[Gluster-users] Unable to upgrade nodes because of cksums mismatch

[Michael Böhm]

Am Mo., 27. Dez. 2021 um 16:25 Uhr schrieb Nikhil Ladha <nladha at redhat.com>:

>
> On Mon, Dec 27, 2021 at 7:24 PM Michael Böhm <dudleyperkins at gmail.com>
> wrote:
>
>> Am Mo., 27. Dez. 2021 um 14:04 Uhr schrieb Nikhil Ladha <
>> nladha at redhat.com>:
>>
>>> Hi Michael
>>>
>>> I think you are hitting a similar issue like this one
>>> https://github.com/gluster/glusterfs/issues/3066.
>>> If so, the fix for the same is under review and could be available in
>>> the next release.
>>>
>>
>> Thanks for your reply. I saw that issue and thought it doesn't really
>> apply to me cause i don't use striping (only Replicate and one
>> Distributed-Replicate). But maybe, if this is a cause for the mismatched
>> cksum, it could be the root cause
>>
>
> Yes, the missing stripe count from the volfile after the upgrade is
> causing the checksum mismatch and hence the "peer-rejected" issue.
>

Aah i see - i subscribed to the issue to keep track when this is
potentially resolved.


>
>
>> I always hesitate to upgrade to a x.0 version - but as those holidays are
>> the best time for these upgrades i thought i might skip v9 this time.
>>
>>
To close this for me, as i have a lot of updates to do and this was one of
the prequisites, i downgraded the host to v9.4 and this didn't cause any
errors. I will now upgrade all other nodes to 9.4.
 As the release schedule changed with v10 and the world we live in atm is
generally slowing everything down i hope to stay on a maintained v9 until
late next year - and that's totally fine with me.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gluster.org/pipermail/gluster-users/attachments/20211227/8a36c857/attachment.html>