[ale] md2_raid1

Brandon Checketts brandon at brandonchecketts.com
Sun Jan 31 20:50:11 EST 2010


I'm not sure if the raid resync is part of the issue, but I recently found out 
that you can adjust the speed at which the rebuild tries to cat'ing new values to

/proc/sys/dev/raid/speed_limit_min
/proc/sys/dev/raid/speed_limit_max

See
http://www.brandonchecketts.com/archives/speed-up-a-linux-software-raid-rebuild

Thanks,
Brandon Checketts

Matt Rideout wrote:
> What's the output of:
> 
> cat /proc/mdstat
> 
> drifter wrote:
>> This FC11 Box has suddenly slowed  to a crawl.
>>
>> Top reports that md2_raid1 is sucking up 20 - 32%
>> of the cpu cycles and md2_resync is taking another
>> 2-5%.
>>
>> The log says this:
>>
>> Sarge kernel: md: data-check of RAID array md2
>> Sarge kernel: md: minimum _guaranteed_  speed: 1000 KB/sec/disk.
>> Sarge kernel: md: using maximum available idle IO bandwidth (but not more
>>   than 200000 KB/sec) for data-check.
>> Sarge kernel: md: using 128k window, over a total of 202001152 blocks.
>> Sarge kernel: md: delaying data-check of md0 until md2 has finished (they
>>   share one or more physical units)
>> Sarge kernel: md: delaying data-check of md1 until md2 has finished (they
>>   share one or more physical units)
>>
>> Is this routine, or an early hint that one of my hard drives may be about 
>> to die?
>>
>> If routine, can I create a cron job that does this work at a more useful 
>> time, such as 2 am?
>>
>> Sean
>>
>> _______________________________________________
>> Ale mailing list
>> Ale at ale.org
>> http://mail.ale.org/mailman/listinfo/ale
>> See JOBS, ANNOUNCE and SCHOOLS lists at
>> http://mail.ale.org/mailman/listinfo
>>   
> _______________________________________________
> Ale mailing list
> Ale at ale.org
> http://mail.ale.org/mailman/listinfo/ale
> See JOBS, ANNOUNCE and SCHOOLS lists at
> http://mail.ale.org/mailman/listinfo


More information about the Ale mailing list