Yes, sorry. It is a problem that started over the weekend. I thought I had patched it by doubling the server’s RAM and adding a core, but that was not enough. Some process is causing the RAM use to spike and the image backend is crashing because of that.
Thanks. I noticed and reset the server a few minutes ago.
Something has been off recently. The CPU is spiking and the RAM gets used up, which crashes the pict-rs container. The pict-rs won’t reconnect until I reset the lemmy Docker container.
I doubled the RAM and added one core, but that was not enough to stop this problem, which means that whatever is causing these spikes is unconstrained. I need to look more deeply into Docker memory management to see if I can limit RAM usage such that the crash can be avoided while remaining functional.
Thanks. I have looked into it a bit more and I think that it is the postgres database grabbing all the memory it can. I have set a hard limit for the postgres container. Hopefully this resolves the problem!
Are images from mander.xyz broken for anyone else?
Yes, sorry. It is a problem that started over the weekend. I thought I had patched it by doubling the server’s RAM and adding a core, but that was not enough. Some process is causing the RAM use to spike and the image backend is crashing because of that.
@Sal@mander.xyz Another whack-a-mole.
Thanks. I noticed and reset the server a few minutes ago.
Something has been off recently. The CPU is spiking and the RAM gets used up, which crashes the pict-rs container. The pict-rs won’t reconnect until I reset the lemmy Docker container.
I doubled the RAM and added one core, but that was not enough to stop this problem, which means that whatever is causing these spikes is unconstrained. I need to look more deeply into Docker memory management to see if I can limit RAM usage such that the crash can be avoided while remaining functional.
Goodluck, comrade.
Thanks. I have looked into it a bit more and I think that it is the postgres database grabbing all the memory it can. I have set a hard limit for the postgres container. Hopefully this resolves the problem!