setiq.muc-central.com downtime

[Mucman]

Sorry about the downtime last night. I believe the server was unresponsive from around 1:00am - 9:00am PST.

It turns out that there is a strange memory leak in JSetiDoor. Or it just handles network connections very poorly.
Here's line 4 from `netstat -m`

128/4576/4576 mbuf clusters in use (current/peak/max)

As you can see the server maxed out its mbuf clusters. I could tune the kernel to get more, but the problem will
still happen. I'm going to configure JSetiDoor to restart every night or every two nights now. When it restarts it
takes about 5-10 minutes to load all the config data and stuff... Hopefully this will have minimal impact on your
proxy connections.

Other than the JSetiDoor restarts every night (probably will occur at around 1-2am) there are no expected downtimes
for the server right now. It's current uptime is 61 days and would be 120 if I didn't accidently hit `shutdown -h now`
through an ssh connection from home 🙂.

 

[Smoke]

Thanks for the update, Mucman.

Every couple of weeks or so I always go to my workshop and restart the SetiQueue machine just for the heck of it.

Heck, I remember not too long ago that is something I had to do to several times @ day just to keep any computer from crashing. 😛 🙂

 

[Mucman]

Hehe, for some reason I don't believe in "rebooting" 🙂. There are many things that can be fixed by HUPing a process and
I prefer to figure it out (albeit the problem might linger while I figure it out 🙂).

You wouldn't believe how hard it was for me to shutdown a linux nameserver at work that was up for 421 days! I soon felt
better when I removed RedHat and put on FreeBSD 😛

My new stats addiction

When I tell the ladies I have remarkable uptime, I'm not lying 😀