1

Recently when I tried to clone a VM on a ESX4 host, I encountered an error message saying that the ESX failed to create journal file. After some research I figured out it's caused by the fact that the /var/log/ filesystem is running out of space on the host.

It seems that we are having a huge log file vmkiscsid.log which occupies almost 2GB space.

vmware does have a KB about this log file

It seems the file will be re-created after a reboot. But I wonder if I can do it without a reboot. Can I safely remove the file instead?

4
  • IIRC the ESX core is based on Linux. In which case: 1) Yes, you can delete it, 2) But it will NOT free up any space since the deleted file is still in use and can even be written to. To delete it close the program writing to it. If that is the hypervisor core then you probably need to restart that. Which means shutting down all VMs and restaring the core. At this point you might as well reboot.
    – Hennes
    Jul 5, 2016 at 11:18
  • thanks Hennes for your comment. do you know if i can just zero out this log file? i am really trying to avoid the reboot .... if possible Jul 6, 2016 at 1:54
  • Zero out as in trim it? Maybe. I read this as the same problem which has been asked before with webservers and large log files. Only from webservers I know they usually support a SIGHUP causing them to reread config files and reopen log file (and closing the old one first, which is what we are after). With esxi I do not know. A comment on this answer on serverfault (serverfault.com/questions/86092/…) seem to indicate that ` truncate -s0 logfile` might work.
    – Hennes
    Jul 6, 2016 at 8:04
  • turned out i chose to use the zero out approach, and it did the trick. no reboot was needed. thx for all the comments! Aug 22, 2016 at 7:30

1 Answer 1

0

turned out i chose to use the zero out approach, and it did the trick. no reboot was needed. thx for all the comments!

You must log in to answer this question.

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