@acebmxer Hello, no. Let me describe my setup.
Esxi + vcenter + veeam. Vcenter wakes up veaam at 9am. Veeam sends wol packet to the synology nas at 9.15. At 9.30 the backup starts and ends more or less at 10.00
Due to inactivity synology stops the hard disks, then after some time it shut itself off. Veeam just see the repo as unavailable. At 12.00 veeam auto shuts down.
Migrating to xcpng + xo, veeam is retired. At first i added the remote to XO, believing it's like veeam. If the backend goes down it simply shows unavailable. Then i discovered is not at all like that when i found an inaccessible XO. After some chatgpt-ing i found out this "bug". I thought that the script would be sufficient, but it's not. Waiting for ping to fail and then disable the remote (cron every 5mins) still ends up in
ps -T -p 738 -o pid,tid,stat,wchan:32,comm
PID TID STAT WCHAN COMMAND
738 738 Ssl do_epoll_wait MainThread
738 753 Ssl do_epoll_wait DelayedTaskSche
738 754 Ssl futex_wait_queue V8Worker
738 755 Ssl futex_wait_queue V8Worker
738 756 Ssl futex_wait_queue V8Worker
738 757 Ssl futex_wait_queue V8Worker
738 758 Ssl futex_wait_queue SignalInspector
738 863 Dsl rpc_wait_bit_killable libuv-worker
738 864 Dsl rpc_wait_bit_killable libuv-worker
738 865 Dsl rpc_wait_bit_killable libuv-worker
738 866 Dsl rpc_wait_bit_killable libuv-worker
And it's even worse because not only XO becomes unavailable, but even if i disable manually (or via script) the remote as soon as the backup ends, with the synology still ON, the linux os does not unmount the nfs.
Today the remote was disabled before the nas shutdown and correctly XO kept working. But if at this point i want to restart the XO vm i must do this before shutdown
umount -l /run/xo-server/mounts/9fc232df-4d0c-4a15-93b5-c0585c1dd9c9
otherwise it stays like this for at least 15 minutes, then finally reboots
[image: image.jpeg]
The ideal behaviour would be that XO does all by itself, but since we can't have it all, the closest thing i have in mind is
Script that sends WOL to the nas, when it pings than enables the remote
Script that monitors the backup acrivity. when the last job is done, it must disable the remote and also umount the linux os nfs mount.
Then the nas could shut down automatically without issue.
Since i'm not a dev i need to take my time with chatgpt and see i what i can do.
Having said that, my concern is not about my homelab, it's about production. In my environment i can and want to simulate all kinds of malfuncions. In the previous message i said i still need to try to shut down the main nfs SR for vms and see how the host behaves.
But if i decide to deploy XEN at my company, i cant have 1 remote unavailability cause the whole XO (or XOA in that case) fall apart.
Hope it helps!!
