Thanks for the responses, (Alan, Alan and Bill)
As expected there were two opinions, what should happen and what does
happen.
The domain should panic on the first I/O error, and consequently become
inaccessible hence data cannot be lost/changed. When the disk is fixed
everything should work ok, although verify should be run.
As for what does happen. Generally the responses indicated that the above
does occur, although occasionally there can be problems resulting in the
need for a restore.
In our case everything was fine. The system was shutdown, battery replaced
in the HSZ40, rebooted and everything was visible and accessible.
Regards
Kevin Jones
> -----Original Message-----
> From: Kevin Jones
> Sent: Monday, May 10, 1999 12:08 PM
> To: 'alpha-osf-managers_at_ornl.gov'
> Subject: Disk failure in ADVFS domain
>
> Hello Managers
>
> We have an AdvFS domain consisting of LSM volumes on RAID 5 sets
> configured behind two HSZ40 's.
> One of the batteries in one of the HSZ40's has failed, hence not all
> volumes are available to the domain (Yes we would love redundant
> controllers). The domain is inaccessible to unix.
>
> My question is :-
> When the battery is replaced and all volumes become available again to the
> domain, should it all jump back into life, or will the data need to be
> restored to the domain.
>
> I suspect there may be two answers here, i.e. What is supposed to happen,
> and what people have experienced happening.
>
> OS is 4.0D patch kit 3.....
>
> Regards
>
> Kevin Jones
>
>
-----------------------------------------------------------------
Hamilton Rentals PLC does not accept legal responsibility for the contents of this message. The views or opinions presented are solely those of the author and do not necessarily represent those of Hamilton Rentals PLC.
This email was scanned for 41838 known viruses by kerberos.hamilton.co.uk. It was deemed clean for transmission.
-----------------------------------------------------------------
Received on Tue May 11 1999 - 11:42:34 NZST