Original Message:
Sent: Aug 13, 2024 06:40 PM
From: chulcher
Subject: Server in which the publisher is hosted maintenance question
Like I said, 2 hours isn't an issue. 24 hours is the threshold. If the subscriber doesn't sync with the publisher for 24 hours then the subscriber will not sync and will need to re-join the cluster once the publisher is available again.
------------------------------
Carson Hulcher, ACEX#110
Original Message:
Sent: Aug 13, 2024 06:30 PM
From: cdelarosa
Subject: Server in which the publisher is hosted maintenance question
Right now the stand by publisher is disabled
But the maintanance will be more than 2 hours.
If the maintenance is more than 2 hours and i have my stand by publisher OFF what will happen in this case? if my publisher is out of sync my subcriber wont be able to rejoin when i turn on again the publisher?
Original Message:
Sent: Aug 13, 2024 06:08 PM
From: chulcher
Subject: Server in which the publisher is hosted maintenance question
After 24 hours the publisher will be out of sync and the subscribers won't re-join. 2 hours isn't an issue. Make sure that standby publisher failover is not enabled.
------------------------------
Carson Hulcher, ACEX#110
Original Message:
Sent: Aug 13, 2024 05:00 PM
From: cdelarosa
Subject: Server in which the publisher is hosted maintenance question
Hello i have a question regarding this
We have 1 publisher central site and 2 subscriber in remote sites
The server in which the publisher is hosted, the server admins needs to do some maintenance to it so we need to turn off that publisher for several hours
in this case i understand that if it takes more than 2 hours it will not join automatically again to the cluster, and you need to rejoin it manually
we need that same server to be the publisher when it gets back
what would be the best practice to achieve this? what would be the steps ?
I guess i need to turn off the publisher but right now we don't have configured the automatic fail over to another subcriber.
I was wondering if its just leave the automatic fail over off and just rejoin the publisher to the cluster? of this is the wrong way to do it?
Any ideas?