Re: Questions about software RAID level 1 (how long should it take to sync, etc?)

[Craig Sanders <cas@taz.net.au>]

On Tue, Apr 04, 2006 at 06:19:27PM +1200, Dave Watkins wrote:

> > "synchronize" the drives, aren't they already sync'd?  Shouldn't
> > this be a much faster process?
>
> No, you copying the data isn't a "sync". You could have left the new
> drive blank and then added it and the same process would have occured,
> but you are right to be concerned about 25hour rebuild time.

yep, the raid system doesn't know (or care) what data is/was on the
mirrored to drive - as far as it's concerned, it's just a blank drive to
mirror every sector onto.


> Judging by the Asterisk problems you mentioned in another post it
> sounds like one or both of your HDD's don't have DMA turned on (hdparm
> -d <device> to check).

also, try unmasking interrupts 'hdparm -u 1 <device>' on both drives.

the man page has this to say about the unmask option:

       -u     Get/set interrupt-unmask flag for the drive.  A setting
              of 1 permits the driver to unmask other interrupts
              during processing of a disk interrupt, which greatly
              improves Linux's responsiveness and eliminates "serial
              port overrun" errors.  Use this feature with caution:
              some drive/ controller combinations do not tolerate the
              increased I/O latencies possible when this feature is
              enabled, resulting in massive filesystem corruption.  In
              particular, CMD-640B and RZ1000 (E)IDE interfaces can be
              unreliable (due to a hardware flaw) when this option is
              used with kernel versions earlier than 2.0.13.  Disabling
              the IDE prefetch feature of these interfaces (usually a
              BIOS/CMOS setting) provides a safe fix for the problem for
              use with earlier kernels.




the comments it makes about using this feature with caution seem to be
overly paranoid or specific to the (antique and crappy) IDE interfaces
mentioned. i've been using '-u 1' on ALL my drives since the mid
1990s with never a problem. hdparm's man page makes similar comments
about enabling DMA, and that's been pretty nearly flawless since the
late 1990s (although it's still possible to run into drive/interface
combinations which dont like it).

craig

-- 
craig sanders <cas@taz.net.au>           (part time cyborg)