I’m getting very long wait times for pages to reload and or navigate etc. Maybe just a side effect of a small private server?
I received “server is too busy” error when trying to search, and then this
It’s gotten progressively slower for me over the last few weeks.
There seems to be two things going on making the server slow:
- The server is a “virtual” server, and it’s storage is a SCSI-over-ethernet server cluster which is shared between many servers. Either it’s become more busy lately, or they are performing maintenance on it.
- It seems like the database the crucible runs on gets slower over time. However, it is possible to defragment it, which makes it faster again. Unfortunately that takes the server offline for a little while, but it seems that I will need to do that.
Hmm, further investigation indicates that it’s a CPU time problem, not a disk speed problem… Running a site discourse update seems to have made things speedier, but I’m not sure if it is temporary or not.
Marked improvement today.
Is it that time again maybe?
“Saving” replies and general navigation loading pages pretty sluggish lately.
I ran some updates last night, not sure how much it helped though.
If it’s still slow, maybe I should try an ‘app rebuild’ to see if that helps.
I’ve had a few of that ‘Error 502’ thing today when I was logging in from work. I just assumed it was an issue with my work’s IT (which I don’t believe is the best, but it has got better lately) but maybe it wasn’t.
It’s been OK today for me.
Until today I was getting regular “this site is under maintenance” messages and generally not loading or slow loading.
Site is still slow, and I’m going to try a rebuild tonight to see if that helps.
In the meantime, I’ve configured nginx to wait longer before timing out.
Hopefully this will allow the crucible to work, even when it is slow.
Did a full rebuild, and a few other optimizations I found on the discourse site.
Seems ok right now, but we’ll see how long it lasts.
Could it be a memory leak? It seems to be OK for a bit then grinds to a halt.
Maybe, but nothing seems to grow HUGE if I leave it running for a while so it doesn’t really make sense for it to be a memory leak.
I have a discussion going about it on “meta” the discourse discourse forum. If anybody wants to follow along, you can find the discussion here:
Tried rebooting the server.
Of course, things seem snappy again, but we’ll see how long it lasts.
One possibility is that it’s a memory leak in linux itself.
According to “top” I was using a lot of memory on my server, but I couldn’t actually figure out where it was going to. Rebooting the server may help that.
Feels like the reboot has helped.
Still pretty snappy.
Hopefully I didn’t just jinx it…
It’s zippy tonight!
It runs fine now