Details
-
New Feature
-
Status: Resolved (View Workflow)
-
Major
-
Resolution: Fixed
Description
In Jenkins projects we have many users complaining that the slave/agent is "already connected", because they have a runaway slave/agent process. It happens when WinSW gets terminated without executing the process shutdown logic (force kill) or when WinSW fails to terminate the process.
As a part of WinSW 2.0, it would be great to create a logic, which...
- records PID of the created process to the disc
- performs status check of the previously spawned process upon restart
- terminates the runaway process if required
It can be done via WinSW 2 "plugin"
Issue: https://github.com/kohsuke/winsw/issues/125
Attachments
Issue Links
- is duplicated by
-
JENKINS-29825 some windows agents unable to reconnect after master restart
-
- Resolved
-
- is related to
-
-
- Open
-
-
-
- Resolved
-
-
-
- Resolved
-
-
-
- Resolved
-
- links to
Activity
krogan Smells like another issue.
Anyway, makes sense to retest after the WinSW 2 integration. You can install the new Windows service wrapper manually
Code changed in jenkins
User: Oleg Nenashev
Path:
src/main/resources/org/jenkinsci/modules/windows_slave_installer/jenkins-slave.xml
http://jenkins-ci.org/commit/windows-slave-installer-module/8dcf02da16e7c95c67c7de95fd078089a8ecf8df
Log:
JENKINS-39231 - Integrate the Runaway Process Killer extension to terminate runaway processes
Code changed in jenkins
User: Oleg Nenashev
Path:
core/pom.xml
core/src/main/resources/windows-service/jenkins-slave.xml
core/src/main/resources/windows-service/jenkins.xml
war/pom.xml
http://jenkins-ci.org/commit/jenkins/e698d1de41d4311bf5f8b1d2c40b591109e696e2
Log:
Update Windows Agent Installer to 1.7 and WinSW to 2.0.2 (#2765)
-
-
- WinSW changes
-
The update includes many fixes and improvements, the full list is provided in the [WinSW changelog](https://github.com/kohsuke/winsw/blob/master/CHANGELOG.md). There are several issues referenced in Jenkins bugtracker:
JENKINS-22692(https://issues.jenkins-ci.org/browse/JENKINS-22692) - Connection reset issues when WinSW gets terminated due to the system shutdownJENKINS-23487(https://issues.jenkins-ci.org/browse/JENKINS-23487)- Support of shared directories in WinSWJENKINS-39231(https://issues.jenkins-ci.org/browse/JENKINS-39231) - Enable Runaway Process Killer by defaultJENKINS-39237(https://issues.jenkins-ci.org/browse/JENKINS-39237) - Auto-upgrade of JNLP agent versions on the slaves
-
-
- Windows Agent Installer changes
-
- Adapt the default configurations to pick fixes above
- Slave => Agent renaming where possible
-
-
- Jenkins core changes
-
- Modify the configuration template, reference advanced options
- Enable Runaway Process Killer by default
- Update Windows Agent Installer to 1.7
- Remove the obsolete jenkins-slave.xml file from the core.
Now it is within windows-slave-installer
- Use the deployed Snapshot for CI
- Pick the release version of windows-slave-installer-1.7
Jenkins master is on 2.32.1
Master and slaves running Win2012
The symptoms sound very familiar to a problem where we've had a jenkins slave up.... then we reboot the windows server (slave).
When the server returns and the slave is automatically started, it hangs around for about 30secs then terminates connection which kills our job.
We've also witnessed the hosting windows service winsw 1.17 (which auto upgrades to 1.18) bombs out but leaves the java process running.
The java process is still keeping the slave active to the master for an indiscriminate amount of time (anywhere between 20secs to 2hrs) before eventually dying of its own accord, with no fresh jobs sent or interaction with the windows service.