Re: System load of 1 with CPU 100% idle

[Roger Pau Monné <roger.pau%entel.upc.edu@localhost>]

Hello,

Here's the output from raidctl -v -s raid0 when load is 1:

odin# raidctl -v -s raid0
Components:
           /dev/sd0a: optimal
           /dev/sd1a: optimal
No spares.
Component label for /dev/sd0a:
   Row: 0, Column: 0, Num Rows: 1, Num Columns: 2
   Version: 2, Serial Number: 2011062201, Mod Counter: 145
   Clean: No, Status: 0
   sectPerSU: 128, SUsPerPU: 1, SUsPerRU: 1
   Queue size: 100, blocksize: 512, numBlocks: 1952476416
   RAID Level: 1
   Autoconfig: Yes
   Root partition: Yes
   Last configured as: raid0
Component label for /dev/sd1a:
   Row: 0, Column: 1, Num Rows: 1, Num Columns: 2
   Version: 2, Serial Number: 2011062201, Mod Counter: 145
   Clean: No, Status: 0
   sectPerSU: 128, SUsPerPU: 1, SUsPerRU: 1
   Queue size: 100, blocksize: 512, numBlocks: 1952476416
   RAID Level: 1
   Autoconfig: Yes
   Root partition: Yes
   Last configured as: raid0
Parity status: clean
Reconstruction is 100% complete.
Parity Re-write is 100% complete.
Copyback is 100% complete.

I don't think there's any parity rewrite going on, raid seems to be
ok. Here's a daily graph of the load of the system:

http://i52.tinypic.com/11943g9.png

Thanks, Roger.

2011/7/2 Manuel Bouyer <bouyer%antioche.eu.org@localhost>:
> On Wed, Jun 29, 2011 at 06:27:41PM +0200, Roger Pau Monné wrote:
>> Hello,
>>
>> I'm seeing some strange behaviour in a NetBSD installation with
>> RAIDframe, the load of the system goes to 1 with no reason, an then
>> goes to 0 again after a period of time (usually after some hours). The
>> output of the "top" command is the following:
>>
>> [...]
>> System is always in "raidiow" state, but at the same time CPU is
>> always 100% idle, so I don't know how I have an average load of 1 if
>> the CPU is not used. I have another system, with the same hardware and
>> configuration (except for the software disk RAID) that is reporting
>> load as expected, and is also running a Xen Dom0, so it makes me thing
>> it has something to do with the software RAID. Could someone provide
>> and explanation of what's happening?
>
> load is more than just a function of CPU usage; for example I/O is also
> used in computation. If there a parity rewrite running when this happens ?
> If so, the load could be related to I/O generated by this ...
>
> --
> Manuel Bouyer <bouyer%antioche.eu.org@localhost>
>     NetBSD: 26 ans d'experience feront toujours la difference
> --
>