Re: How to replace or add a monitor in stretch cluster?

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



yes, I think so. I think the context in which this originally came up was
somebody trying to replace the tiebreaker mon.

On Fri, Dec 2, 2022 at 9:08 AM Sake Paulusma <sake1989@xxxxxxxxxxx> wrote:

> That isn't a great solution indeed, but I'll try the solution. Would this
> also be necessary to replace the Tiebreaker?
> ------------------------------
> *From:* Adam King <adking@xxxxxxxxxx>
> *Sent:* Friday, December 2, 2022 2:48:19 PM
> *To:* Sake Paulusma <sake1989@xxxxxxxxxxx>
> *Cc:* ceph-users@xxxxxxx <ceph-users@xxxxxxx>
> *Subject:* Re:  How to replace or add a monitor in stretch
> cluster?
>
> This can't be done in a very nice way currently. There's actually an open
> PR against main to allow setting the crush location for mons in the service
> spec specifically because others found that this was annoying as well. What
> I think should work as a workaround is, go to the host where the mon that
> failed to join the quorum but failed due to lack of crush location is, open
> the /var/lib/ceph/<fsid>/mon.<mon-id>/unit.run file, then at the very end
> of the last line (the last line should be a long podman/docker run command)
> append "--set-crush-location <crush-loc>".  Then, still on that host, a
> "systemctl restart <mon-service>" where mon-service is the systemd unit
> listed for the mon in "cephadm ls --no-detail". That should allow the
> monitor to at least join the quorum as it now has a crush location, and
> then you should be able to make other alterations a bit easier.
>
> On Fri, Dec 2, 2022 at 7:40 AM Sake Paulusma <sake1989@xxxxxxxxxxx> wrote:
>
> I succesfully setup a stretched cluster, except the CRUSH rule mentioned
> in the docs wasn't correct. The parameters for "min_size" and "max_size"
> should be removed, or else the rule can't be imported.
> Second there should be a mention about setting the monitor crush location
> takes sometime and know other ceph command can be used.
>
> But now I need to replace a few monitors (it's virtualized and machines
> need to be replaced). I use cephadm and have a label "mon" which is
> assigned to the monitor services.
> With the command "ceph orch host add <hostname> <ip-address> --labels=mon"
> I add normally a new monitor to the cluster.
>
> Only this results in the following error in de logs:
> 12/2/22 1:27:19 PM [INF] <hostname> attempted to join from <hostname>
> [v2:<ip-address>:3300/0,v1:<ip-address>:6789/0]; but lacks a crush_location
> for stretch mode
>
> Next I tried to set the CRUSH location like used in the docs for stretch
> mode with the command "ceph mon set_location <hostname> datacenter=<dc1>".
> This only results in the following error:
> Error ENOENT: mon.oqsoel11437 does not exist
>
> So how can I add/replace a monitor in a stretched cluster?
>
> Best regards,
> Sake
> _______________________________________________
> ceph-users mailing list -- ceph-users@xxxxxxx
> To unsubscribe send an email to ceph-users-leave@xxxxxxx
>
>
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx



[Index of Archives]     [Information on CEPH]     [Linux Filesystem Development]     [Ceph Development]     [Ceph Large]     [Ceph Dev]     [Linux USB Development]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [xfs]


  Powered by Linux