March 9, 2017 at 9:17 am
Perry Whittle - Tuesday, March 7, 2017 12:17 PMszejiekoh - Saturday, March 4, 2017 11:02 AMJust some last confirmation on patching and upgrading ->
q1) when we patched the secondary replica, beside the binaries getting updated, does its databases get "updated" as well only when the AG failover to it ? or the databases get updated as soon as the binaries are patchedif i remember correctly only when they go through recovery.
szejiekoh - Saturday, March 4, 2017 11:02 AM
@@version will return the binary version
is there any command to check "database" version ?Not while the database is not recovered.
szejiekoh - Saturday, March 4, 2017 11:02 AM
q2) reason for asking the above is that I have DR instance/replica which we will seldom failover to it, but we will still patched it anyway.I am worried that if we just patched it without failing over to it, will its "databases" get updated as well ? (and if it doesn't, will we have trouble failing over to it as over time more and more patches are being patched .
Regards,
NoobWhen they go through recovery which would usually be fairly quick
Hi Perry,
Thank you for following this through!.
But when does "recovery" actually happen ? Is a read only secondary replica syncing from the primary database doing "recovery" all the while ? or it is only doing recovery when it become the primary ?
My main concern is just that if i don't failover to the DR at all on every patching, will I have problem failing over to it after "n" patches ? Do i need to failover to it to "upgrade" the "database" ?
Regards,
Noob
March 18, 2017 at 10:57 pm
Hi all,
Sorry for the slow update. I just come back to post this in case anyone encounter the same problem like me.
Like what GilaMonster said,
1) patching an SQL server with a CU does not upgrade the database version.
=====
However, I still do not know if i will ever encounter a issue if i am patching a DR(secondary node) but have never failover to it during the patching.
E.G. 3 node setup ( 1 node in DR site)
T1) patch DR node (node3)
T2) patch Local Secondary node (node2)
T3) failover to Local Secondary node (node2 become primary, node1 become secondary)
T4) patch previous primary node (node1)
T5) failback to the previous primary node. (node1)
-- never failover to DR node (node3) at all during the whole patching activity. Not sure if i will encounter issue after N patches and then come a need to failover to DR.
Anyone ever does a failover to their DR site during patching (despite the fact that they can failover to a local secondary node) ?
Regards,
Noob
Viewing 2 posts - 16 through 16 (of 16 total)
You must be logged in to reply to this topic. Login to reply