[sheepdog-users] Redundancy and High Availability

[Hitoshi Mitake]

On Sun, Sep 4, 2016 at 10:04 AM, Lindsay Mathieson <
lindsay.mathieson at gmail.com> wrote:

> Hi all, playing around with sheepdog on Proxmox, which packages the 8.3.1
> version, but will be upgrading to 1.0 soon.
>
>
> I've setup a test 3 node cluster and formatted it with copies=3, installed
> a couple of VM's using the block driver - write performance is really only
> acceptable if I turn sync off. Haven't tried a SSD journal yet.
>
>
> From what I saw in the docs, and in testing - you can only write to the
> cluster with copies=3 when all three nodes are up, which isn't terribly
> useful. Kinda negates the whole point of redundancy. With Ceph and Gluster
> I can specify replica 3 and if a node goes down it allows writes and heals
> the data intelligently when the ndoe comes back up.
>
>
> Is this possible with sheepdog?


Sorry for my late reply.

If you have more than 3 nodes, losing single node doesn't prevent write. If
you cannot have a cluster with more than 3 nodes, how about create a
cluster with --copies=2?

Thanks,
Hitoshi


>
>
> --
> Lindsay Mathieson
>
> --
> sheepdog-users mailing lists
> sheepdog-users at lists.wpkg.org
> https://lists.wpkg.org/mailman/listinfo/sheepdog-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.wpkg.org/pipermail/sheepdog-users/attachments/20160928/3d0869b8/attachment.html>