From: Robert.Rakowski@hboc.com
Date: Thu Sep 05 2002 - 11:24:53 EDT
Long story short -- my backup administrator said "oops" when I asked for
restores to a rebuilt server and the metadata (LSM private region) on
the array is gone so I'm trying to rebuild it by hand with some signs of
success...
Short story long -- the server is an ES40 running Sybase that was
connected to a BA370 array via an HSZ70 (shows up as 2 disks, rz8 and
rzb8). The BA370 was physically removed from the server and replaced
with an SW4354(?) shelf with 14-72G uni-disks connected via a KZPCC 3
channel controller, after which the OS was rebuilt. Once the Sybase
database was rebuilt and ready to be restored, my backup administrator
changed his mind on his earlier statement that the backups had been
successful prior to the rebuild. Oops. The BA370 has been reconnected
to the ES40 and I am trying to rebuild the metadata on the array using
an old volprint output.
On the bright side, the data (LSM public region) is apparently still
there, I have an old volprint output, this is a server that hasn't
changed much at all since the volprint other than an upgrade from 4.0f
to 4.0g, the array was broken into only 2 RAID5 raidsets (rz8 and rzb8
to the OS...). Even better, according to volprint only rz8 is used. I
also have df -k output from just prior to the rebuild that verifies what
I see in the old volprint output.
I have been able to recreate the first two volumes (FS swap and a normal
FS) by hand and I see the data in the second volume, but the third
volume (of 70...), which is also a normal FS, is giving me problems.
When I run a showfsets against the domain of the third volume I get:
showfsets: can't show set info for domain 'pdscustom_domain'
showfsets: error = E_BAD_MAGIC (-1167)
Compaq tells me this is "probably from wrong offset", but the math looks
good to me, especially since the first 2 volumes appear to have been
successfully rebuilt. I used an offset of 0 for rz8-01 (swap
volume...), and 2125824 (the size of -01) for rz8-02. For -03, the one
that gives me the error, I used the sum of the sizes of -01 and -02.
Am I missing something here? Any help would be greatly appreciated.
Rob
Replyto: Robert.Rakowski@hboc.com
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 10:48:51 EDT