[Gluster-users] Rebooting cluster nodes - GFS3.8

[Bernhard Dübi]

Hi,
just wanted to write the same thing.
there was once a post that suggested to kill the gluster processes
manually but I guess rebooting the machine will do the same. the
clients will stall for a while and then continue do access the volume
from the remaining node.
it is very important that you checj the heal status before you bring
down the next node otherwise you could end up in a split brain
situation.
hope this helps
Bernhard

2017-12-05 18:36 GMT+01:00 Andrew Kester <aikester at aikester.com>:
> On my setup at least, just issuing the reboot command works without any
> issue.  I've done a number of rolling reboots for software / kernel upgrades
> in the manner you've described this way.
>
> The one gotcha I've found is when the node comes back online.  I manually
> check healing to ensure that everything is synced and back online before
> taking other nodes offline.
>
> ---
> Thanks,
>
> Andrew Kester
> The Storehouse
> https://sthse.co
>
> On 12/5/17 10:40 AM, Mark Connor wrote:
>>
>> I am running gluster ver 3.8 in a distributed replica 2 config. I need to
>> reboot all my 8 cluster nodes to update my bios firmware.  I would like to
>> do a rolling update to my bios and keep up my cluster so my clients don't
>> take an outage. Do I need to shutdown all gluster services on each node
>> before I reboot? Or just issue the reboot.
>>
>>
>> _______________________________________________
>> Gluster-users mailing list
>> Gluster-users at gluster.org
>> http://lists.gluster.org/mailman/listinfo/gluster-users
>>
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://lists.gluster.org/mailman/listinfo/gluster-users