Re: Re: fr1-2.16 patch for 2.6.10 kernel

"Also sprach denis bonnenfant:"
> It's better. fast resync 5-6 times successivly, but now,

OK.

> sda1 (read) [events: 000000e6]
> md: *** lecture uuid  800001
> md: old uuid 5dc1267b 788f7499 cd49d3c0 d06f0b96
> md: new uuid 5dc1267b 788f7499 cd49d3c0 d06f0b96
> md0: new disk 800001 too old for repair (disk 230 < bitmap 232)  <****** 
> still a problem ?

That's a difficult one.  The problem is that I don't exactly know the
algorithm that is used to stamp the event counter on the disk
superblocks and on the array superblock.  I assume that removing a
device or setting it faulty is an "event" that increments the array
superblock event counter, but obviously can't affect the faulted or
removed disk.  So I am willing to assume that a returning disk is OK if
its counter is "one-behind" the value shown on the bitmap as being the
array event count when the bitmap was started.

But two behind? 

> md: bind<sda1>
> md0: notifying dev 800001 it is now in array
> md0: set repair bit to 0 on superblock f7fc8480
> md0: md_check_recovery 2: repair bit 0 on sb f7fc8480 preserved
> 
> it occurs just one time, and now it's ok again.

Peter