Backup failed, high IO delay

Discussion in 'Proxmox VE: Installation and configuration' started by spearox, Feb 12, 2019.

  1. spearox

    spearox Member

    Joined:
    Jun 6, 2013
    Messages:
    81
    Likes Received:
    1
    Hi all,

    I got a problem.
    I have an NFS mount on the server where backup goes.
    Once the backup failed and the vzdump lock file not deleted and the vzdump still alive.
    I tryed to stop the task but not success, then I killed the vzdump process, now its not running but the task still show its running and the IO delay on server fix 25% after the process failed.

    How can I kill, stop without server restart? The NFS mount is live, I think the problem of process fail is the nfs mount gone once.
     
  2. Chris

    Chris Proxmox Staff Member
    Staff Member

    Joined:
    Jan 2, 2019
    Messages:
    152
    Likes Received:
    12
    Hi,
    as a first step you might want to check out what instances of vzdump are running with `ps aux | grep vzdump`.
     
    Stop hovering to collapse... Click to collapse... Hover to expand... Click to expand...
  3. spearox

    spearox Member

    Joined:
    Jun 6, 2013
    Messages:
    81
    Likes Received:
    1
    I killed the processes already as i mentioned above.
    The output:
    Code:
    root@node1:~# ps aux | grep vzdump
    root      6069  0.0  0.1 522688 86812 ?        Ds   Feb07   0:14 task UPID:node1:000017B5:1D2B434D:5C5B66F1:vzdump::root@pam:
    root     21004  0.0  0.0  12788   936 pts/0    S+   08:23   0:00 grep vzdump
     
  4. Chris

    Chris Proxmox Staff Member
    Staff Member

    Joined:
    Jan 2, 2019
    Messages:
    152
    Likes Received:
    12
    The status of the task is "Ds" so it is waiting for IO, causing your high IO delays. You can try to stop the task in the WebUI, this should send a SIGTERM to the task and clean up afterwards. If this does not work, you might want to try to send a SIGTERM to the task directly.
     
    Stop hovering to collapse... Click to collapse... Hover to expand... Click to expand...
  5. spearox

    spearox Member

    Joined:
    Jun 6, 2013
    Messages:
    81
    Likes Received:
    1
    Its not work. :/
    Code:
    root@node1:~# ps -ef|grep vz
    root      6069     1  0 Feb07 ?        00:00:14 task UPID:node1:000017B5:1D2B434D:5C5B66F1:vzdump::root@pam:
    root     23083 19501  0 07:18 pts/0    00:00:00 grep vz
    root@node1:~# kill -15 6069
    root@node1:~# ps -ef|grep vz
    root      6069     1  0 Feb07 ?        00:00:14 task UPID:node1:000017B5:1D2B434D:5C5B66F1:vzdump::root@pam:
    root     23532 19501  0 07:18 pts/0    00:00:00 grep vz
    root@node1:~#
     
  6. Chris

    Chris Proxmox Staff Member
    Staff Member

    Joined:
    Jan 2, 2019
    Messages:
    152
    Likes Received:
    12
    Then you might be more drastic and send a SIGKILL
     
    Stop hovering to collapse... Click to collapse... Hover to expand... Click to expand...
  7. spearox

    spearox Member

    Joined:
    Jun 6, 2013
    Messages:
    81
    Likes Received:
    1
    Still the same.
     
  8. Chris

    Chris Proxmox Staff Member
    Staff Member

    Joined:
    Jan 2, 2019
    Messages:
    152
    Likes Received:
    12
    Then unfortunately you will have to reboot the machine to get rid of the process as it is in a uninterruptible sleep.
     
    Stop hovering to collapse... Click to collapse... Hover to expand... Click to expand...
  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice