[Gluster-users] Geo-replication adding new master node

[David Cunningham]

Hi Aravinda,

Thank you very much - we will give that a try.


On Mon, 31 May 2021 at 20:29, Aravinda VK <aravinda at kadalu.io> wrote:

> Hi David,
>
> On 31-May-2021, at 10:37 AM, David Cunningham <dcunningham at voisonics.com>
> wrote:
>
> Hello,
>
> We have a GlusterFS configuration with mirrored nodes on the master side
> geo-replicating to mirrored nodes on the secondary side.
>
> When geo-replication is initially created it seems to automatically add
> all the mirrored nodes on the master side as geo-replication master nodes,
> which is fine. My first question is, if we add a new master side node how
> can we add it as a geo-replication master?
> This doesn't seem to happen automatically, according to the output of
> "gluster volume geo-replication gvol0 secondary::gvol0 status". If we use
> the normal "gluster volume geo-replication gvol0 secondary::slave-vol
> create push-pem force" it says that the secondary side volume is not empty,
> which is true because we're adding a master node to the existing
> geo-replication.
>
>
> This is not automatic. Run `gluster-georep-sshkey generate` and georep
> create push-pem with force option to push the keys from new nodes to
> secondary nodes.
>
> You can also try this tool instead of georep create command.
>
> https://github.com/aravindavk/gluster-georep-tools
>
> $ gluster-georep-setup gvol0 secondary::slave-vol --force
>
>
> My second question is whether we can geo-replicate to multiple nodes on
> the secondary side? Ideally we would normally have something like:
> master A -> secondary A
> master B -> secondary B
> master C -> secondary C
> so that any master or secondary node could go offline but geo-replication
> would keep working.
>
>
> Geo-replication command needs one Secondary node to establish the session.
> Once session starts, Geo-rep starts one worker process per master brick.
>
> These worker processes gets the list of secondary nodes by running the
> `ssh <secondary-host> gluster volume info <secondary-volume>`. Then Geo-rep
> distributes the secondary nodes connection in round robin way. For example,
> if Master volume contains three nodes and secondary volume 3 nodes as you
> mentioned then Geo-rep makes connection as Master A -> Secondary A, Master
> B -> Secondary B and Master C -> Secondary C.
>
> Secondary node failover: If a node goes down in secondary cluster then
> Master worker connects to other secondary node and continues replication.
> One known issue is if the secondary node specified in the Geo-rep create
> command goes down then it fails to get the Volume info(To get list of
> secondary nodes). This can be solved by providing the list of secondary
> nodes as config(Not yet available).
>
>
> Thank you very much in advance.
>
> --
> David Cunningham, Voisonics Limited
> http://voisonics.com/
> USA: +1 213 221 1092
> New Zealand: +64 (0)28 2558 3782
> ________
>
>
>
> Community Meeting Calendar:
>
> Schedule -
> Every 2nd and 4th Tuesday at 14:30 IST / 09:00 UTC
> Bridge: https://meet.google.com/cpu-eiue-hvk
> Gluster-users mailing list
> Gluster-users at gluster.org
> https://lists.gluster.org/mailman/listinfo/gluster-users
>
>
> Aravinda Vishwanathapura
> https://kadalu.io
>
>
>
>

-- 
David Cunningham, Voisonics Limited
http://voisonics.com/
USA: +1 213 221 1092
New Zealand: +64 (0)28 2558 3782
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gluster.org/pipermail/gluster-users/attachments/20210601/485af931/attachment.html>