Server disk space/memory problem causes scm-manager to crash - questions about work dirs

after we rebooted our linux server, scm-manager appeared to be down. In the logs it mentioned cleaning up work dirs and the disk space usage decreased substantially. Can someone educate me on these work docs and how to keep them from filling up my disk again?

Hey Marlene,

do you mean the working directories? Their are used for internal operations for repositories in SCM-Manager. Normally they will be cleaned up automatically. If you have restarted the server, there might be some corrupt files left? Do you have a stacktrace or specific errors?

Here some more informations on this topic: https://scm-manager.org/docs/2.40.x/en/administration/workdir_caching/

Not sure. this is what I saw in the log.

] INFO sonia.scm.repository.work.WorkdirProvider - deleting 0 old work dirs in /var/lib/scm/repositories/BMTCo4wUGSM/work

2023-01-26 13:07:16.186 [main] INFO sonia.scm.repository.work.WorkdirProvider - deleting 0 old work dirs in /var/lib/scm/repositories/BgTRuELJUTc/work

2023-01-26 13:07:16.186 [main] INFO sonia.scm.repository.work.WorkdirProvider - deleting 0 old work dirs in /var/lib/scm/repositories/4kTFMDYmFK1/work

2023-01-26 13:07:16.186 [main] INFO sonia.scm.repository.work.WorkdirProvider - deleting 0 old work dirs in /var/lib/scm/repositories/8iTFMCz1mJX/work

2023-01-26 13:07:16.186 [main] INFO sonia.scm.repository.work.WorkdirProvider - deleting 0 old work dirs in /var/lib/scm/repositories/BITTUwEcynX/work

2023-01-26 13:07:16.186 [main] INFO sonia.scm.repository.work.WorkdirProvider - deleting 0 old work dirs in /var/lib/scm/repositories/4iTFMDIiyJm/work

2023-01-26 13:07:16.187 [main] INFO sonia.scm.repository.work.WorkdirProvider - deleting 2 old work dirs in /var/lib/scm/repositories/5wSpUzQVUZ/work

And the disk space usage went from almost 1TB down to 90GB.

This is the message that was on the vm’s console before we rebooted.

~WRD0348.jpg

image002.png

image003.png

image004.png

image005.png

image006.png

Hey @mcote

workdirs will be deleted with each restart.
Question regarding your reported issue: is it true that you have one/two large repository? With almost one TB allocated space?

No, we have many repositories, but one disk that is dynamic and currently has 1TB of allocated space.