[GitHub] spark pull request: [SPARK-4006] Block Manager - Double Register C...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/2854#issuecomment-108803453 @pwendell Done --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] Block Manager - Double Register C...
Github user tsliwowicz closed the pull request at: https://github.com/apache/spark/pull/2854 --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-6305] add log4j2 profile and prevent lo...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/4998#issuecomment-80878911 @srowen We don't know of an option to run side by side with two log4j versions. It conflicts on both slf4j and log4j classes. In any case, I believe it won't create a bigger problem because it just adds a build option. How is that a problem at all? For us, the log4j version is fixed for all drivers. The rest of the organisation switched to log4j 2, and we would like to do the same for Spark. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-6305] add log4j2 profile and prevent lo...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/4998#issuecomment-78868836 @srowen this is in essence the same as the ability to control hadoop version only through classpath manipulation. Since there are many flavors, its being done by using different profiles. While Spark may choose to leave the default log4j support for 1.2, the only way to fully switch to log4j 2 is via changing the classpath. Using the same logic that was applied by the spark designer for hadoop, @liorchaga created the profile for log4j 2. Adding a build profile does not complicate things because the default (that is not specifying the profile) means the build stays the same. Only those who need it, will use the profile. Same as for various hadoop versions. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] In long running contexts, we enco...
Github user tsliwowicz closed the pull request at: https://github.com/apache/spark/pull/2914 --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] In long running contexts, we enco...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/2914#issuecomment-67451754 hurray :-) On Thu, Dec 18, 2014 at 12:13 AM, andrewor14 notificati...@github.com wrote: Finally. I'm merging this into branch-1.0 thanks for your patience @tsliwowicz https://github.com/tsliwowicz â Reply to this email directly or view it on GitHub https://github.com/apache/spark/pull/2914#issuecomment-67404758. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] In long running contexts, we enco...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/2914#issuecomment-66446290 No problem. Glad to help :-) On Wed, Dec 10, 2014 at 4:44 AM, andrewor14 notificati...@github.com wrote: Hey sorry @tsliwowicz https://github.com/tsliwowicz for using your PRs as the battleground in fixing our builds against older branches. There aren't a lot of PRs opened against older branches so these tests aren't run in this context very often. So far I think all of these test failures have nothing to do with your patch so there is no action needed on your side. On our side, we'll keep investigating why the tests are failing all the time. â Reply to this email directly or view it on GitHub https://github.com/apache/spark/pull/2914#issuecomment-66396333. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] In long running contexts, we enco...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/2914#issuecomment-65783784 Seems like an issue with Jenkins --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] Block Manager - Double Register C...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/2854#issuecomment-65783822 Seems like an issue with Jenkins --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] In long running contexts, we enco...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/2914#issuecomment-60739832 Hi @andrewor14 - can I help somehow? I see that the PRs were not yet merged into 0.9 and 1.0 --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] Block Manager - Double Register C...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/2854#issuecomment-60389860 there seems to be some technical issue with the build. (not a real failure with the pull request itself) --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] In long running contexts, we enco...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/2914#issuecomment-60389844 there seems to be some technical issue with the build. (not a real failure with the pull request itself) --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] In long running contexts, we enco...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/2914#issuecomment-60410765 I was asked by @andrewor14 to open separate PRs because it does not merge cleanly. https://github.com/apache/spark/pull/2886 was approved and merged. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] In long running contexts, we enco...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/2914#issuecomment-60411019 @srowen I don't have a login to Jenkins so someone else needs to restart the build. Is there a way to get a login? I would gladly do it. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] In long running contexts, we enco...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/2914#issuecomment-60448957 @andrewor14 - thanks for your help! --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] In long running contexts, we enco...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/2886#issuecomment-60221362 @andrewor14 - thanks for the comments. I believe I fixed them all. Let me know! --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] In long running contexts, we enco...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/2886#issuecomment-60222733 the failure seems technical (not related to my fix), I think. Local maven build works fine for me. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] In long running contexts, we enco...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/2886#issuecomment-60306506 will do. Can you also merge into the 0.9 branch? I will update the PR I already have for it. https://github.com/apache/spark/pull/2854 --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] In long running contexts, we enco...
GitHub user tsliwowicz opened a pull request: https://github.com/apache/spark/pull/2914 [SPARK-4006] In long running contexts, we encountered the situation of d... ...ouble registe... ...r without a remove in between. The cause for that is unknown, and assumed a temp network issue. However, since the second register is with a BlockManagerId on a different port, blockManagerInfo.contains() returns false, while blockManagerIdByExecutor returns Some. This inconsistency is caught in a conditional statement that does System.exit(1), which is a huge robustness issue for us. The fix - simply remove the old id from both maps during register when this happens. We are mimicking the behavior of expireDeadHosts(), by doing local cleanup of the maps before trying to add new ones. Also - added some logging for register and unregister. This is just like https://github.com/apache/spark/pull/2886 except it's on master You can merge this pull request into a Git repository by running: $ git pull https://github.com/taboola/spark branch-1.0-block-mgr-removal Alternatively you can review and apply these changes as the patch at: https://github.com/apache/spark/pull/2914.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #2914 commit 1014493621016c596eb02eba4cf5228b0b834ef7 Author: Tal Sliwowicz ta...@taboola.com Date: 2014-10-23T20:26:26Z [SPARK-4006] In long running contexts, we encountered the situation of double registe... ...r without a remove in between. The cause for that is unknown, and assumed a temp network issue. However, since the second register is with a BlockManagerId on a different port, blockManagerInfo.contains() returns false, while blockManagerIdByExecutor returns Some. This inconsistency is caught in a conditional statement that does System.exit(1), which is a huge robustness issue for us. The fix - simply remove the old id from both maps during register when this happens. We are mimicking the behavior of expireDeadHosts(), by doing local cleanup of the maps before trying to add new ones. Also - added some logging for register and unregister. This is just like https://github.com/apache/spark/pull/2854 except it's on master Author: Tal Sliwowicz ta...@taboola.com Closes #2886 from tsliwowicz/master-block-mgr-removal and squashes the following commits: 094d508 [Tal Sliwowicz] some more white space change undone 41a2217 [Tal Sliwowicz] some more whitspaces change undone 7bcfc3d [Tal Sliwowicz] whitspaces fix df9d98f [Tal Sliwowicz] Code review comments fixed f48bce9 [Tal Sliwowicz] In long running contexts, we encountered the situation of double register without a remove in between. The cause for that is unknown, and assumed a temp network issue. (cherry picked from commit 6b485225271a3c616c4fa1231c20090a95c86f32) Conflicts: core/src/main/scala/org/apache/spark/storage/BlockManagerMasterActor.scala (cherry picked from commit d122236252d63635df7a112d92e90a2654702fc4) Conflicts: core/src/main/scala/org/apache/spark/storage/BlockManagerMasterActor.scala --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] In long running contexts, we enco...
GitHub user tsliwowicz opened a pull request: https://github.com/apache/spark/pull/2915 [SPARK-4006] In long running contexts, we encountered the situation of d... ...ouble registe... ...r without a remove in between. The cause for that is unknown, and assumed a temp network issue. However, since the second register is with a BlockManagerId on a different port, blockManagerInfo.contains() returns false, while blockManagerIdByExecutor returns Some. This inconsistency is caught in a conditional statement that does System.exit(1), which is a huge robustness issue for us. The fix - simply remove the old id from both maps during register when this happens. We are mimicking the behavior of expireDeadHosts(), by doing local cleanup of the maps before trying to add new ones. Also - added some logging for register and unregister. This is just like https://github.com/apache/spark/pull/2886 except it's on branch-1.1 You can merge this pull request into a Git repository by running: $ git pull https://github.com/taboola/spark branch-1.1-block-mgr-removal Alternatively you can review and apply these changes as the patch at: https://github.com/apache/spark/pull/2915.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #2915 commit d122236252d63635df7a112d92e90a2654702fc4 Author: Tal Sliwowicz ta...@taboola.com Date: 2014-10-23T20:13:41Z [SPARK-4006] In long running contexts, we encountered the situation of double registe... ...r without a remove in between. The cause for that is unknown, and assumed a temp network issue. However, since the second register is with a BlockManagerId on a different port, blockManagerInfo.contains() returns false, while blockManagerIdByExecutor returns Some. This inconsistency is caught in a conditional statement that does System.exit(1), which is a huge robustness issue for us. The fix - simply remove the old id from both maps during register when this happens. We are mimicking the behavior of expireDeadHosts(), by doing local cleanup of the maps before trying to add new ones. Also - added some logging for register and unregister. This is just like https://github.com/apache/spark/pull/2854 except it's on master Author: Tal Sliwowicz ta...@taboola.com Closes #2886 from tsliwowicz/master-block-mgr-removal and squashes the following commits: 094d508 [Tal Sliwowicz] some more white space change undone 41a2217 [Tal Sliwowicz] some more whitspaces change undone 7bcfc3d [Tal Sliwowicz] whitspaces fix df9d98f [Tal Sliwowicz] Code review comments fixed f48bce9 [Tal Sliwowicz] In long running contexts, we encountered the situation of double register without a remove in between. The cause for that is unknown, and assumed a temp network issue. (cherry picked from commit 6b485225271a3c616c4fa1231c20090a95c86f32) Conflicts: core/src/main/scala/org/apache/spark/storage/BlockManagerMasterActor.scala --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] In long running contexts, we enco...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/2886#issuecomment-60312692 @andrewor14 I created PR for 1.0, 1.1 and updated the 0.9 PR - can you please review and merge if ok? --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] Block Manager - Double Register C...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/2854#issuecomment-59922468 @andrewor14 - thanks, and sure - I will fix your comments and do a PR against master. However, re your logging comments, it really isn't that much. It adds a few lines of logging per run, which is insignificant, and it helps greatly to track registration and removal of block managers, which is really helpful in production to track issues. If you still think it's too much I will leave them out. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] In long running contexts, we enco...
GitHub user tsliwowicz opened a pull request: https://github.com/apache/spark/pull/2886 [SPARK-4006] In long running contexts, we encountered the situation of double registe... ...r without a remove in between. The cause for that is unknown, and assumed a temp network issue. However, since the second register is with a BlockManagerId on a different port, blockManagerInfo.contains() returns false, while blockManagerIdByExecutor returns Some. This inconsistency is caught in a conditional statement that does System.exit(1), which is a huge robustness issue for us. The fix - simply remove the old id from both maps during register when this happens. We are mimicking the behavior of expireDeadHosts(), by doing local cleanup of the maps before trying to add new ones. Also - added some logging for register and unregister. This is just like https://github.com/apache/spark/pull/2854 except it's on master You can merge this pull request into a Git repository by running: $ git pull https://github.com/taboola/spark master-block-mgr-removal Alternatively you can review and apply these changes as the patch at: https://github.com/apache/spark/pull/2886.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #2886 commit f48bce9cc25fa2672ea36bd90e64854159de8ead Author: Tal Sliwowicz ta...@taboola.com Date: 2014-10-21T14:29:39Z In long running contexts, we encountered the situation of double register without a remove in between. The cause for that is unknown, and assumed a temp network issue. However, since the second register is with a BlockManagerId on a different port, blockManagerInfo.contains() returns false, while blockManagerIdByExecutor returns Some. This inconsistency is caught in a conditional statement that does System.exit(1), which is a huge robustness issue for us. The fix - simply remove the old id from both maps during register when this happens. We are mimicking the behavior of expireDeadHosts(), by doing local cleanup of the maps before trying to add new ones. Also - added some logging for register and unregister. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: [SPARK-4006] Block Manager - Double Register C...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/2854#issuecomment-60017882 Created another pull request - https://github.com/apache/spark/pull/2886 - this time on master and also fixed the comments above. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: Block Manager - Double Register Crash
GitHub user tsliwowicz opened a pull request: https://github.com/apache/spark/pull/2854 Block Manager - Double Register Crash In long running contexts, we encountered the situation of double register without a remove in between. The cause for that is unknown, and assumed a temp network issue. However, since the second register is with a BlockManagerId on a different port, blockManagerInfo.contains() returns false, while blockManagerIdByExecutor returns Some. This inconsistency is caught in a conditional statement that does System.exit(1), which is a huge robustness issue for us. The fix - simply remove the old id from both maps during register when this happens. We are mimicking the behavior of expireDeadHosts(), by doing local cleanup of the maps before trying to add new ones. Also - added some logging for register and unregister. https://issues.apache.org/jira/browse/SPARK-4006 You can merge this pull request into a Git repository by running: $ git pull https://github.com/taboola/spark branch-0.9.2-block-mgr-removal Alternatively you can review and apply these changes as the patch at: https://github.com/apache/spark/pull/2854.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #2854 commit efd93f2026ddc427e84fa03e8a595ded2b1a81ce Author: Tal Sliwowicz ta...@taboola.com Date: 2014-10-12T08:35:20Z In long running contexts, we encountered the situation of double register without a remove in between. The cause for that is unknown, and assumed a temp network issue. However, since the second register is with a BlockManagerId on a different port, blockManagerInfo.contains() returns false, while blockManagerIdByExecutor returns Some. This inconsistency is caught in a conditional statement that does System.exit(1), which is a huge robustness issue for us. The fix - simply remove the old id from both maps during register when this happens. We are mimicking the behavior of expireDeadHosts(), by doing local cleanup of the maps before trying to add new ones. Also - added some logging for register and unregister. commit 81d69f088e421b19e47495d06e8b187a0ec29075 Author: Tal Sliwowicz ta...@taboola.com Date: 2014-10-12T08:41:53Z fixed comment --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request: mesos executor ids now consist of the slave id...
Github user tsliwowicz commented on the pull request: https://github.com/apache/spark/pull/1358#issuecomment-59724452 @mateiz - @KashiErez and I went on a different route. The killer issue was that there is a System.exit(1) in BlockManagerMasterActor which was a huge robustness issue for us. @taboola we are running some pretty large clusters (process many tera bytes of data / day) which do real time calculations and are mission critical. So - we fixed the issue and it's been running successfully in our production for a while now. I opened a new ticket - https://issues.apache.org/jira/browse/SPARK-4006 And a pull request - https://github.com/apache/spark/pull/2854 What do you think about our fix? --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
