Best practices for small/edge/IoT deployments? (remote management, power)
-
We're evaluating xcp-ng & XOA (right now XO from sources), and are unsure if we're going about this the right way. It's all working great, but we get the feeling that it's more targeted towards data centers and other "larger" deployments, and our needs are in some ways smaller.
Most of our deployments are single or dual machines, running on site in a more or less business critical role for our clients. In xcp-ng terms, we're running multiple small pools, managed over connections of varying quality. (sometimes even 4G) The sites running dual machines have continous replication set up, to speed up recovery in case of catastrophic failure. All machines are running local storage, with no shared storage available. Everything is backed up via XO.
To improve backup/replication performance we're running XO proxy, which works great. But it seems to not be possible to run redundant proxies? If the pool master (also running the VM with the proxy) were to fail we'd have to remote in via some other channel, make the backup machine master, and then start the replicated proxy VM to regain control. (and then go on to start the VM:s the client actually cares about via XO) Or is there any easier way?
We're also looking into monitoring the UPS at each site, to handle power outages gracefully. The idea right now is to have a small VM running NUT, among some other things for housekeeping. NUT would then log in via SSH on each host in the pool and issue a shutdown command. Is this the "xcp-ng way"?
So far we really like everything, but in questions like these we get the feeling that we're trying on shoes we haven't grown into yet. Single pane of glass, simple migration between pools, and integrated backups are very tempting though.
-
Hi,
We have similar deployments (and they are more and more common with Broadcom leaving a huge void even in this context).
There's probably some improvements possible, worth thinking about it. Have you already contacted Vates to get potential more advanced discussions about this use case?
-
@johnny I think the Vates VMS (whether you go with paid support or not) would fit your use case perfectly. Since each host is essentially a pool of one, you would pretty much have multiple pools managed by a single Xen Orchestra instance (giving you a single pane of glass, if you will). You could augment this with XCP-ng Center.
As @olivierlambert correctly inquired, you should contact the Vates Team and have an in-depth discussion into your situation.
