I would hate that: we have lots of users on Windows and I'd like to enable them to use the latest, greatest. S. So not happy with 2.4. Op 11 jan. 2012 om 16:37 heeft Jim Jagielski j...@jagunet.com het volgende geschreven: Things look stable. Things look good. Things look happy. So are we ready to finally make this happen?? I offer to RM. -- scte...@apache.orghttp://www.temme.net/sander/ PGP FP: FC5A 6FC6 2E25 2DFD 8007 EE23 9BB8 63B0 F51B B88A View my availability: http://tungle.me/sctemme
Re: Time for 2.4.0 GA??
Yep. I supplied logs when asked. And Stefan was asked to provide a trace log, not me. Op 12 jan. 2012 om 18:11 heeft Rainer Jung rainer.j...@kippdata.de het volgende geschreven: On 12.01.2012 11:24, Steffen wrote: We have at least 4 hard bugs in 2.3.16. Known for a long time, and no need to exposure more for these. Fine a GA, with a big note that it is not ready for Windows and advising to run 2.2.21 as proven stable. So not happy with 2.4. Understood, but we repeatedly asked whether you are anyone who can reproduce the problem could please provide an ErrorLog on LogLevel trace8 for one broken and one good request. We didn't get an answer to that. Can you provide it? Note that neither Bill nor me were able to reproduce on our local Windows systems. So we need some help to hunt the bug. Regards, Rainer
Re: Time for 2.4.0 GA??
It is the SSL issue when AcceptFilter https none (not httpd what Gregg says). More info, see the thread: Win 2.3.16 :: SSL and AcceptFilter Better to discuss there and not clutter this thread. Steffen Op 13 jan. 2012 om 01:29 heeft Daniel Ruggeri drugg...@primary.net het volgende geschreven: On 1/12/2012 5:50 PM, Gregg L. Smith wrote: Either apachehaus.com or apachelounge.com have 2.3.16 binaries available for Windows. The problem is with the directive; AcceptFilter httpd none That is the only non-stardard config option. Greg; Thanks for the overview - if I understand correctly this seems to manifest with socket reuse and incomplete or empty responses - correct? And that this is a matter of recycling existing listening sockets rather than anything keepalive-related. If so, I'm guessing a simple perl LWP script to make request after request for a known-size resource (and verifying the size returned) would do. Any other pointers or things to look for? I seem to recall SSL being brought up, or is that a separate issue? -- Daniel Ruggeri
Re: Time for 2.4.0 GA??
Better to discuss there and not clutter this thread. Even better to open a bug report.
Re: Time for 2.4.0 GA??
On 13.01.2012 10:18, Steffen wrote: Yep. I supplied logs when asked. And Stefan was asked to provide a trace log, not me. I know you provided some, but it was only LogLevel info and without any timestamps. What would be very helpful in addition would be the full trace8 log files of a good and a bad response. The name Stefan was a typo at that type and was corrected shortly after on the same thread to be meant as Steffen ;) I think Stefan doesn't have a Win build and environment to reproduce. Regards, Rainer Op 12 jan. 2012 om 18:11 heeft Rainer Jungrainer.j...@kippdata.de het volgende geschreven: On 12.01.2012 11:24, Steffen wrote: We have at least 4 hard bugs in 2.3.16. Known for a long time, and no need to exposure more for these. Fine a GA, with a big note that it is not ready for Windows and advising to run 2.2.21 as proven stable. So not happy with 2.4. Understood, but we repeatedly asked whether you are anyone who can reproduce the problem could please provide an ErrorLog on LogLevel trace8 for one broken and one good request. We didn't get an answer to that. Can you provide it? Note that neither Bill nor me were able to reproduce on our local Windows systems. So we need some help to hunt the bug. Regards, Rainer
Re: Time for 2.4.0 GA??
We have at least 4 hard bugs in 2.3.16. Known for a long time, and no need to exposure more for these. Fine a GA, with a big note that it is not ready for Windows and advising to run 2.2.21 as proven stable. So not happy with 2.4. Op 11 jan. 2012 om 16:37 heeft Jim Jagielski j...@jagunet.com het volgende geschreven: Things look stable. Things look good. Things look happy. So are we ready to finally make this happen?? I offer to RM.
Re: Time for 2.4.0 GA??
On monday (Jan 16th), I plan to TR 2.4.0...
Re: Time for 2.4.0 GA??
On 12.01.2012 11:24, Steffen wrote: We have at least 4 hard bugs in 2.3.16. Known for a long time, and no need to exposure more for these. Fine a GA, with a big note that it is not ready for Windows and advising to run 2.2.21 as proven stable. So not happy with 2.4. Understood, but we repeatedly asked whether you are anyone who can reproduce the problem could please provide an ErrorLog on LogLevel trace8 for one broken and one good request. We didn't get an answer to that. Can you provide it? Note that neither Bill nor me were able to reproduce on our local Windows systems. So we need some help to hunt the bug. Regards, Rainer
Re: Time for 2.4.0 GA??
On 12 Jan 2012, at 18:10, Jim Jagielski j...@jagunet.com wrote: On monday (Jan 16th), I plan to TR 2.4.0... +1. Let's do this. Regards, Graham --
Re: Time for 2.4.0 GA??
+1 non binding. Please note in the announcement that for Windows there are still issues pending, special not working SSL, hanging workers, balancer and Rewrite Proxy. For production use, is 2.2.21 advised. Op 12 jan. 2012 om 18:10 heeft Jim Jagielski j...@jagunet.com het volgende geschreven: On monday (Jan 16th), I plan to TR 2.4.0...
Re: Time for 2.4.0 GA??
On 12.01.2012 19:10, Jim Jagielski wrote: On monday (Jan 16th), I plan to TR 2.4.0... +1
Re: Time for 2.4.0 GA??
On 1/12/2012 10:11 AM, Rainer Jung wrote: On 12.01.2012 11:24, Steffen wrote: We have at least 4 hard bugs in 2.3.16. Known for a long time, and no need to exposure more for these. Fine a GA, with a big note that it is not ready for Windows and advising to run 2.2.21 as proven stable. So not happy with 2.4. Understood, but we repeatedly asked whether you are anyone who can reproduce the problem could please provide an ErrorLog on LogLevel trace8 for one broken and one good request. We didn't get an answer to that. Can you provide it? Note that neither Bill nor me were able to reproduce on our local Windows systems. So we need some help to hunt the bug. Regards, Rainer Yes, and I currently have my logging of mod_ssl and the mpm at trace, unfortunately, I have not been lucky to see a failure (partial/blank screen) while trace logging is on, odd, but bad luck. It's not that I have not tried. I am seeing this though; [Thu Jan 05 14:59:55.161420 2012] [ssl:info] [pid 1232:tid 856] [client 192.168.1.1:62717] AH01964: Connection to child 57 established (server www.somedomain.tld:443) [Thu Jan 05 14:59:55.161420 2012] [ssl:trace3] [pid 1232:tid 844] ssl_engine_kernel.c(1840): [client 192.168.1.1:62716] OpenSSL: Loop: before/accept initialization [Thu Jan 05 14:59:55.161420 2012] [ssl:trace2] [pid 1232:tid 856] ssl_engine_rand.c(123): Seeding PRNG with 144 bytes of entropy [Thu Jan 05 14:59:55.161420 2012] [ssl:trace4] [pid 1232:tid 844] ssl_engine_io.c(2000): [client 192.168.1.1:62716] OpenSSL: I/O error, 11 bytes expected to read on BIO#12e0ef0 [mem: 27eb6a8] [Thu Jan 05 14:59:55.161420 2012] [ssl:trace3] [pid 1232:tid 844] ssl_engine_kernel.c(1869): [client 192.168.1.1:62716] OpenSSL: Exit: error in SSLv2/v3 read client hello A [Thu Jan 05 14:59:55.161420 2012] [ssl:trace4] [pid 1232:tid 844] ssl_engine_io.c(2000): [client 192.168.1.1:62716] OpenSSL: I/O error, 11 bytes expected to read on BIO#12e0ef0 [mem: 27eb6a8] [Thu Jan 05 14:59:55.161420 2012] [ssl:trace3] [pid 1232:tid 856] ssl_engine_kernel.c(1832): [client 192.168.1.1:62717] OpenSSL: Handshake: start [Thu Jan 05 14:59:55.161420 2012] [ssl:trace3] [pid 1232:tid 844] ssl_engine_kernel.c(1869): [client 192.168.1.1:62716] OpenSSL: Exit: error in SSLv2/v3 read client hello A [Thu Jan 05 14:59:55.161420 2012] [ssl:trace3] [pid 1232:tid 856] ssl_engine_kernel.c(1840): [client 192.168.1.1:62717] OpenSSL: Loop: before/accept initialization [Thu Jan 05 14:59:55.161420 2012] [ssl:trace4] [pid 1232:tid 856] ssl_engine_io.c(2000): [client 192.168.1.1:62717] OpenSSL: I/O error, 11 bytes expected to read on BIO#142eae0 [mem: 27f9158] [Thu Jan 05 14:59:55.161420 2012] [ssl:trace3] [pid 1232:tid 856] ssl_engine_kernel.c(1869): [client 192.168.1.1:62717] OpenSSL: Exit: error in SSLv2/v3 read client hello A [Thu Jan 05 14:59:55.161420 2012] [ssl:trace4] [pid 1232:tid 856] ssl_engine_io.c(2000): [client 192.168.1.1:62717] OpenSSL: I/O error, 11 bytes expected to read on BIO#142eae0 [mem: 27f9158] [Thu Jan 05 14:59:55.161420 2012] [ssl:trace3] [pid 1232:tid 856] ssl_engine_kernel.c(1869): [client 192.168.1.1:62717] OpenSSL: Exit: error in SSLv2/v3 read client hello A [Thu Jan 05 14:59:55.208311 2012] [ssl:trace4] [pid 1232:tid 848] ssl_engine_io.c(1989): [client 192.168.1.1:62713] OpenSSL: read 5/5 bytes from BIO#12c87f0 [mem: 133f643] (BIO dump follows) [Thu Jan 05 14:59:55.208311 2012] [ssl:trace7] [pid 1232:tid 848] ssl_engine_io.c(1922): +-+ I'll hammer on it as time permits the next couple of days. I'm using a real cert, not self-signed. I'm +1 for 2.4.0, as to GA or beta can be decided then. Regard, Gregg
Re: Time for 2.4.0 GA??
On 1/11/2012 5:52 PM, Daniel Ruggeri wrote: On 1/11/2012 6:26 PM, Noel Butler wrote: Also, in relation to the windows stuff, I think Jim's suggestion is most appropriate in the absence of more windows users to test. If all we need is testing and some logging, I'd be happy to assist. If someone can provide the compiled build and any non-standard config directives with an overview of how to duplicate, I have Win7, Server 2003 and Server 2008 I can test with. I fall a bit short in how much volume I could throw at an installation, though Hi Daniel, Either apachehaus.com or apachelounge.com have 2.3.16 binaries available for Windows. The problem is with the directive; AcceptFilter httpd none That is the only non-stardard config option. Cheers, Gregg
Re: Time for 2.4.0 GA??
On 1/12/2012 5:50 PM, Gregg L. Smith wrote: Either apachehaus.com or apachelounge.com have 2.3.16 binaries available for Windows. The problem is with the directive; AcceptFilter httpd none That is the only non-stardard config option. Greg; Thanks for the overview - if I understand correctly this seems to manifest with socket reuse and incomplete or empty responses - correct? And that this is a matter of recycling existing listening sockets rather than anything keepalive-related. If so, I'm guessing a simple perl LWP script to make request after request for a known-size resource (and verifying the size returned) would do. Any other pointers or things to look for? I seem to recall SSL being brought up, or is that a separate issue? -- Daniel Ruggeri
Re: Time for 2.4.0 GA??
On 1/12/2012 5:29 PM, Daniel Ruggeri wrote: On 1/12/2012 5:50 PM, Gregg L. Smith wrote: Either apachehaus.com or apachelounge.com have 2.3.16 binaries available for Windows. The problem is with the directive; AcceptFilter httpd none That is the only non-stardard config option. Greg; Thanks for the overview - if I understand correctly this seems to manifest with socket reuse and incomplete or empty responses - correct? And that this is a matter of recycling existing listening sockets rather than anything keepalive-related. If so, I'm guessing a simple perl LWP script to make request after request for a known-size resource (and verifying the size returned) would do. Any other pointers or things to look for? I seem to recall SSL being brought up, or is that a separate issue? Hi Daniel, I typo'd that again. I am bad at that, I should just copy and paste from my config. AcceptFilter https none SSL has everything to do with it. It is just in SSL with AcceptFilter for that protocol set to none I find with trace logging on it doesn't want to manifest itself. When I turn off the trace logging, it happens often. It's odd, I do not get it. Yes, incomplete or empty responses. The empty ones are easy to spot. Just an image here or there is a possible incomplete. It may be keepalive. On some windows flavors, we do not have a huge pool of connections and can be DOSed rather easy. I have seen some errors while trace logging was on, that may be suggesting that. Looking, Keepalive is high, at 30, the default has been whittled down since I configured the server the 1st time (and have kept these config files) . Timeout is at 60. Even at 60, I have seen this; [Thu Jan 12 15:38:31.812658 2012] [ssl:info] [pid 3504:tid 720] (OS 10060)A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond. : [client 192.168.1.1:51417] AH01991: SSL input filter read failed. However, I can say this does not happen to just me, Steffen has seen same behavior. It seems to happen more pronounced on my SNI hosts that the default SSL host. So maybe it is an artifact of SNI. A lot of possibilities I just cannot quite pin down. II do not really have much problems with *not* setting AcceptFilter to none on https, not like I have on standard http where it has to be set to none or the server just stops responding after some time.
Time for 2.4.0 GA??
Things look stable. Things look good. Things look happy. So are we ready to finally make this happen?? I offer to RM.
Re: Time for 2.4.0 GA??
On Wed, Jan 11, 2012 at 11:37 AM, Jim Jagielski j...@jagunet.com wrote: Things look stable. Things look good. Things look happy. So are we ready to finally make this happen?? I offer to RM. Only outstanding issue from last series of thread I am aware of is windows ssl socket reuse thing that causes sporadic empty responses. (I may be taking some serious liberties with this description) I don't know how we'd communicate that to early windows adopters outside of e.g. the apachelounge community.
Re: Time for 2.4.0 GA??
On 11 Jan 2012, at 7:38 PM, Eric Covener wrote: On Wed, Jan 11, 2012 at 11:37 AM, Jim Jagielski j...@jagunet.com wrote: Things look stable. Things look good. Things look happy. So are we ready to finally make this happen?? I offer to RM. +1. Only outstanding issue from last series of thread I am aware of is windows ssl socket reuse thing that causes sporadic empty responses. (I may be taking some serious liberties with this description) I don't know how we'd communicate that to early windows adopters outside of e.g. the apachelounge community. This is a bug that will benefit from wider exposure, and declaring v2.4.0 done will give us that wider exposure. Regards, Graham -- smime.p7s Description: S/MIME cryptographic signature
Re: Time for 2.4.0 GA??
On 1/11/2012 10:37 AM, Jim Jagielski wrote: Things look stable. Things look good. Things look happy. So are we ready to finally make this happen?? I offer to RM. I'm voting -1 on GA until we see something that we agree is GA :) Totally +1 on making 2.4.0 happen, whether it is alpha, beta or GA caliber.
Re: Time for 2.4.0 GA??
On Jan 11, 2012, at 2:48 PM, William A. Rowe Jr. wrote: On 1/11/2012 10:37 AM, Jim Jagielski wrote: Things look stable. Things look good. Things look happy. So are we ready to finally make this happen?? I offer to RM. I'm voting -1 on GA until we see something that we agree is GA :) What currently is not GA-worthy?
Re: Time for 2.4.0 GA??
On Jan 11, 2012, at 12:38 PM, Eric Covener wrote: On Wed, Jan 11, 2012 at 11:37 AM, Jim Jagielski j...@jagunet.com wrote: Things look stable. Things look good. Things look happy. So are we ready to finally make this happen?? I offer to RM. Only outstanding issue from last series of thread I am aware of is windows ssl socket reuse thing that causes sporadic empty responses. I don't know how we'd communicate that to early windows adopters outside of e.g. the apachelounge community. Same way we did with 1.3.0 maybe? At present, the Win95/NT port of Apache is not as stable as the UNIX version. Further releases of the 1.3.x tree will bring the Win95/NT port closer to parity.
Re: Time for 2.4.0 GA??
On 1/11/2012 2:17 PM, Jim Jagielski wrote: What currently is not GA-worthy? We vote on releases, not subversion. That was my point. It is not yet the best available version, I believe 2.2.21 is. There is far too much unexercised code in 2.4.0, new regressions and so forth, that will only be detected by adoption of 2.4.0. But the 2.1 and 2.3 cycles have taught us users aren't adopting our 2.odd releases. Sigh.
Re: Time for 2.4.0 GA??
On 1/11/2012 5:51 PM, Noel Butler wrote: On Wed, 2012-01-11 at 14:41 -0600, William A. Rowe Jr. wrote: cycles have taught us users aren't adopting our 2.odd releases. Sigh. This is likely a carry over from the old days of the kernel, no-one dares run a major.odd on production boxes :) Which is why I've been hinting at 2.4.0-beta, but I don't the the message is received. Maybe I'm being too opaque?
Re: Time for 2.4.0 GA??
On Wed, 2012-01-11 at 18:00 -0600, William A. Rowe Jr. wrote: On 1/11/2012 5:51 PM, Noel Butler wrote: On Wed, 2012-01-11 at 14:41 -0600, William A. Rowe Jr. wrote: cycles have taught us users aren't adopting our 2.odd releases. Sigh. This is likely a carry over from the old days of the kernel, no-one dares run a major.odd on production boxes :) Which is why I've been hinting at 2.4.0-beta, but I don't the the message is received. Maybe I'm being too opaque? I'm all for that, better to nut out as much as possible before declaring a new version as recommended GA. Please lets not get into the mindset of certain distros where release often is a policy even when its not ready with bugs - release when stable and proved, is far far far better. Also, in relation to the windows stuff, I think Jim's suggestion is most appropriate in the absence of more windows users to test. signature.asc Description: This is a digitally signed message part
Re: Time for 2.4.0 GA??
On 1/11/2012 6:26 PM, Noel Butler wrote: Also, in relation to the windows stuff, I think Jim's suggestion is most appropriate in the absence of more windows users to test. If all we need is testing and some logging, I'd be happy to assist. If someone can provide the compiled build and any non-standard config directives with an overview of how to duplicate, I have Win7, Server 2003 and Server 2008 I can test with. I fall a bit short in how much volume I could throw at an installation, though. -- Daniel Ruggeri
