Seafile has jumped from 300MB RAM consumption to 950MB in a few hours

Seafile Server
1

Hi there…So I’ve had a few crashes on Seafile, and initially I didn’t pay much attention, until I’ve come to realize its memory consumption has climbed quite dramatically and in a very short period of time. This is now stable at around 940ish MB of RAM in my LXC container, but it was hovering for months at 256~300MB of RAM, with some few peaks to 500-600MB (hence the crashes, since my container was set for 1GB of RAM).
Is there a way to analyze what’s most likely taking so much RAM, specially all of the sudden? I’m kinda suspecting MySQL, but not sure how to troubleshoot for this.

Thanks!

2

You can login to the server to check which process takes most memory.

Which Seafile version do you use? It may be memory leak in older versions.

3

Thanks for your reply. I’m currently on 11.0.9. Not the latest, but not a really old one. Not sure if htop is the most accurate way to show it, but I can see a bunch of mysqld processes all labelled at the same RAM usage, 454MB right now, and 22% of memory. Then there’s a few python processes running gunicorn, clocking at around 135MB and 118MB. Total usage is 992MB right now. This is a dedicated LXC container, nothing else runs in it.

4

You may consider adding more memory to the container.

5

Thanks, Jonathan. I did that in the first place. But my question still kinda remains, what could have made the MySQL to double up in memory needs in a day? Just so I know what to prepare for next time.