1

I looked at https://unix.stackexchange.com/search?q=jenkins & Jenkins support https://issues.jenkins.io

I didn't find out there how to solve that.

initial state

Nodes & main server had disk space issues, and no pipeline were triggered anymore.

After space issue has been solved nodes are still offline ignoring the new status.

how can I refresh those nodes to online state?

Architecture :

  • host connections uses java / git & ssh depending on with step/pipeline is running
  • pipelines usually uses git to get last groovy & script files ; & also run the jenkins groovy file on nodes
  • steps are calling ssh on either bash or powershell access when needed to access nodes & run scripts.

Done :

  • cleaned /var/lib on main jenkins server

    find . -name "xxxx" -atime +365 -exec rm {} \;

  • cleaned application drives on each nodes

    $limit = (Get-Date).AddDays(-30)
    $path = "xxxxxxx"
    Get-ChildItem -Path $path -Recurse -Force -filter XXXXX | Where-Object { !$_.PSIsContainer -and $_.CreationTime -lt $limit } | Remove-Item -Force
  • restarted jenkins service on main jenkins server

    sudo systemctl restart jenkins

Current Status :

  • Those nodes remains offline (and still display as error the same space amount warning, this is like it doesn't read/detect the new available space)
    Disk space is too low. Only 0.000GB left on F:\jenkins2.
    
    Agent is connected.

on machines for example :

DriveType    : 3
ProviderName :
FreeSpace    : 255938342912
Size         : 449998483456
VolumeName   : Disk2

so jenkins ignore issue about disk space is solved & keeps nods offline how to make them online again ?

13
  • 2
    Question is hard to understand: Edit question to show, What did you do?, What have you tried?, What went wrong? (show don't describe). First link is just search results (not specific enough to be of use), 2nd link is broken. Commented Feb 3, 2021 at 10:23
  • 1
    Edit your question and include how the nodes connect to the master (SSH? Java?)
    – Panki
    Commented Feb 3, 2021 at 10:31
  • @ctrl-alt-delor I don't want to show (external client defense environnement)
    – francois P
    Commented Feb 3, 2021 at 11:03
  • Thanks for including more info. Now proof read. What can you do to make it shorter and clearer? But do include the command that you typed (obfuscating any private names). Commented Feb 3, 2021 at 14:08
  • 2
    Are you using a Windows environment in any part of your Jenkins usage scenario?
    – AdminBee
    Commented Feb 9, 2021 at 14:02

1 Answer 1

2
+150

Based on your comments and question edits. It appears the space issue isn't with your host RHEL system.

The problem is with your "Windows Images" (I presume you mean virtual machines). The space issue is clearly talking about a Windows path F:\jenkins2 containing a Windows drive letter F: and path backslash path seperator \. If it were a linux path there would be no : in the path and it would use a forward slash / as a path separator.

You've not given an information on what type of VM these are beyond being a "Minimal Windows image", so I can't offer any advice about how to clear down the existing images. To do that you'd probably need to get a command line running inside the image and manually delete files.

What you can do is...

  • Confirm that your existing builds are still accessible (and not only stored in the build nodes). To do this, just log into Jenkins and check you can still download build artefacts from Windows builds.
  • Destroy your build images and recreate them. Depending on your VM software it might be possible to revert them to an earlier version.

I'd also recommend that for future you check your Jenkins options for disk space management. I suspect that your build nodes have filled up with old builds which they do not need to keep.

2
  • Thanks to all of you, especially to Philip. So yes both windows AND linux had space issues. 2cd point, builds are accessible. 3rd point, destroying (after snapshot), restoring & getting back differential backup datas permit to solve the whole solution jenkins now detects the change of space & ran the stuck jobs one remains running but I really think it will be OK now. (BTW I am not a windows guy, so I don't understand how the "scale images for windows were build, they just are hosted on that seems to be a modified "scale" hypervisor.)
    – francois P
    Commented Feb 11, 2021 at 15:10
  • of course I also took note about disk space management :)
    – francois P
    Commented Feb 11, 2021 at 15:12

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .