Interesting this is the first time I’ve seen this bug posted. I’m still having problems with bricks going down in my hci setup. It had been 1-2 bricks dropping 2-4 times per day on different volumes.  However when all bricks are up everything is working ok and all bricks are healed and seemingly in sync. Could this stil be related to the mentioned bug in such a case?

On Fri, Mar 15, 2019 at 11:37 PM Darrell Budic <budic@onholyground.com> wrote:
Upgrading gluster from version 3.12 or 4.1 (included in ovirt 3.x) to 5.3 (in ovirt 4.3) seems to cause this due to a bug in the gluster upgrade process. It’s an unfortunate side effect fo us upgrading ovirt hyper-converged systems. Installing new should be fine, but I’d wait for gluster to get https://bugzilla.redhat.com/show_bug.cgi?id=1684385 included in the version ovirt installs before installing a hyper converged cluster. 

I just upgraded my 4.2.8 cluster to 4.3.1, leaving my separate gluster 3.12.15 servers along, and it worked fine. Except for a different bug screwing up HA engine permissions on launch, but it looks like that’s getting fixed on a different bug.

Sandro, it’s unfortunate I can’t take more part in testing days, but the haven’t been happening at times where I can participate, and a one test test isn’t really something i can participate in often. I sometimes try and keep up with the RCs on my test cluster, but major version changes wait until I get time to consider it, unfortunately. I’m also a little surprised that a major upstream issue like that bug hasn’t caused you to issue more warnings, it’s something that is going to affect everyone who’s upgrading a converged system. Any discussion on why more news wasn’t released about it?

  -Darrell


On Mar 15, 2019, at 11:50 AM, Jayme <jaymef@gmail.com> wrote:

That is essentially the behaviour that I've seen.  I wonder if perhaps it could be related to the increased heal activity that occurs on the volumes during reboots of nodes after updating.

On Fri, Mar 15, 2019 at 12:43 PM Ron Jerome <ronjero@gmail.com> wrote:
Just FYI, I have observed similar issues where a volume becomes unstable for a period of time after the upgrade, but then seems to settle down after a while.  I've only witnessed this in the 4.3.x versions.  I suspect it's more of a Gluster issue than oVirt, but troubling none the less.  

On Fri, 15 Mar 2019 at 09:37, Jayme <jaymef@gmail.com> wrote:
Yes that is correct.  I don't know if the upgrade to 4.3.1 itself caused issues or simply related somehow to rebooting all hosts again to apply node updates started causing brick issues for me again. I started having similar brick issues after upgrading to 4.3 originally that seemed to have stabilized, prior to 4.3 I never had a single glusterFS issue or brick offline on 4.2

On Fri, Mar 15, 2019 at 9:48 AM Sandro Bonazzola <sbonazzo@redhat.com> wrote:


Il giorno ven 15 mar 2019 alle ore 13:38 Jayme <jaymef@gmail.com> ha scritto:
I along with others had GlusterFS issues after 4.3 upgrades, the failed to dispatch handler issue with bricks going down intermittently.  After some time it seemed to have corrected itself (at least in my enviornment) and I hadn't had any brick problems in a while.  I upgraded my three node HCI cluster to 4.3.1 yesterday and again I'm running in to brick issues.  They will all be up running fine then all of a sudden a brick will randomly drop and I have to force start the volume to get it back up. 

Just to clarify, you already where on oVirt 4.3.0 + Glusterfs 5.3-1 and upgraded to oVirt 4.3.1 + Glusterfs 5.3-2 right?


 

Have any of these Gluster issues been addressed in 4.3.2 or any other releases/patches that may be available to help the problem at this time?

Thanks!
_______________________________________________
Users mailing list -- users@ovirt.org
To unsubscribe send an email to users-leave@ovirt.org
Privacy Statement: https://www.ovirt.org/site/privacy-policy/
oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/
List Archives: https://lists.ovirt.org/archives/list/users@ovirt.org/message/746CU33TP223CFYS6BFUA2C4FIYZQMGU/


--
SANDRO BONAZZOLA

MANAGER, SOFTWARE ENGINEERING, EMEA R&D RHV

sbonazzo@redhat.com   

_______________________________________________
Users mailing list -- users@ovirt.org
To unsubscribe send an email to users-leave@ovirt.org
Privacy Statement: https://www.ovirt.org/site/privacy-policy/
oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/
List Archives: https://lists.ovirt.org/archives/list/users@ovirt.org/message/RXHP4R5OXAJQ3SOUEKXYGOKTU43LZV3M/
_______________________________________________
Users mailing list -- users@ovirt.org
To unsubscribe send an email to users-leave@ovirt.org
Privacy Statement: https://www.ovirt.org/site/privacy-policy/
oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/