QUESTION: is it correct that /sbin/advfs/salvage (du4.0d) does not salvage
device special files with correct major/minor device numbers?
Recently, the root filesystem (using Advfs) on an Alphastation 500
(DU4.0d, jumbopatch 3) was corrupted, probably due to SCSI termination
problems during the nightly defragcron run.
/sbin/advfs/verify run from the DU4.0d CD simply reported an inconsistency
that it could not handle. The usr and var domains were OK. Unfortunately,
I had no backup so I used /sbin/advfs/salvage to recover as much as
possible. Salvage ran with no error messages:
==========================================
fenchel:/var/fixit/> cat salvage.log
Recovery of fileset root, fileset quotas:
Hard block limit = 0, soft block limit = 0
Hard file limit = 0, soft file limit = 0
3763 files processed, including 117 directories
3763 successfully recovered, 0 partially recovered
0 recovered larger, 0 unrecovered
0 had no attributes.
fenchel:/var/fixit/>
==========================================
It turned out, however that while /vmunix etc. in fact was salvaged
correctly, device special files had wrong major/minor device numbers:
crw-rw-rw- 2 root system 1024,134688 Apr 4 15:44 ttyp1
crw-rw-rw- 1 root system 1024,133248 Apr 4 15:41 null
crw-rw-rw- 2 root system 1024,135008 Apr 3 15:16 ttyp2
crw-rw-rw- 2 root system 1024,138848 Mar 30 15:40 ttype
crw------- 1 root system 1024,146336 Feb 2 16:01 rrz0b
brw------- 1 root system 1024,146304 Feb 2 16:01 rz0b
and so on. (Unfortunately, it took us some time to realize that this was
the reason the machine wouldn't boot with the salvaged root domain).
To test if this was a problem with the salvage program and not a result
of the corruption, I created a new Advfs domain with just a device special
file on it and tried to slavage it. The result was like the above: wrong
device numbers.
yours
Niels Joergen Kokholm | email: kokholm_at_math.ku.dk
Institut for Matematiske Fag | phone: +45 3532 0759/+45 2128 6932
Universitetsparken 5 | fax: +45 3532 0704
DK-2100 Kobenhavn OE, Denmark | www:
http://www.math.ku.dk/~kokholm
Received on Thu Apr 08 1999 - 13:53:41 NZST