Re: Re: fr1-2.16 patch for 2.6.10 kernel

Peter T. Breuer a écrit :
> "Also sprach denis bonnenfant:"
> (of the patch I made)
> 
>>It seems ok. there are some little changes from my files, but as far as 
>>i can see, nothing harmful.
>>
>>I recompile a fresh kernel...
> 
> 
> It occurs to me to wonder if it is a caching effect.  You didn't see the
> effect with 9MB loopback devices, and I don't see it with 8MB hda
> partitions. Maybe a struct is recycled under memory pressure that
> remains stable when not under pressure.

I think you are right :  problem occurs with 500mb partitions.

> Can you make yourself a couple of small 8MB partitions on your disk and
> try it with that?  

Done. 8 mb on Sda2 and sdb2  No problems ! fast resyncs all the time !

Another clue about a struct not properly locked or released  : when 
mounting my 500 mb raid, and accessing it, the problem seemed to be less 
frequent. I umount it, and then total resync occurs 4 times ! the fifth 
one, disk was recognised, but a very old bitmap  was used ! causing 
resync of already clean sectors...


  raid1: Disk failure on sda1, disabling device.
         Operation continuing on 1 devices
raid1: made bitmap f7cbb740 at events 0:c8
md0: md_check_recovery 2: repair bit 0 on sb f7fc8480 preserved
RAID1 conf printout:
  --- wd:1 rd:2
  disk 0, wo:1, o:0, dev:sda1
  disk 1, wo:0, o:1, dev:sdb1
RAID1 conf printout:
  --- wd:1 rd:2
  disk 1, wo:0, o:1, dev:sdb1
md: trying to hot-add unknown-block(8,1) to md0 ...
md0: repair of faulty disk 800001!
md: trying to remove unknown-block(8,1) from md0 ...
md0: notifying dev 800001 it is no longer in array
md: unbind<sda1>
md: export_rdev(sda1)
md 1205: unlock_rdev nulls bdev
md 1197: lock_rdev restores bdev 800001
md0: repairing old mirror component 800001 (disk 201 >= bitmap 78)
md: bind<sda1>
md0: notifying dev 800001 it is now in array
md0: set repair bit to 1 on superblock f7fc8480
md0: md_check_recovery 2: repair bit 100 on sb f7fc8480 preserved
RAID1 conf printout:
  --- wd:1 rd:2
  disk 0, wo:1, o:1, dev:sda1
  disk 1, wo:0, o:1, dev:sdb1
md: syncing RAID array md0 (sb f7fc8480)
md: minimum _guaranteed_ reconstruction speed: 10000000 KB/sec/disc.
md: using maximum available idle IO bandwith (but not more than 200000 
KB/sec) for reconstruction.
md: using 128k window, over a total of 489856 blocks.
md0: repair bit set on sb so retained bitmap f7cbb740
raid1: synced dirty sectors 0-255
raid1: skipped clean sectors 256-255
raid1: synced dirty sectors 256-511
raid1: skipped clean sectors 512-490111
raid1: synced dirty sectors 490112-490367
raid1: skipped clean sectors 490368-979711
md: md0: sync done.
RAID1 conf printout:
  --- wd:2 rd:2
  disk 0, wo:0, o:1, dev:sda1
  disk 1, wo:0, o:1, dev:sdb1
md0: md_check_recovery 1: repair bit 0 on sb f7fc8480 preserved
md0: md_check_recovery 2: repair bit 0 on sb f7fc8480 preserved