Updated all nodes

[moon-buddy]

By mistake I updated all nodes and pointed all to one domain.

Easily fixed, I thought. Restored to the last DB backup and when I looked into it, the nodes are still pointing to the single domain!

Running 7.1.7.1 on Redhat.

Am I wrong on concluding that the domain should go back to what it was used to be? If not, this blows away all of my 'understanding of how things work'! A bug?

 

[marclant]

Probably did a most recent state restore. The logs rollforward and all the transactions after the DB backup got replayed. You'd have to do a PIT restore to avoid rolling forward the logs.

 

[moon-buddy]

I deleted the TSM DB and did just a plain restore db. I should have done a point in time?

 

[marclant]

If the archive log directory is there and contains the logs since the last backup, it will roll them forward. That's what's called "restore to must current state". Need a PIT to restore to a point in time, in case, that time is your last DB backup. You can do PITD=now and it will take the most recent backup and will NOT replay the logs.

 

[moon-buddy]

Doing restore db todate=11/09/2017 totime=06:00:00 and should work

 

[marclant]

Assuming your DB backup started before 6AM. I'd put the time as now just to make sure you get the latest, and not the one prior. It will take the most recent backup that occurred before the date/time you provide.

 

[moon-buddy]

Assuming your DB backup started before 6AM. I'd put the time as now just to make sure you get the latest, and not the one prior. It will take the most recent backup that occurred before the date/time you provide.

All good now - I know what went wrong.

I never use PIT but I do delete the TSM DB and clear the LOGS. I use a script to do this and there was a misstep on the script - it did not clear the LOGS.