From: Dave Warchol (Warchol@harthosp.org)
Date: Thu May 22 2003 - 08:45:14 EDT
Hello Managers:
Two sets of LUNs defined on a CLARiiON. On any given day, the
database on a "reporting" host is using one of the sets. On a daily
basis, the set that is not being used is refreshed from a production
database and the reporting database switched to point at set that has
been refreshed. Production database is mirrored across primary and
disaster recovery CX600's. SnapShot is performed off mirror (production
DB quiesced first), SanCopy is used to copy from the snapshot LUNs to
the set of reporting LUNs that are being refreshed. The entire process
is scripted (Korn Shell), the "flip-flop" piece works as follows:
- reporting database is brought down
- disk group is deported (vxdg deport <dgname>)
- private regions on the disks that we won't be using are initialized
(vxdisk -f init <each disk>)
- disk group is imported (vxdg -C import <dgname>)
- volumes are started (vxvol -g <dgname> startall)
- reporting database is brought up
For each Veritas command line step, I check the return code from the
command. Following the "startall" step, I parse through output from
"vxprint -g <dgname>" and verify that the volumes, plexes and subdisks
are all "ENABLED" and "ACTIVE" (except for sd's, where it should be
"ENABLED and '-'), prior to bringing the database back up. I'd be very
interested in what else (if anything) that I could check via Veritas
commands for the other steps (deport, init, import) that would help me
verify that things were as they needed to be following those steps. I
will summarize.
Thanks very much,
Dave
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:26:27 EDT