[JIRA] (JENKINS-48933) Unstoppable Multibranch Pipeline Scan

8 views
Skip to first unread message

DTaylorBusiness@gmail.com (JIRA)

unread,
Apr 16, 2018, 12:14:02 PM4/16/18
to jenkinsc...@googlegroups.com
David Taylor commented on Bug JENKINS-48933
 
Re: Unstoppable Multibranch Pipeline Scan

Having this exact same issue. No updates? It's been a few months

Add Comment Add Comment
 
This message was sent by Atlassian JIRA (v7.3.0#73011-sha1:3c73d0e)
Atlassian logo

kevin.lannen@gmail.com (JIRA)

unread,
Apr 17, 2018, 12:57:03 PM4/17/18
to jenkinsc...@googlegroups.com
Kevin Lannen updated an issue
 
Jenkins / Bug JENKINS-48933
Unstoppable Multibranch Pipeline Scan
Change By: Kevin Lannen
Priority: Minor Major

kevin.lannen@gmail.com (JIRA)

unread,
Apr 17, 2018, 12:57:03 PM4/17/18
to jenkinsc...@googlegroups.com

We are running into this same issue, This is a major issue for us since restarting Jenkins is not a valid workaround during the workday.

kevin.lannen@gmail.com (JIRA)

unread,
Apr 17, 2018, 12:59:02 PM4/17/18
to jenkinsc...@googlegroups.com

This seems similar to this issue: https://issues.jenkins-ci.org/browse/JENKINS-32767 We are using Bitbucket for our source control.

kevin.lannen@gmail.com (JIRA)

unread,
Apr 17, 2018, 1:31:03 PM4/17/18
to jenkinsc...@googlegroups.com
Kevin Lannen updated an issue
Change By: Kevin Lannen
Component/s: git-plugin

svanoort@cloudbees.com (JIRA)

unread,
May 2, 2018, 5:48:03 PM5/2/18
to jenkinsc...@googlegroups.com

Today I learned that multibranch indexing is The Juggernaut and there is no stopping it.

rsandell Could it be that we need to include SCM timeouts or better exception handling somewhere?

kevin.lannen@gmail.com (JIRA)

unread,
May 2, 2018, 7:46:01 PM5/2/18
to jenkinsc...@googlegroups.com

There definitely needs to be some sort of timeout here, this happens less now since we have backed off our polling interval to avoid it (not ideal). We have been able to link all the times that these have occurred to connection issues with our Bitbucket server.

A connection error should not require a restart of Jenkins though

rsandell@cloudbees.com (JIRA)

unread,
May 3, 2018, 6:57:02 AM5/3/18
to jenkinsc...@googlegroups.com
rsandell commented on Bug JENKINS-48933

Did a quick dive into the depths of JGit and there are possibilities to specify connection and read timeouts, but the Jenkins git-client-plugin doesn't provide any facilities to set it from what I can see. Strange though that JGit doesn't respond to any Thread.interrupt(), from what I could see there are no explicit checks for it.

 

As a workaround have you tried switching to cli git instead of using JGit?

kevin.lannen@gmail.com (JIRA)

unread,
May 3, 2018, 10:22:04 AM5/3/18
to jenkinsc...@googlegroups.com

++We have used both JGit and the cli git and both have this issue. We would prefer to use cli git as it seems to be more reliable in other areas but it seemed to hang more during branch indexing then JGit does.

kevin.lannen@gmail.com (JIRA)

unread,
May 3, 2018, 10:22:04 AM5/3/18
to jenkinsc...@googlegroups.com
Kevin Lannen edited a comment on Bug JENKINS-48933
++ We have used both JGit and the cli git and both have this issue. We would prefer to use cli git as it seems to be more reliable in other areas but it seemed to hang more during branch indexing then JGit does.

shen3lu4@gmail.com (JIRA)

unread,
May 14, 2018, 1:40:02 PM5/14/18
to jenkinsc...@googlegroups.com
Lu Shen commented on Bug JENKINS-48933

Having the exact the same issue. Is there a way to delete these multibranch indexing without restarting Jenkins?

DTaylorBusiness@gmail.com (JIRA)

unread,
May 23, 2018, 4:21:02 PM5/23/18
to jenkinsc...@googlegroups.com

ryan_gates@ultimatesoftware.com (JIRA)

unread,
Jun 25, 2018, 3:57:02 PM6/25/18
to jenkinsc...@googlegroups.com

I too am having the same issue and would be interested in any updates.

This message was sent by Atlassian JIRA (v7.10.1#710002-sha1:6efc396)

andgoldin@gmail.com (JIRA)

unread,
Oct 5, 2018, 5:33:02 PM10/5/18
to jenkinsc...@googlegroups.com

I am also seeing this issue and would be happy to know about any updates. We have Jenkins running in a Docker container, and so far the only solution is to restart the Docker daemon on the host, as the stuck jobs cause the container to refuse to stop on its own.

This message was sent by Atlassian Jira (v7.11.2#711002-sha1:fdc329d)

dthomash@live.com (JIRA)

unread,
Jan 21, 2019, 5:07:04 PM1/21/19
to jenkinsc...@googlegroups.com
Dylan H commented on Bug JENKINS-48933

I am seeing the same issue with Perforce Stream multibranch scanning. Restarting Jenkins seems to resolve it for us, but it is certainly a less than elegant workaround.

mark.earl.waite@gmail.com (JIRA)

unread,
Jul 24, 2019, 9:06:10 AM7/24/19
to jenkinsc...@googlegroups.com
Mark Waite updated an issue
Change By: Mark Waite
Component/s: git-plugin

jgrant216@gmail.com (JIRA)

unread,
Oct 15, 2019, 6:47:03 AM10/15/19
to jenkinsc...@googlegroups.com
Jeff G commented on Bug JENKINS-48933

Same issue.  Does anyone have a way to abort these?  I tried using Scriptler as though it were a build but could not find the right combination.

This message was sent by Atlassian Jira (v7.13.6#713006-sha1:cc4451f)
Atlassian logo

bitwiseman@gmail.com (JIRA)

unread,
Jan 9, 2020, 12:00:08 PM1/9/20
to jenkinsc...@googlegroups.com

Is this specific to git based multibranch projects? Mark Waite does this looks familiar to you?

mark.earl.waite@gmail.com (JIRA)

unread,
Jan 9, 2020, 2:11:03 PM1/9/20
to jenkinsc...@googlegroups.com

Liam Newman the stack trace in the bug report is specific to git based multibranch projects. It is using the JGit implementation to retrieve the remote symbolic references.

This report is based on git client plugin 2.x and git plugin 3.x (based on when it was submitted). One or more of the submitters might want to try git client plugin 3.0.0 and git plugin 4.0.0. They switched to JGit 5.5.0 instead of the JGit 4.x version that was included in git client plugin 2.x and git plugin 3.x.

They might also want to try with git client plugin 3.1.0-beta and git plugin 4.1.0-beta so that they could test JGit 5.6.0 instead of JGit 4.x.

Reply all
Reply to author
Forward
0 new messages