I'm running a couple of iSCSI Targets with OCFS2. I had to reboot my iSCSI Target (the physical iSCSI Target) due to some hardware changes and it caused all the initiators to reboot. I want to prevent that or at least make them wait a lot longer before they reboot.

Is there some sysctl variables I can use to modify the targets?

I have multiple iSCSI Targets on my main system, due to need to share them out with multiple systems I'm running OCFS2 as the Filesystems on the Initiators

  • I am not sure I get your terminology. You are running a few servers with iSCSI targets and I presume they get rebooted as well when you reboot a server working as iSCSI initiator? Or do the initiators reboot when you reboot the server running the targets?
    – Sven
    Sep 11 '11 at 0:28
  • Just for clarity - An iSCSI Target is on the iSCSI 'server' that serves up disk. The iSCSI 'Initiators' are on the 'clients' that make use of the disk that's being provided. With this in mind, you need to clarify your question as Sven said. It's possible that your terminology is correct and you're running multiple targets on one server, and didn't really mean 'reboot'. Sep 11 '11 at 0:39
  • Did the initiators write anything into their logs before rebooting? Maybe the answer is in there. Sep 11 '11 at 5:03
  • Just a log message saying rebooting. I was also watching the initiators when I rebooted my Target. Pretty much as soon as I shutdown my Target my initiators rebooted.
    – Squidly
    Sep 11 '11 at 14:10

Are the initiators at question using the iSCSI storage as a system/boot device? If so, the only option that comes to mind would be to suspend or hibernate the initiators prior to rebooting the target.

  • No the initiators are just using it for storage.
    – Squidly
    Sep 11 '11 at 4:09
  • If you can't isolate the source of the reboot via logs, it would be interesting to set up an otherwise similarly configured initiator that has no OCFS2 targets, but does have a generic ext3 (or whatever) iSCSI target mounted, then reboot the iSCSI server. If the initiator doesn't reboot, you can suspect that the issue is specific to OCFS2; if it does reboot, you can suspect that it's fundamentally an iSCSI issue.
    – Skyhawk
    Sep 11 '11 at 16:17
  • I'll do that, I have a file system I can export via iSCSI that should work as a good test.
    – Squidly
    Sep 11 '11 at 17:43
  • Any revelations?
    – Skyhawk
    Sep 26 '11 at 20:20
  • Not yet. I have not had time to test it. Work's been very busy..
    – Squidly
    Sep 26 '11 at 22:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.