RAIDframe troubles

[Saad Kadhi <saad@docisland.org>]

Hi there,

I have an OPENBSD 3.2-stable system with 4 IDE  drives  (wd[0123]).  wd0
and wd2 are used as components of two RAIDframe RAID devices (raid0  and
raid1). wd1 and wd3 are used as components  of  another  RAIDframe  RAID
device (raid2).

All was running fine until I experienced a power outage  two  days  ago.
When  /etc/rc  called  'raidctl  -P  all',  the  parity  re-writing  was
successful on raid0 and raid1 but it failed on raid2 with the  following
message:
  
  [...]/bsd: raid2: Error re-writing parity!

At the prompt, executing 'raidctl -P raid2' return this:
  
  
  # raidctl -P raid2
  raid2: Parity status: DIRTY
  raid2: Initiating re-write of parity
  # raidctl -Pv raid2
  raid2: Parity status: DIRTY
  raid2: Initiating re-write of parity
  Parity Re-write status:

  raid2: Parity Re-write complete

but the '/bsd: raid2: Error re-writing parity!'  message  is  logged  by
syslog and at the console, the following message appears (in blue bg):
  
  raid2: Error re-writing parity!

And 'raidctl -[pP]' still shows the parity status as DIRTY. However  the
file systems created on raid2 can be mounted and used.

Reading the raidctl man page and the list archives didn't  lead  to  any
way to solve this situation save for a message from P T Washington dated
Nov 27, 1999:
  <quote>
  I'm still confused about the use of raidctl -i. On previous advice,  I
  upgraded to 2.6 (October snapshot). The man page for raidctl says that
  raidctl -i must be used when initializing  an  array  _and_  after  an
  unclean shutdown, before fsck.
  </quote>

Would 'raidctl -i' solve my problem _without_ destructing existing  data
on the raid device ?

TIA.
-- 
Saad Kadhi -- [saad@docisland.org] [saad.kadhi@hapsis.fr]
[pgp keyid: 35592A6D http://pgp.mit.edu]
[pgp fingerprint: BF7D D73E 1FCF 4B4F AF63  65EB 34F1 DBBF 3559 2A6D]
---