[ClusterLabs] Antw: Re: GFS2 problem after host name change

Ulrich Windl Ulrich.Windl at rz.uni-regensburg.de
Mon Jan 15 04:09:07 EST 2018


I can't tell you what was wrong, but I could tell you what not to do in
production environments:

* Change host names

It's not that it cannot be done, but there are too many places where a
hostname may be configured, and it's very likely to miss one of those.

Our strategy had always been: New configurations on new hardware.
(Well, virtual machines add an interesting new aspect to that)


>>> Digimer <lists at alteeve.ca> schrieb am 15.01.2018 um 03:48 in Nachricht
<3b063678-7e7b-0bab-19f1-5f19ca1f2bdd at alteeve.ca>:
> On 2018-01-14 12:48 PM, Digimer wrote:
>> On 2018-01-14 12:29 PM, Digimer wrote:
>>> I recently changed the host name of a cluster. It may or may not be
>>> related, but after I noticed that I can cleanly start gfs2 when the node
>>> boots. However, if the node is withdrawn and then I try to rejoin it
>>> without a reboot, it hangs with this in syslog;
> A bit more info...
> I tried reformatting the GFS2 partition, and the problem remained.
> Then I tried zero'ing out the LV, and it was writing at about 4 KB/sec
> (dd if=/dev/zero of=/dev/vg/lv bs=1M oflag=dsync, run for about two
> minutes).
> When I saw how slow it was going, I decided to delete the LV (original
> was 40 GiB) and created a new LV (20 GiB). Then I formatted the new LV
> GFS2 and the problem remained.
> After this, I could no longer mount the GFS2 partition at all. Even on a
> fresh boot of both nodes, it immediately hung.
> At this point, I was losing my maintenance window, so I migrated the
> hosted servers to other Anvil! systems and the rebuilt the cluster.
> After the rebuild, the GFS2 partition mounted properly.
> A frightening experience, to be sure. Thankfully I had enough overhead
> on other systems to avoid an interruption. That said, I would love some
> input on why this might have happened.
> cheers,
> digimer
> -- 
> Digimer
> Papers and Projects: https://alteeve.com/w/ 
> "I am, somehow, less interested in the weight and convolutions of
> Einstein’s brain than in the near certainty that people of equal talent
> have lived and died in cotton fields and sweatshops." - Stephen Jay Gould
> _______________________________________________
> Users mailing list: Users at clusterlabs.org 
> http://lists.clusterlabs.org/mailman/listinfo/users 
> Project Home: http://www.clusterlabs.org 
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf 
> Bugs: http://bugs.clusterlabs.org

More information about the Users mailing list