Gus Wirth wrote:
> There was some discussion at the December meeting about replacing the
> existing KPLUG server. I can't remember who got the last one, I think
> Joshua? Anyway, we have the funds to get a new one so I say Neil
> delegates this to whoever wants to volunteer and we go ahead and do it.
> How much capability do we need? That will determine the price.
I have LogWatch logs going back to March 17. I could make a graph
showing disk usage, over time. I seem to see some spikes in September,
as far as / goes. I was not paying as much attention to the other
partitions as much.
Current disk usage:
Wed Dec 20 06:25:38 2006
/dev/sda3 7.5G 3.3G 4.3G 44% /
/dev/sda1 89M 5.3M 79M 7% /boot
/dev/sda7 16G 1.4G 14G 10% /home
/dev/sda6 957M 50M 908M 6% /tmp
/dev/sda5 9.4G 7.2G 2.2G 77% /var
A further breakdown of /var:
1.2G /var/amavisd
1.1G /var/amavisd/quarantine
5.3G /var/lib
1.6G /var/lib/mailman
1.4G /var/lib/zope2.7
2.3G /var/lib/zope2.8
The trend appears to be relatively stable, with a couple spikes in
September. I can generate more precise statistics if anyone is
interested.
The system load is currently:
top - 08:28:47 up 299 days, 13:39, 3 users, load average: 1.45, 1.14, 0.73
Top five processes at that time:
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
508 jaqque 19 0 1104 1104 844 R 3.8 0.2 0:00.35 top
28560 zope 11 0 337m 134m 92m S 1.6 27.6 22:39.81 python2.3
6 root 10 0 0 0 0 S 0.3 0.0 35:47.88 kupdated
28556 zope 10 0 337m 134m 92m S 0.3 27.6 1:16.37 python2.3
31136 www-data 9 0 3440 2160 1964 S 0.3 0.4 0:01.09 apache2
du's seem to take a long time, so I would say that the disk i/o is the
bg bottleneck vice CPU horsepower.
-john
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-steer
