iSCSId stopping on rack controllers after a while
jim at tilander.org
Thu Feb 9 22:47:27 UTC 2017
> On Feb 9, 2017, at 1:23 PM, Andres Rodriguez <andres.rodriguez at canonical.com> wrote:
> Hi Jim,
> On Thu, Feb 9, 2017 at 3:34 PM, Jim Tilander <jim at tilander.org <mailto:jim at tilander.org>> wrote:
> I’ve got a separate region / rack controller setup and I’ve noticed that after a while, the rack controllers seem to fail to serve iSCSI, causing the boot agents to hang during their boot up (see attached screenshot). Has anyone seen this? Any potential workarounds / fixes?
> I'd really need a bit more detail, but I've not particularly seen an issue where tgtd would just stop on the rack controllers. A few questions though:
> 1. How often does this happen ? (multiple times a day, once a day, etc?)
It only happened once so far it seems.
> 2. Is MAAS updating its images while machines are booting that could cause this? (for e.g. rack controllers syncing new images while machines are booting).
No, this is during a steady state — nothing really was happening in the system (don’t have many images nor nodes yet).
> 3. Check the UI (images page, or rack controller details) to see if there are any errors (i.e. rack controller disconnected, tgt not running, error while downloading images, etc).
It all looked green.
> 4. Can you provide logs? (/var/log/maas/rackd.log and /var/log/maas/regiond.log on the region).
I can dig up some logs and attach to a bug.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Maas-devel