Re: idle RAID1 cpu usage

[Paul Slootman]

Jan Engelhardt  <jengelh@xxxxxxxxxxxxxxx> wrote:
>On Dec 27 2006 19:09, Paul Slootman wrote:
>>
>>This works fine, but I noticed that quite some time was being used by
>>the mdX_raid1 threads; even on a partition that's not even being used at
>>this time... (it was an empty mounted filesystem, but I umounted it to
>>be sure the filesystem code wasn't causing some IO).
>>
>>My question is: why is CPU being used by the RAID1 threads, even for
>>those devices that are otherwise unused? What are they doing?
>
>First-time synchronization (if applies), otherwise I suspect some
>housekeeping (bitmap perhaps?).

No, I had rebooted after the initial synchronization was complete.

>Otherwise it seems like a question who does what. On a raid5 array (i.e.
>cpu-heavy), both smbd and mdx_raid5 accumulate time for, of course, xor
>calculation.

On those devices that are actually in use, I'd understand CPU being
used; but where the device is not even mounted or used for swap or
whatever, I find that amount of CPU usage very high for just checking
the bitmap... I may take the idle device apart and reassemble without a
bitmap to see whether that makes any difference.


Paul Slootman
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/