Scheduling a slave node reboot

[Chris Packham]

Hi,

We have a problem with our build nodes and while we know the cause we haven't found a fix.

One workaround we're looking at is to periodically reboot the node. Is there a safe way to reboot a node i.e. prevent new jobs being assigned, reboot and rejoin the master.

[Mark Waite]

One technique is to reduce the number of executors on the agent to 1, then schedule a Jenkins job to perform the reboot.  That assures the job performing the reboot is the only job on that agent.

Mark Waite

--
You received this message because you are subscribed to the Google Groups "Jenkins Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-use...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/jenkinsci-users/CAFOYHZB1GahEozJ_DKPuVit26YpqbEsaUXrK9i%3DN%2By3Fz_tUDg%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

[Simon Richter]

Hi,

On 23.09.2017 01:40, Chris Packham wrote:

> One workaround we're looking at is to periodically reboot the node. Is
> there a safe way to reboot a node i.e. prevent new jobs being assigned,
> reboot and rejoin the master.

Evil hack: use the heavyweight job plugin to make sure all executors are
used by the job that issues the reboot command.

Simon