Hi all,
I’d like to pick your brains for a moment if I may.
I’ve just setup mailcow on a physical 4 core CPU system with 16GB of ram and mirrored 1TB M2 SSDs and the only other thing it’s doing is a couple of containers for PDNS and pdnsAdmin. It’s on a gigbabit connection connected from the router via fibre. The base OS is a minimal ALMalinux (rhel 10.2 clone):
cat /etc/redhat-release
AlmaLinux release 10.2 (Lavender Lion)
The mailcow dashboard shows between 1 and 10 percent CPU usage but closer to 1 most of the time.. it jumped a couple of percent when I expanded every container shown in the dashboard to see if any were spiking. but overall it’s reporting well below 10% CPU and 24% of ram used. None of the containers show high usage on the dashboard either.
Mailcow has 4 domains setup but only about 8 or 10 users. When it drops a connection on a machine, it will no longer show Sogo or allow IMAPs connections. A port 143 connection to the old mail server (so they can drag stored mail over to the new Imap account before I shut it down) still works fine so the issue is the new box and not the connection, but I’ll go through it briefly anyway.
AlmaLinux (docker host) own system display shows 12GB of free ram and 1% CPU and the CPU is 18 degress (Its cold today) Network utilisation on the host is low in the kb too. The Pdns containers are not showing much resource usage either.
Host top:
top - 13:41:12 up 23:43, 3 users, load average: 0.06, 0.07, 0.08
Tasks: 402 total, 1 running, 401 sleeping, 0 stopped, 0 zombie
%Cpu(s): 1.9 us, 2.3 sy, 0.0 ni, 95.4 id, 0.0 wa, 0.2 hi, 0.2 si, 0.0 st
MiB Mem : 15505.5 total, 9145.9 free, 4146.1 used, 2575.2 buff/cache
MiB Swap: 7948.0 total, 7948.0 free, 0.0 used. 11359.4 avail Mem
The router is a 4core job showing very low utilisation. An internet speed test showed tripple digits and low response time.
I do not know what the bottleneck is here. At first I thought it was the client machines but its happened to a few and on different networks and the only thing they have in common is that they are running Windows 11. I tried turning off the windows firewalls when it first happened and it made no difference. The only thing that seems to help is restarting mailcow…. and minutes after that it stops for those machines again.
Does anyone have any suggestions as to what could be causing the bottleneck? This is not a commercial server, so it’s never going to have more than 20 users on it.. All parts of the hardware should be able to handle that fine.
Any help would be much appreciated.
cheers,
Frank