Swap didn't come up by itself on staging storage node

After rebooting staging node storage001 15 mins ago, swap did not come up cleanly.

journalctl -u mkswap-var-swapfile.service brings up:

Sep 23 11:08:44 storage001 mkswap-var-swapfile-start[550]: Attaching loopback device failed (loop device with autoclear flag is required).

... which I guess might be a new thing with the updated kernel or so.

I issued a systemctl restart mkswap-var-swapfile.service and swapon -a, which made it work.

TBD:

• Investigate what's up,
• make sure it doesn't happen again,
• understand whether this happens on the production grid too.