I've seen a lot of thoughts on what may have caused this problem, but I don't
remember seeing any ideas concerning how to fix it.
Since the DB is mirrored shouldn't you be able to take the failing DB volume
offline (leaving the mirror copy that is good online), reformat it and then put
it back online and the DB will resync with the mirror copy? Of course you would
likely want to do it during a period of inactivity where you could take a full
DB backup immediately beforehand.
Steve Branch
Phillips Petroleum
e-mail: swbranc AT ppco DOT com
"Mauro M. TINELLI" <Mauro.TINELLI AT ST DOT COM> on 08/20/99 01:53:13 AM
Any replies will be addressed to: "ADSM: Dist Stor Manager"
<ADSM-L AT VM.MARIST DOT EDU>
To: ADSM-L AT VM.MARIST DOT EDU
cc:
(bcc: S W Branch/Sweeny/Phillips Petroleum/us)
Subject: Re: Page address mismatch ANR0207E
Roger,
"bad block susbtitution" may replace the block but if the error was on a
disk-READ operation, I wouldn't bet on the its content being the original
one, but then the mirror should have done the rest, shouldn't it?
Mauro, STM
> Alas, this version of the server does not have an AUDIT DB command.
Or is ther
> e another
> command I am missing?
>
> Having a defect in the HD sounds plausible, but if you understand HPFS,
it can h
> andle certain
> faults, mark the blocks bad, and continue on.
>
> Any other suggestions are welcome!
>
> Roger
>
> On Wed, 18 Aug 1999 11:07:10 -0500, Wu, Jie wrote:
>
> >Mauro,
> >
> >well, like I said, it is solely based on my personal experience. It is
lucky
> >that Roger mirrored the database, otherwise, the server was going to
crash
> >when the database page mismatch error was encontered. As pointed out
from
> >Roger's email, the server looked for a morror copy from
E:\DBVOL\DB4.DSM
> >while a page mismache happened on D:\ADSMV2\DB1.DSM. I do not know how
much
> >database knowledge you have, but anyway, one really need some database
> >knowledge in order to understand a page mismatch error. (logical page
-->
> >physical page).
> >physical page).
> >
> >Actually you would find that several people have reported the same
error
> >before if you search this archive. It was suggest that bad hard disk
might
> >be the problem. However, other possibilities exist. For example, a
corrupted
> >database. I remember an adsm developer from IBM suggested auditing the
> >database. I believe that IBM has no sure answer to this problem and is
still
> >dealing with it.
> >
> >I would like to repeat myself, the above is only my understanding. I
might
> >be wrong.
> >
> >Jie Wu
> >System and Storage Administrator
> >ROW Sciences Inc @NCTR FDA
> >phone: 870-543-7958
> >fax: 870-543-7382
> >
> >
> >> -----Original Message-----
> >> From: Mauro M. TINELLI [SMTP:Mauro.TINELLI AT ST DOT COM]
> >> Sent: 1999
|
