[Mageia-sysadm] [sysadmin-reports] Hobbit [727252] valstar.mageia.org:disk CRITICAL (RED)

nicolas vigier boklm at mars-attacks.org
Sat Sep 15 19:05:47 CEST 2012


On Sat, 15 Sep 2012, Thomas Backlund wrote:

> Thomas Backlund skrev 15.9.2012 18:47:
>> nicolas vigier skrev 15.9.2012 17:51:
>
>>> I also added 30GB to /tmp.
>>
>> That will help not hitting this issue again...
>
> Here is a thought....
>
> What about sticking a 240GB SSD in valstar,
> that would carry /tmp and /var/lib/shedbot ?

Yes, that could be nice. If we take a 400 or 500GB SSD, we could also
have /distrib/bootstrap/distrib/cauldron on SSD, to speed up genhdlists
on cauldron.

>
> I mean...  looking at this:
> /dev/md0               20G  9.9G  8.4G  55% /
> /dev/sda1             981M   91M  840M  10% /boot
> /dev/sdb1             981M   91M  841M  10% /boot2
> /dev/mapper/vg0-tmp    60G  243M   56G   1% /tmp
> /dev/mapper/vg0-svn   394G  295G   80G  79% /svn
> /dev/mapper/vg0-var    18G  8.2G  8.5G  50% /var
> /dev/mapper/vg0-distrib
>                       886G  677G  165G  81% /distrib
> /dev/mapper/vg0-schedbot
>                       100G   31G   64G  33% /var/lib/schedbot
> /dev/mapper/vg0-binrepo
>                       247G  144G   90G  62% /var/lib/binrepo
>
>
> and filesystems tend to get slover when they hit ~85% mark...
>
> and according to vgdisplay we have only 105G left for extending the lv's
>
> so sticking /tmp and /var/lib/shedbot on the SSD would:
> - free up 160G for other lv's
> - speed up rpmlints and schedbot work
>
> WDYT ?
>
>
> btw, how many disks do valstar support? I mean at the point we branch mga3 
> we will need to split out /distrib to separate disks...
> (this would also help rsync processes not bog down i/o for the rest
>  of the workload happening on valstar)

valstar supports 8 disks. So we can add 2 disks to have more space,
and an SSD.



More information about the Mageia-sysadm mailing list