[jira] [Commented] (OAK-2609) Thread.interrupt seems to stop repository
[ https://issues.apache.org/jira/browse/OAK-2609?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14371139#comment-14371139 ] Michael Dürig commented on OAK-2609: +1 Thread.interrupt seems to stop repository - Key: OAK-2609 URL: https://issues.apache.org/jira/browse/OAK-2609 Project: Jackrabbit Oak Issue Type: Bug Affects Versions: 1.1.7 Environment: Java 7 Reporter: Carsten Ziegeler Priority: Critical Fix For: 1.3.0 We have a sporadic problem with Sling's JCR installer 3.3.8 and Oak (tar mk). It seems to timing related: the JCR installer does a Thread#interrupt at one point and sometimes this brings the hole instance to stop. Nothing else is going on any more. While of course, a workaround is to remove the Thread.interrupt call in the JCR installer (which we did, see SLING-4477), I have the fear that this can happen with any code that is using the repository and gets interrupted. This error is hard to reproduce, however with three people testing we could see this several times happening -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (OAK-2609) Thread.interrupt seems to stop repository
[ https://issues.apache.org/jira/browse/OAK-2609?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14368763#comment-14368763 ] Michael Marth commented on OAK-2609: would it make sense to simply document the implications of Thread.interrupt (and advise not to use it)? Thread.interrupt seems to stop repository - Key: OAK-2609 URL: https://issues.apache.org/jira/browse/OAK-2609 Project: Jackrabbit Oak Issue Type: Bug Affects Versions: 1.1.7 Environment: Java 7 Reporter: Carsten Ziegeler Priority: Critical Fix For: 1.3.0 We have a sporadic problem with Sling's JCR installer 3.3.8 and Oak (tar mk). It seems to timing related: the JCR installer does a Thread#interrupt at one point and sometimes this brings the hole instance to stop. Nothing else is going on any more. While of course, a workaround is to remove the Thread.interrupt call in the JCR installer (which we did, see SLING-4477), I have the fear that this can happen with any code that is using the repository and gets interrupted. This error is hard to reproduce, however with three people testing we could see this several times happening -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (OAK-2609) Thread.interrupt seems to stop repository
[ https://issues.apache.org/jira/browse/OAK-2609?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14363255#comment-14363255 ] Michael Dürig commented on OAK-2609: I don't think we should do anything here on the Oak side as we are dealing with [documented side effects|http://docs.oracle.com/javase/7/docs/api/java/nio/channels/InterruptibleChannel.html] of {{Thread.interrupt}}: Thread.interrupt seems to stop repository - Key: OAK-2609 URL: https://issues.apache.org/jira/browse/OAK-2609 Project: Jackrabbit Oak Issue Type: Bug Affects Versions: 1.1.7 Environment: Java 7 Reporter: Carsten Ziegeler Priority: Critical Fix For: 1.1.8, 1.0.13 We have a sporadic problem with Sling's JCR installer 3.3.8 and Oak (tar mk). It seems to timing related: the JCR installer does a Thread#interrupt at one point and sometimes this brings the hole instance to stop. Nothing else is going on any more. While of course, a workaround is to remove the Thread.interrupt call in the JCR installer (which we did, see SLING-4477), I have the fear that this can happen with any code that is using the repository and gets interrupted. This error is hard to reproduce, however with three people testing we could see this several times happening -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (OAK-2609) Thread.interrupt seems to stop repository
[ https://issues.apache.org/jira/browse/OAK-2609?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14363484#comment-14363484 ] Michael Dürig commented on OAK-2609: bq. Can this at least be catched somehow with a big log message? The problem is probably that there is no single place nor a single exception to catch. To cover them all we'd need to go through each and every place where something in Oak does a blocking operation and check the side effects. Maybe a good compromise is to come up with a test case provoking the problem and just cover the cases we see coming up with that assuming those will be the most frequent ones. Thread.interrupt seems to stop repository - Key: OAK-2609 URL: https://issues.apache.org/jira/browse/OAK-2609 Project: Jackrabbit Oak Issue Type: Bug Affects Versions: 1.1.7 Environment: Java 7 Reporter: Carsten Ziegeler Priority: Critical Fix For: 1.1.8, 1.0.13 We have a sporadic problem with Sling's JCR installer 3.3.8 and Oak (tar mk). It seems to timing related: the JCR installer does a Thread#interrupt at one point and sometimes this brings the hole instance to stop. Nothing else is going on any more. While of course, a workaround is to remove the Thread.interrupt call in the JCR installer (which we did, see SLING-4477), I have the fear that this can happen with any code that is using the repository and gets interrupted. This error is hard to reproduce, however with three people testing we could see this several times happening -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (OAK-2609) Thread.interrupt seems to stop repository
[ https://issues.apache.org/jira/browse/OAK-2609?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14363343#comment-14363343 ] Carsten Ziegeler commented on OAK-2609: --- Just to be clear: I'm not encouraging the use of Thread.interrupt - in many cases it can be replaced by a better way. However, as noted, I've seen a Thread.interrupt call bringing down Oak and while I agree that replacing this with something else is a better solution, I'm not sure about how practible this is. If you own all the code in the world, you can do this. Can this at least be catched somehow with a big log message? Thread.interrupt seems to stop repository - Key: OAK-2609 URL: https://issues.apache.org/jira/browse/OAK-2609 Project: Jackrabbit Oak Issue Type: Bug Affects Versions: 1.1.7 Environment: Java 7 Reporter: Carsten Ziegeler Priority: Critical Fix For: 1.1.8, 1.0.13 We have a sporadic problem with Sling's JCR installer 3.3.8 and Oak (tar mk). It seems to timing related: the JCR installer does a Thread#interrupt at one point and sometimes this brings the hole instance to stop. Nothing else is going on any more. While of course, a workaround is to remove the Thread.interrupt call in the JCR installer (which we did, see SLING-4477), I have the fear that this can happen with any code that is using the repository and gets interrupted. This error is hard to reproduce, however with three people testing we could see this several times happening -- This message was sent by Atlassian JIRA (v6.3.4#6332)