AdvFS Domain Panic

Caspar Williams <[email protected]>
Newsgroups gmane.os.tru64.managers
Message-ID <[email protected]>
Dear Tru64 Managers,

An AlphaServer DS20E running Tru64 5.1A (never patched) recently 
experienced a Domain Panic on one of the AdvFS domains.

I don't normally take care of this machine, but was asked to assist about 
a week ago, which was some 10 days after the Domain Panic occurred - hence 
some vageries about the exact hardware details in the following sections.

>From what I understand the failed domain is half of a 1.5 TB RAID system; 
this domain was 96% full. I know no further technical details of this RAID 
system, not even what level of RAID it is. It does look like (I have been 
told) the other half is still operational and not experiencing problems 
(it is about 20% full). To clarify: the failed domain was dsk6h; the still 
operational other half is dsk6g.

We extracted more up-to-date versions of verify, fixfdmn and salvage from 
5.1A Patch Kit 6, as the originals pretty much keeled over straight away.

After this, verify still doesn't see filesets in the domain, and thus 
can't do anything.

Fixfdmn sees corruption in RBMT/BMT0 root tag mcell, but says it can't 
continue as it can't find a volume with a root tag file. The fixfdmn 
log-file mentions it found 1026 errors while validating tag page on volume 
1 LBN 96.

I've been told backups of this domain are inaccessible, unavailable or 
unusable.

Is there anything we could or should do, other than run salvage ? Given 
the amount of data to be recovered, and a lack of sufficient spare disk 
space, any in-place repair would be preferrable, if at all possible.

If salvage is the only option, I suppose we'll have to run it in piecemeal 
fashion, moving recovered data out of the way in between runs. Is there 
any other way to do this besides giving it (top-level) directory names to 
look for, one at a time (assuming we can remember these directory names) ?

Thanks in advance for any advice or recommendations you can give.
Cheers,

Caspar
--
Caspar Williams

e-mail: [email protected]         mail: Department of Mathematical 
phone (office): +1 780 492 8921                  and Statistical Sciences
phone (lab):    +1 780 492 1049               CAB 434a, University of Alberta
fax:            +1 780 492 6826               Edmonton, AB T6G 2G1
                                               Canada
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.