A horde of SOGO processes

jetkins

My server has been consuming more memory than normal recently, and today I finally decided to see why. 🙂 Turns out I have over 20 instances of sogod running, each consuming at least 0.8% memory - a total of over 16% of my available RAM. I don’t recall seeing that before, but maybe I overlooked it. Is that normal? (I don’t even have that many user accounts.)

Looking at the SOGO log, I see the following message recurring:
Jul 25 18:10:30 4b94a7025a1e sogod [74]: [ERROR] <0x0x55af9bcae280[SOGoDomainDefaults]> current source cannot synchronize defaults
so something isn’t quite right, but I’m not sure if/how this would cause so many processes to spawn.

Ideas and suggestions welcomed!

pkernstock

It’s normal that SOGo spawns multiple workers and threads in parallel - that’s nothing to be concerned about. Roughly 20 processes/threads from SOGo doesn’t sound that concerning either. Please take into account that when you’re using EAS/Carddav/caldav or other fancy stuff from SOGo, it will also consume more RAM.

I’m unsure what exactly your issue is here. So far as I’ve understood you think you have higher memory usage and also no issues at all, right?

jetkins

Huh, okay. My issue was that my server, which normally sits at around 65% memory utilization, was triggering an alert because it was hovering around 80%. After a reboot it’s back down to 65% again, though the sogod processes remain so apparently they weren’t the culprit.

I’m still curious about the occasional flood of the error message I posted, though - is that a cause for concern? Doesn’t happen constantly, but when it does, there’s a small flood of them.

jetkins

Answering my own second question: I found this SOGO bug report which suggests it was something that snuck in in a recent build and was fixed on 7/16. I ran a mailcow update.sh, and that message appears to have gone away.