Absurd Services RAM use.

[Whattteva]

Is this normal? I'm used to seeing ZFS cache taking that much, but Services taking that much? Leaving practically nothing for ZFS cache.
I don't think I did anything different other than renaming my pool using the GUI export -> zfs export -> zfs import old_name new_name -> GUI import.

I may have done other things after that, but a few minutes later I noticed the chart looks absurd. Honestly, this caught me totally by surprise because I have never seen this in my 10 years running TrueNAS since the days of FreeNAS 8.3!

 

[winnielinnie]

What does the command arc_summary | head -n 19 reveal?

 

[Davvo]

Wow.
Did you reboot? This looks so extreme that could very well be a GUI bug.

 

[artlessknave]

uh. ya, that is not normal. it would appear you have a service that might have a leak or something?

 

[Jailer]

I'd check top before a reboot to see what's eating all your memory.

 

[Whattteva]

Yeah unfortunately, I rebooted cause I was in the middle of doing other things and didn't have time to troubleshoot it. I definitely wish I could've tried to see what was going on. Maybe it'll come around again when I have time to tinker with it. It wasn't a big issue cause it seems like a one-off thing but I just wanted to post it cause I thought the chart looked really interesting.

 

[Whattteva]

I didn't mention this, but what made that chart even more absurd is the fact that I don't even run much on it. Just SSH, UPS, SMART, SMB, and NFS. No jails nor VM's. I think that's such a light usage that RAM for services probably shouldn't even exceed 4 GB. Of course, prior to today, I have never tried to rename my pool either. So this is basically the only time I've ever done that.

 

[qq8554650]

scale also has the same problem, you see what I posted

 

[Whattteva]

scale also has the same problem, you see what I posted

You need to link it.

 

[cengjingcaihua]

I've noticed if I delete snapshots services mem usage will go up and ARC mem usage will go down.