[JIRA] (JENKINS-49736) Remote ssh slave disconnection
Title: Message Title Ivan Fernandez Calvo assigned an issue to Ivan Fernandez Calvo Jenkins / JENKINS-49736 Remote ssh slave disconnection Change By: Ivan Fernandez Calvo Assignee: Ivan Fernandez Calvo Add Comment This message was sent by Atlassian JIRA (v7.3.0#73011-sha1:3c73d0e) -- You received this message because you are subscribed to the Google Groups "Jenkins Issues" group. To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-issues+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
[JIRA] (JENKINS-49736) Remote ssh slave disconnection
Title: Message Title Ivan Fernandez Calvo closed an issue as Not A Defect Trilead library it is really sensitive to network performance/issues Jenkins / JENKINS-49736 Remote ssh slave disconnection Change By: Ivan Fernandez Calvo Status: Open Closed Resolution: Not A Defect Add Comment This message was sent by Atlassian JIRA (v7.3.0#73011-sha1:3c73d0e) -- You received this message because you are subscribed to the Google Groups "Jenkins Issues" group. To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-issues+unsubscr...@googlegroups.com. For more options, visit h
[JIRA] (JENKINS-49736) Remote ssh slave disconnection
Title: Message Title Chris Amis commented on JENKINS-49736 Re: Remote ssh slave disconnection I said earlier that the 1G connection made things better, after some time now with not another occurence I think it fixed it. I did not just change the switch, we stripped the whole lot down and rebuilt it in a new home, we check all cables were nice and snug and so on. My guess is that we must have had a bad connection and the jenkins SSH connection was really susceptible (remember that the tests would run the full 60 hours if I ran them from a cmd prompt on the jenkins machine). Chris Add Comment This message was sent by Atlassian JIRA (v7.3.0#73011-sha1:3c73d0e) -- You received this message because you are subscribed to the Google Groups "Jenkins Issues" group. To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-issues+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
[JIRA] (JENKINS-49736) Remote ssh slave disconnection
Title: Message Title Francois Aichelbaum commented on JENKINS-49736 Re: Remote ssh slave disconnection Hi Ivan Fernandez Calvo On our side, we have the very exact same issues since beginning of January and kernel patches for Spectre/Meltdown. Sysctl on Jenkins (core) and the various slaves) have been tweaked multiple times without luck. Your figures are not the best for our setup and the number of jobs we run, but though, did not help either when we used such (those values are very conservative but are not of the best use in production environments with high load. Cheers Add Comment This message was sent by Atlassian JIRA (v7.3.0#73011-sha1:3c73d0e) -- You received this message because you are subscribed to the Google Groups "Jenkins Issues" group. To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-issues+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
[JIRA] (JENKINS-49736) Remote ssh slave disconnection
Title: Message Title Ivan Fernandez Calvo commented on JENKINS-49736 Re: Remote ssh slave disconnection Did you try to tune the keepalive int the TCP stack? the default values are terrible sysctl -w net.ipv4.tcp_keepalive_time=120 sysctl -w net.ipv4.tcp_keepalive_intvl=30 sysctl -w net.ipv4.tcp_keepalive_probes=8 sysctl -w net.ipv4.tcp_fin_timeout=30 Add Comment This message was sent by Atlassian JIRA (v7.3.0#73011-sha1:3c73d0e) -- You received this message because you are subscribed to the Google Groups "Jenkins Issues" group. To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-issues+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
[JIRA] (JENKINS-49736) Remote ssh slave disconnection
Title: Message Title Oleg Nenashev assigned an issue to Unassigned Bulk issue update: The plugin connectivity is still unstable from what I see in this and other reports. Probably the recent patches in 1.24-1.25 caused some extra instability by getting rid of interlocks between agent connection and termination logic. Apparently it impacts some reconnection scenarios due to the race conditions. Unfortunately I do not have capacity to work on the plugin in medium-term. So for now I am unassigning issues from myself. Ivan Fernandez Calvo was very kind to take ownership of the plugin and to handle some workload in it. Probably he will have some capacity to review the backlog I was unable to triage. Jenkins / JENKINS-49736 Remote ssh slave disconnection Change By: Oleg Nenashev Assignee: Oleg Nenashev Add Comment This message was sent by Atlassian JIRA (v7.3.0#73011-sha1:3c73d0e)
[JIRA] (JENKINS-49736) Remote ssh slave disconnection
Title: Message Title Chris Amis commented on JENKINS-49736 Re: Remote ssh slave disconnection We have managed to very much improved matters by putting all the machines on a single 1G switch. Before this we had Jenkins-1G-100M-Slave which does not sound too bad. The 1G switch is the same device, the 100M is a big Cisco job in a rack. We have plans to remote the Jenkins server into the cloud, should I say we cannot risk it? Add Comment This message was sent by Atlassian JIRA (v7.3.0#73011-sha1:3c73d0e) -- You received this message because you are subscribed to the Google Groups "Jenkins Issues" group. To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-issues+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
[JIRA] (JENKINS-49736) Remote ssh slave disconnection
Title: Message Title Chris Amis commented on JENKINS-49736 Re: Remote ssh slave disconnection I updated the problem box to the latest LTS release, same as the working box. Much improved, but still fails 50% of the time. I suspect one of the corporate forced updates that happened on the day it went wrong is the culprit. http://support.microsoft.com/?kbid=4056894 http://support.microsoft.com/?kbid=4056568 http://support.microsoft.com/?kbid=4054176 http://support.microsoft.com/?kbid=4054998 Somehow something has made the jenkins ssh comms more susceptible to something. Add Comment This message was sent by Atlassian JIRA (v7.3.0#73011-sha1:3c73d0e) -- You received this message because you are subscribed to the Google Groups "Jenkins Issues" group. To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-issues+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
[JIRA] (JENKINS-49736) Remote ssh slave disconnection
Title: Message Title Chris Amis created an issue Jenkins / JENKINS-49736 Remote ssh slave disconnection Issue Type: Bug Assignee: Oleg Nenashev Components: ssh-slaves-plugin Created: 2018-02-26 10:25 Environment: Jenkins 2.89.4 on Windows 7. ssh slaves, 14.04 LTS and 16.04 LTS. Labels: slave Priority: Major Reporter: Chris Amis This is a bit of an epic, sorry. For a year or so we have had a system configured that after build triggers an ssh slave (14.04 LTS) in a test chamber to run hours of tests, an hour after a build, 12 hours overnight and 60 hours over the weekend. All stable and happy. For the last couple of months we have been recreating the setup for a new product, pretty much a copy/paste of the existing system except this is on 16.04 LTS. Here is the twist, adding the new system seems to have broken the old. I have about logs of about 600 good jobs up to 16th Feb, after that about 2 in 50 have worked. I cannot even work out which end is failing. Over the last weekend I disconnected Jenkins from the slaves (I changed the addresses not just offlined). On the jenkins server I ran 2 command windows, in the windows I ran ssh connections to the slaves and executed the commands manually to run 60 hour tests on both. Both stayed up for 60 hours and ran to conclusion. Tried automated again this morning, broken in minutes. I spotted this morning a new version of the ssh plugin, so I upgraded from 1.25.1 to 1.26, same results. Any ideas... The jenkins log looks like this