Bug#336153: Reproduced, also using ext2 + dm-crypt on top of RAID5

[maximilian attems]

severity 336153 important
stop

2.6.8 is known to have an fucked md and doesn't stand heavy fileload
unless you are lucky.

the archive features now 2.6.14 please try it out,
beware that the new initramfs generating tools are still in dev

-- 
maks


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#336153: Reproduced, also using ext2 + dm-crypt on top of RAID5

[kiu]

Is it correct what Max wrote ?

The debian people are shipping a kernel in their stable production line, which
is known to crash your data (if you use a system invented to prevent data loss)
and there is no warning at all in the debian installer or after apt-get install
mdadm has been run.

Imho debian just told its users: We dont care about your data at all - use our
system and we will crash it for sure.

Why is there no new version of the kernel-image without any code changes just
displaying a big warning message stating that MD shouldnt be used ?
-- 
kiu


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#336153: Reproduced, also using ext2 + dm-crypt on top of RAID5

[Tillmann Steinbrecher]

Hello,

I managed to reproduce the bug using a newer kernel (actually, I didn't 
_try_ to reproduce it, I ran into it and a google search brought up this 
bug, which occured under almost identical circumstances).


My configuration is very similar to that of the original submitter; I 
also use dm-crypt with AES on top of a 4x160GB RAID5 system, and the 
problem also occurs when copying large amounts of data on the ext2 
partition.


There might possibly be an issue with ext2 on RAID5+dm-crypt. After 
reboot, the ext2 filesystem is erronous, and the above message appears 
even when reading it.


I have previously used the same encrypted volume with Reiser4, and there 
were no problems whatsoever (despite Reiser4 being experimental and ext2 
super-stable :)). I also trust my hardware very much; this server has 
been super stable for many months. SMART status of the disks is also OK.


Unfortunately, I don't use a debian-supplied kernel; I use a 
selfcompiled 2.6.13 with archck patchset (2.6.13-archck2). The rest of 
the system, including the ext2 tools, is a standard sarge system.


I am now switching the partition to ReiserFS 3.6 and will do further 
tests. In the unlikely case that there will still be problems when not 
using ext2, I will report back.


bye,
Tillmann


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]