[CentOS] RAID rebuild time and disk utilization....

Ross Walker rswwalker at gmail.com
Mon Sep 27 22:54:18 UTC 2010


On Sep 27, 2010, at 1:24 PM, Les Mikesell <lesmikesell at gmail.com> wrote:

> On 9/27/2010 12:19 PM, Benjamin Franz wrote:
>> On 09/27/2010 08:15 AM, Tom Bishop wrote:
>>> So I'm in the process of building and testing a raid setup and it
>>> appeared to take along time to build I came across some settings for
>>> setting the min amount of time and that helped but it appears that one
>>> of the disks is struggling  (100 utilization) vs the other one...I was
>>> wondering if anyone else has seen this and if so, is their a solution
>>> for it...my 2 disks are 1 Samsung F3 1tb /dev/sdb and 1 Seagate
>>> 7200.12 1Tb /dev/sdc...smartctl looks good on both....
>>> 
>>> 
>> 
>> [...]
>> 
>> What is the output from 'cat /proc/mdstat'?
> 
> Also, head position/motion is the usual limiting factor with disk 
> performance.  Do you have other disk activity that will keep yanking the 
> head on the active drive away from the track currently needed for the 
> rebuild?

Also, also, if any of the drives are operating in PIO mode (Legacy Mode) then rebuild times will be horrific, as well as performance afterward. Make sure all drives are in SATA/AHCI mode.

-Ross
 
PS look at svc_tm in iostat and make sure all drives are performing correctly. Maybe a bad cable or flakey port expander...


More information about the CentOS mailing list