Bad AdvFS magic number
Didier Godefroy <[email protected]> Mon, 20 Dec 2010 05:43:12 +0100
| Newsgroups | gmane.os.tru64.managers |
|---|---|
| Message-ID | <C9349D70.343D6%[email protected]> |
Hi all once more, I still haven't got any hints on how to fix my big problem with advfs corruption and lost lsm configuration. I've been working on this nonstop trying to figure this out now for more than 24hours straight. I need something to get me passed that hurdle. In my previous posts from the past 24 hours or so, I explained what happened I think in details and that I was trying to find a way to "patch up" a corrupted advfs filesystem. In fact, this posting can somewhat be like a partial summary, although I didn't get any definite fix and no hints as to how to get this done right, I was able after a lot of tries to fix up (so it seems so far) the corrupted root domain. Don't ask me how this got done, I didn't use anything I didn't use before and that failed at first. I'm working on a duplicate copy of the drive that has two advfs domains corrupted. Those domains are root and var. On that same drive there are also swap and usr (dskxa=root,dskxb=swap,dskxd=usr,dskxf=var), the dskxe partition is just a small one that's for lsm to put one more config copy (dskx as dsk...whatever). At first I couldn't even mount the root domain, then it mounted, after a few tries and caused advfs panics. The utilities like verify and fixfdmn didn't work at all and caused a panic rendering the domain inaccessible. I haven't done anything else special, no other commands to try, as I didn't get any help on that, but finally at some point by using fixfdmn with one -m flag at a time, I was able to get it to work on it and fix corruption, so after that verify ran fine without errors on it. It looks like all the domain's files contents are there, so it might be bootable, but I haven't tried this yet because there is still var to fix. The usr domain passes verify and mounts fine. Both verify and fixfdmn don't want to get passed a "Bad AdvFS magic number" or "invalid volume number", I even tried nvlogpg with the same errors. I have the impression that if I could just get passed that wrong id number, things would start getting fixed. But how to I fix this bad number???? I looked up the undocumented commands, but they're not really well documented of course and they're mostly info gathering tools and don't fix anything. Can someone point me to a possible solution please? I'm about to drop after lack of sleep and the machine has been down for more than a day now. Since I can't get my emails from that machine any more, please reply to my address below if you can help. Thanks all, -- Didier Godefroy mailto:[email protected]