Multibranch Pipeline job creating multiple Checkout Directories in Workspace

[Ven H]

Hi,

I have a Multibranch Pipeline job setup for SVN. In Trunk, I have a directory called Platform which has a Jenkins file. The workspace structure is something like Drive\Workspace_Dir\trunk%2FPlatform.

After a few builds, I can see some strange things happening with this.

The Workspace_Dir has these many directories now.

trunk%2FPlatform

trunk%2FPlatform@tmp

trunk%2FPlatform@2

trunk%2FPlatform@2@tmp

trunk%2FPlatform@3

trunk%2FPlatform@3@tmp

trunk%2FPlatform@4

trunk%2FPlatform@4@tmp

and so on...

Although, this doesn't always happen, it happens every now and then. I am not sure of the reason and would like to fix this. Can anyone please help me with this? Because of this, the drive itself runs out of space sometimes, since the code base is very large.

 

Regards,
Venkatesh

[Björn Pedersen]

This standard behaviour if jobs run in parallel. The first run uses:

trunk%2FPlatform => the workdir for the job

trunk%2FPlatform@tmp => the checkout for the Jenkinsfile to set up the job.

the @n dirs are created for further parallel runs. 

Björn

[Ven H]

Thanks a lot. Can you please advise on how to stop / control this behavior?

Regards,
Venkatesh

--
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/5219e400-ad91-48c8-ac50-dd9f81767c3bn%40googlegroups.com.

[Anton Shepelev]

Bjorn Pedersen to Ven H:

> > I have a Multibranch Pipeline job setup for SVN. In
> > Trunk, I have a directory called Platform which has a
> > Jenkins file. The workspace structure is something like
> > Drive\Workspace_Dir\trunk%2FPlatform.
> >
> > After a few builds, I can see some strange things
> > happening with this. The Workspace_Dir has these many
> > directories now.
> >
> > trunk%2FPlatform
> > trunk%2FPlatform@tmp
> > trunk%2FPlatform@2
> > trunk%2FPlatform@2@tmp
> > trunk%2FPlatform@3
> > trunk%2FPlatform@3@tmp
> > trunk%2FPlatform@4
> > trunk%2FPlatform@4@tmp
> > and so on...
>

> This standard behaviour if jobs run in parallel.

Does the official documentation tell about it, or is it
considred an internal part of the implementation?

Ven H:

> Can you please advise on how to stop / control this
> behavior?

Try limiting the number of executors to unity.