How can I see which of the services is eating my memory ?
Memory used by services is continuously increasing
htop is probably a good place to start. Or top -o RES. Latter is more cross-platform and the -o RES sorts by resident memory for the process (in KiB). You can press the q button to exit when done looking. It’s easiest to inspect this via SSH (not webshell). Do note that in some older releases memory being used by ZFS was reported as “services” and so the UI stuff was misleading.
“Services” includes whatever Apps you have installed, and you can see what memory they’re using in the Apps section of the web UI:
Sorry, not relevant. OP didn’t initially specify CORE
I’m on core so it is top -o res
But ok if this is all …
PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU COMMAND
2115 root 14 20 0 6951M 5711M kqread 10 8:30 4.78% bhyve
2223 root 30 20 0 9222M 5696M kqread 10 15:36 9.56% bhyve
453 root 34 20 0 475M 276M kqread 8 0:31 0.24% python3.9
583 root 3 20 0 264M 178M usem 10 0:08 0.00% python3.9
782 root 3 22 0 231M 178M usem 2 0:08 0.00% python3.9
779 root 3 20 0 231M 178M usem 4 0:07 0.00% python3.9
781 root 3 20 0 231M 178M usem 3 0:08 0.00% python3.9
780 root 3 20 0 221M 176M piperd 10 0:07 0.00% python3.9
1998 root 1 21 0 148M 120M kqread 3 8:47 2.37% smbd
2353 root 1 20 0 144M 114M kqread 3 0:00 0.00% smbd
1872 root 1 20 0 140M 112M kqread 6 0:00 0.00% smbd
1888 root 1 20 0 138M 109M kqread 2 0:00 0.00% smbd
1889 root 1 20 0 138M 109M kqread 0 0:00 0.00% smbd
1887 root 1 20 0 91M 65M kqread 1 0:00 0.00% winbindd
610 root 5 20 0 92M 58M usem 5 0:01 0.00% python3.9
1935 root 11 20 0 114M 52M nanslp 10 0:06 0.68% collectd
byhve is guilty and I do have to upgrade to 13.3 cos there the mem hole should be solved
1 Like
Without further context on what you’re seeing, it looks like the bhyve VMs are the pigs.
Yes they are… one is linux on is windows and a memory eating task on linux let both explode.
It was funny to see that both vm res mem starts to increase at the time where the linux vm task was started.