RE: Taking care of Fail Back in HDR
Scenario: Old primary (dbsvr_a) failed, old secondary (dbsvr_b) was switched
to 'standard' mode, and traffic was moved over to dbsvr_b. Now need to get
dbsvr_a re-sync'd with dbsvr_b, and then get traffic moved back to dbsvr_a.
1) level-0 archive on dbsvr_b, transfer archive over and do physical restore
on dbsvr_a
2) start hdr replication *from* dbsvr_b *to* dbsvr_a, wait for "server
operational" msgs
3) disconnect all client connections from dbsvr_b, if possible
4) on dbsvr_b:
-- onmode -uy
-- onmode -l
-- onmode -l
-- onmode -c
-- onmode -d standard
-- onmode -ky
(note: make sure ckpt does *not* occur between last two commands)
5) on dbsvr_a:
-- onmode -ky
-- oninit -v
(wait for "Recovery Mode" in msg log)
-- onmode -d standard
(wait for "Quiescent Mode" in msg log)
-- onmode -m
6) allow client connections to dbsvr_a
7) on dbsvr_b:
-- oninit -PHY (brings up server just thru physical recovery, then
"pauses")
-- onmode -d secondary dbsvr_a
(note: this will *not* work if noted ckpt occurs above in step 4)
8) on dbsvr_a:
-- onmode -d primary dbsvr_b
9) wait for "server operational" msgs
HTH,
Paul Mosser
-----Original Message-----
From: Francisco Roldan [mailto:froldan@5b.com.gt]
Sent: Thursday, August 14, 2003 3:02 PM
To: informix-list@iiug.org
Subject: RE: Taking care of Fail Back in HDR
What about the Synchronization method after
recovering the master failure?
I too work with HDR .
I am in the process of making my recovering manual,
Your suggestions are welcome ...
thanks
sending to informix-list
sending to informix-list