Problems after moving Sophos DB to other cluster node

Question

Hi,

we had a little bit trouble because our database department had to do some maintenance last night on the SQL-Cluster running the Sophos database.

They switched the database to the other cluster node, so the database was a few minutes offline.

The Management Server lost connection to the database and didn't reconnect automatically.

The endpoints were up to date, but the Enterprise Console showed them "Not updated since..."

After a reboot of the primary Update-Manager and the Management Server I had to do a manual update and then all was fine again.

For future maintenance works of the SQL-Cluster I like to know what I have to do to avoid these trouble.

Here is a short overview of our environment:

- Database is on a dedicated SQL-Cluster

- Server A: management server + secondary Update-Manager

- Server B: primary Update-Manager

Is it enough to disable the management services on Server A before maintenance starts an enable them afterwards?

What about the endpoints that updated from Server B, this is the primary update-manager, during the database is offline?

Is this information cached somewhere and transferred to the database after reconnect?

Thanks in advance.

This thread was automatically locked due to age.

QC · Accepted Answer

Hello shawn_38,

it's the Management Service which communicates with the database. Where the clients update from doesn't matter - all status messages (including those from the primary SUM) are sent via RMS to the management server and if necessary stored when they can't be passed upstream. It takes some time for normal operation to resume and to process the backlog after the connection has been re-established.

Can't say how the management service deals (and should deal) with a connection loss (or switch) in a cluster environment though and why it apparently didn't recover in your situation.

Christian