We have seen the problem an AIX 3.2.4 as well (once).  We have a procedure
that allows us to get the filesystem on line and use cp to copy the data to
new volumes, then delete and rebuild the filesystem.

        # get the old superblock and replace the current one
        dd count=1 bs=4k skip=31 seek=1 if=/dev/lv?? of=/dev/lv??
        # run fsck to clean up the filesystem
        fsck -y /vicep?
        # find the hdisk number
        lslv -l lv??
        # find the group name
        lspv hdisk??
        #find the log disk for the group
        lsvg -l ?group?
        # format the log disk for the group
        logform loglv??
        # mount the file system
        mount /vicep?
        # run salvager by hand
        /usr/afs/bin/salvager /vicep? -force
        # restart the file server
        bos restart codesrv? fs

---------------------------------------------------------------------------
Ken Paquette                         IBM Internet: [EMAIL PROTECTED]
IBM Microelectronics                     Internet: [EMAIL PROTECTED]
Information Solutions & Integration       IBMMAIL: USIB1X62 at IBMMAIL
Distributed Computing Services             BITNET: KEN at VNET
D54/615-1                                    VNET: KEN at BTV
1000 River Road                             Voice: (802) 871-4014 (tie 435)
Essex Junction, VT.; 05452                    Fax: (802) 871-4253 (tie 435)
X.400  c=us; a=ibmx400; p=ibmmail; s=paquette; g=paquetk

------------------ Referenced E-Mail ----------------------------

Date:    Fri, 24 Sep 93 19:03:34 O

To:      [EMAIL PROTECTED]
From:    <[EMAIL PROTECTED]> (Michael Niksch)
Subject: Re: lost AFS partitions
Message-Id: <[EMAIL PROTECTED]>
In-Reply-To: [EMAIL PROTECTED]
"lost AFS partitions" (Sep 24, 10:28)
References: <[EMAIL PROTECTED]>
X-Face: (Oyg$iEzVArw$j8ML8^l(f55jC3xM#U1fZ[]!$t(dS0:/\AvEHX~4(tPe>"n=v>*u.Cq?)&
`%~i=+1{p~|j;,&6o3AF[n(ZsY.gmAl_HWb^;/:#F@ioA%#gc{X^?'k#4V0b#Mbs^hx(xH;'8rhc7|
`&)oAM(P0J~sDBnum&ayJ7P(dYFSEjxF,{;w}QN<|;:IRE%yE31"51y:-[(*}8KztQ[%jW:'UbG8o!
YGNO:%y2jwlRF4$aFLRP]7ovX{m@o%%8mTp23[ECI/qpyxH/%X5nC95`;Fo$YSBs]9B's
X-Mailer: Z-Mail (2.1.4 02apr93)
On Sep 24, 10:28, [EMAIL PROTECTED] wrote:
> There is a known AIX problem that could cause a server to lose a
> partition.
>-- End of excerpt from [EMAIL PROTECTED]

None of the suggested fixes ever helped. The problem eventually
disappeared when we reinstalled the machine with AIX 3.2.4. There was a
rumour from IBM Austin saying that the cause could be a corrupted log
device for a volume group. In fact, our machine had been reporting
errors with the configuration of its rootvg since we had replaced a disk
in the rootvg. If this was in fact the cause, it would mean that a
corrupted log device of one volume group could cause loss of an AFS
partition on a different volume group. I would highly recommend to
install AIX 3.2.4.


--
Michael Niksch                     TEL:               +41-1-7248-913
IBM Zurich Research Laboratory     FAX:               +41-1-7103608
Saeumerstrasse 4                   BITNET/VNET:       NIK at ZURICH
CH-8803 Rueschlikon / Switzerland  INTERNET/IBMIPNET: [EMAIL PROTECTED]

Reply via email to