Friends of #BSDCafe and the #Fediverse, the BSD Cafe VPS is starting to show some signs of load. The average is low, but there are moments, especially during hourly backups, when RAM is under pressure. This morning, there seemed to be an OOM issue that killed Ruby. Since the services are managed by daemon(8) on #FreeBSD, everything returned to normal operation, but I believe it's time to give the system a little more power.

A user encountered some access issues, so I delved into the logs for further investigation. It appears that the increased RAM load during backups might be causing these intermittent problems. I will be proactive in addressing this and optimizing the system to ensure smooth access and operation.

For now, I've increased the RAM of the current VPS, but since I have available hardware, I will split the services. I don't want to change the IP address at this time, so the reverse proxy and mail server will stay on the current VPS, while the rest (PostgreSQL, Redis, #Mastodon itself) will run on another server - of course, in VPN with the current one.

The process will be straightforward, as I'll simply move (zfs-send / zfs-receive) the respective jails. There will be some brief downtimes during the service restarts (very short, restarting one service at a time). I'll likely proceed in the evening and will write something before to give a heads-up.

#ServerUpgrade #ServiceMigration #ServerManagement #Infrastructure #BSDCafeUpdates