Random crashing

[agentgreen]

ag@bigassnas:~$ top -bn1 | grep -i -e USER -e sabnzbd -e rar
top - 08:52:21 up 25 days, 20:37, 2 users, load average: 0.31, 0.09, 0.07
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
20312 ag 20 0 262m 54m 3444 S 2 3.8 3:55.35 sabnzbdplus



ag@bigassnas:~$ free -m
total used free shared buffers cached
Mem: 1443 1218 224 0 374 196
-/+ buffers/cache: 647 795
Swap: 90 90 0

[agentgreen]

So I can't access Sabnzbd+ right now. The service is running:


ag@bigassnas:~$ ps axuwwww|grep -i sabnzbd
ag 26275 0.0 2.2 140380 33432 ? Sl Aug03 0:39 /usr/bin/python -OO /usr/bin/sabnzbdplus --daemon


But is not listening on 9090:

ag@bigassnas:~$ netstat -na|grep 9090 |grep -v CLOSE
ag@bigassnas:~$

The other strange thing, and this may be a problem with Ubuntu, I have a bunch of CLOSE_WAIT to port 9090 in netstat from the ip of the workstation I'm trying to access sabnzbd from:

ag@bigassnas:~$ netstat -na|grep 9090 |grep CLOSE_WAIT |grep 10.8.0.10|wc -l
1009


Also, I'm seeing a bunch of these in the error log, not sure if it's related or not:

FatalSSLAlert: [('SSL routines', 'SSL3_WRITE_BYTES', 'ssl handshake failure')]


Restat Sab, and all is well.

[sander]

Wow! Those 1009 CLOSE_WAIT sessions sounds bad; I checked my Ubuntu (11.04) and there are ... none.

So what if you disable the HTTPS connection in the sab config? Does it still occur?

[agentgreen]

Wow! Those 1009 CLOSE_WAIT sessions sounds bad; I checked my Ubuntu (11.04) and there are ... none.

So what if you disable the HTTPS connection in the sab config? Does it still occur?

Yep, had it happen again just now. If I stop Sabnzbd+ they all go away.

It's almost like Sabnzbd is hitting a connection limit and stops responding. By restarting Sabnzbd+ it clears up the connections and I am good to go again.

My problem doesn't seem to be the same as OPs, so I'll start a new thread...

[agentgreen]

You know what, I'm running Sabconnect++, which I would assume tries over and over to connect to Sab to update.

I uninstalled the add-on in Chrome, we'll see if that resolves it.

[agentgreen]

So because I was inquisitive, I decided to uninstall Sabconnect+ and reinstall it. Then I watched my netstat for CLOSE_WAIT. At no point did I go to Sabnzbd+ from my browser, the only thing connecting was Sabconnect+.

An hour or so later, 900 CLOSE_WAIT messages.

Must be my plugin....goodbye.

[sander]

So because I was inquisitive, I decided to uninstall Sabconnect+ and reinstall it. Then I watched my netstat for CLOSE_WAIT. At no point did I go to Sabnzbd+ from my browser, the only thing connecting was Sabconnect+.

An hour or so later, 900 CLOSE_WAIT messages.

Must be my plugin....goodbye.

Very interesting. Have you informed the maker of Sabconnect+?

BTW: what was the message you got in your webbrowser when trying to connect to the unreachable SAB?

[agentgreen]

I have not informed him yet, but I will.

I don't exactly remember the message, but I believe it was a timeout error.

[coatezy]

Thanks for the heads up on the Sabconnect+ issue.. Sab crashing on my server daily was really starting to get annoying.. Didn't even think that something connected via the api could be causing the problem. Removed the extension and all sorted!

[shypike]

What an app goes opening sockets like mad and not closing them, any system will break down.

[agentgreen]

So this problem has cropped up again. Not sure what's causing it now, maybe Sabmonitor...

I wrote this quick script to check number of CLOSE_WAIT's and if it's greater than 20, it will restart SAB. It's a hack but I like having a Sabmonitor type plugin, so it will work for me.

#!/bin/bash

NUMBER=`netstat -na |grep 9090 |grep CLOSE_WAIT | wc -l`

if [ $NUMBER -gt 20 ]
then
echo "" | mail -s "Sabnzbd had to be restarted" [email protected]
/etc/init.d/sabnzbdplus restart
fi

[wonderchimp]

Hello again, the problem cropped up again for me. An update had solved my problem previously but then started again. This time I seem to have permanently solved my problem and have been running with no down time for a month now.

I was using a NAS but my complete/incomplete folders were stored locally and my "movies","TV", "Games" folders were stored on the NAS so when I downloaded anything sab would move that category file to the appropriate folder. The problem was that the drives would spin down after 30 minutes of inactivity, I changed the settings so that they did not spin down and the problem has been solved.