Gluster users - Aug 2020 - Upgrade from 6.9 to 7.7 stuck (peer is rejected)

Hello Mabi

You don't need to follow the offline upgrade procedure. Please do follow
the online upgrade procedure only. Upgrade the nodes one by one, you will
notice the `Peer Rejected` state, after upgrading one node or so, but once
all the nodes are upgraded it will be back to `Peer in
Cluster(Connected)`.  Also, if any of the shd's are not online you can try
restarting that node to fix that. I have tried this on my own setup so I am
pretty sure, it should work for you as well.
This is the workaround for the time being so that you are able to upgrade,
we are working on the issue to come up with a fix for it ASAP.

And, yes if you face any issues even after upgrading all the nodes to 7.7,
you will be able to downgrade in back to 6.9, which I think you have
already tried and it works as per your previous mail.

Regards
Nikhil Ladha
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.gluster.org/pipermail/gluster-users/attachments/20200824/cb10f42b/attachment.html>

Dear Nikhil,

Thank you for your answer. So does this mean that all my FUSE clients where I
have the volume mounted will not loose at any time their connection during the
whole upgrade procedure of all 3 nodes?

I am asking because if I understand correctly there will be an overlap of time
where more than one node will not be running the glusterfsd (brick) process so
this means that the quorum is lost and then my FUSE clients will loose
connection to the volume?

I just want to be sure that there will not be any downtime.

Best regards,
Mabi

??????? Original Message ???????
On Monday, August 24, 2020 11:14 AM, Nikhil Ladha <nladha at redhat.com>
wrote:
> Hello Mabi
>
> You don't need to follow the offline upgrade procedure. Please do
follow the online upgrade procedure only. Upgrade the nodes one by one, you will
notice the `Peer Rejected` state, after upgrading one node or so, but once all
the nodes are upgraded it will be back to `Peer in Cluster(Connected)`. Also, if
any of the shd's are not online you can try restarting that node to fix
that. I have tried this on my own setup so I am pretty sure, it should work for
you as well.
> This is the workaround for the time being so that you are able to upgrade,
we are working on the issue to come up with a fix for it ASAP.
>
> And, yes if you face any issues even after upgrading all the nodes to 7.7,
you will be able to downgrade in back to 6.9, which I think you have already
tried and it works as per your previous mail.
>
> Regards
> Nikhil Ladha
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.gluster.org/pipermail/gluster-users/attachments/20200824/02ab61d1/attachment.html>