Hey Amit,
There's no need typically to shut down cluster nodes in any particular order, same with bringing them up again - unless of course a Connection Manager, configured as a failover arbitrator, is watching somewhere and might be tempted to draw the wrong conclusions ;-) In that case, make sure the Connection Manager is down first and up last.
There also is no need for a forced checkpoint (the shutdown will perform one anyway), nor for a log switch - or what exactly are we talking about?
Then, in a scenario where the whole cluster gets shut down only to be restarted again some time later, you'll never incurr a "DR: Start failure recovery from tape ..." situation; this really only can occur if one of the secondaries fell so far behind the primary that the next logical log to be received from primary isn't available any more on primary.
In such case, you'd simply prepare for and then start a logical log restore while the secondary is a secondary (don't do this before turning the secondary into a secondary).
HTH,
Andreas
------------------------------
Andreas Legner
------------------------------
Original Message:
Sent: Mon May 17, 2021 03:06 AM
From: AMIT PATEL
Subject: HDR shutdown and Restart
Dear Andreas,
This time I have to shutdown all 3 server PRIMARY , HDR 2ndary and RSS . so shutting all 3 servers sequence must be RSS --> HDR SECONDARY --> PRIMARY. And while up the instance sequence -- PRIMARY --> SECONDAY --RSS.
Am I correct?
Before starting Primary I will run "onmode -c" and "onmode -l" command to Force checkpoint and log switch.
But if ""DR: Start failure recovery from tape ..." error comes on SECONDARY server , then what action need to take to solve the issue?
Kindly revert.
Thanks
Amit Patel
------------------------------
AMIT PATEL
Original Message:
Sent: Wed April 14, 2021 01:03 PM
From: Andreas Legner
Subject: HDR shutdown and Restart
Not entirely true.
If your secondary server (HDR or RSS) was down for so long that the oldest log it requires to start resyncing with primary isn't available any more on the primary, or if the backup you used to instantiate the secondary using physcial restore was so old, or restore took so long, the secondary simply would say:
"DR: Start failure recovery from tape ..."
in which case you now could perform a logical restore using ontape or onbar, so all you'd need are the log backups available.
------------------------------
Andreas Legner
Original Message:
Sent: Wed April 14, 2021 12:43 PM
From: Kirit Rana
Subject: HDR shutdown and Restart
Amit
Another aside to bear in mind is that when you've patched the Primary node and brought it back online, don't leave it too long to patch the Secondary node and reestablish HDR connectivity. What you can not afford is for the Logical Logs on the Primary node to wrap around as that would mandate that you rebuild the Secondary from a fresh ontapeL0 of the Primary, or ifxclone if you prefer that method.
------------------------------
Kirit Rana
Original Message:
Sent: Wed April 07, 2021 06:54 AM
From: AMIT PATEL
Subject: HDR shutdown and Restart
Dear Art,
Yes , Currently DRAUTO set to 2 (Automatic).
Thanks for clarification.
Thanks
Amit
------------------------------
AMIT PATEL
Original Message:
Sent: Wed April 07, 2021 06:26 AM
From: Art Kagel
Subject: HDR shutdown and Restart
Amit:
No, there is no need to run onmode -d again just to restart either the primary or the secondary. You just restart both. If you have auto-upgrade set in the ONCONFIG file or through the Connection Manager, you will want to start the primary first so that the secondary doesn't upgrade itself or be upgraded to primary mode by the CM, otherwise it's just a restart.
Art
------------------------------
Art S. Kagel, President and Principal Consultant
ASK Database Management Corp.
www.askdbmgt.com
Original Message:
Sent: Wed April 07, 2021 05:20 AM
From: AMIT PATEL
Subject: HDR shutdown and Restart
Dears,
Please let me know, if have HDR setup for PRIMARY AND SECONDARY and if we have to shutdown for patching activity both server then can we directly shutdown (onmode -ky) and then restart (oninit -vy) ?
In this case if we start both servers directly thru oninit -vy then will they automatically start as PRIMARY and Secondary?
OR
again when starting PRIMARY then also need to run onmode -d PRIMARY <sec_server> and for Secondary onmode -d SECONDARY <prim_server>.
kindly suggests.
Thanks
Amit
------------------------------
AMIT PATEL
------------------------------
#Informix