VC upgrade from 2.0.2 to 2.5 - Database Update Failed

[J_Wood]

I have tried to upgrade twice now with the same result. I get a message that the database upgrade has failed. The database upgrade log file has this line:

The specified @job_name ('Past Week stats rollup') does not exist.

The database rollback also fails.

The account used for the upgrade has the necessary permissions. The database is in bulk-logged recovery mode.

Suggestions? Please...

[scerazy]

Bit more details please, what was the original DB format?

In place upgrade on same server or remote?

Did you have a backup (or two)

Seb

[Dave_Mishchenko]

Are you using SQL 2000 or 2005 and which edition (MSDE / Express / Standard / Enterprise)?

[J_Wood]

This is an in-place upgrade on the same server going from VC 2.0.2 Update 2 to VC 2.5. It is SQL 2000 Standard Edition SP4 on a remote SQL server. I did take a fresh DB backup before I started the upgrade so I was able to get a functioning VC server back fairly quickly.

The VC Server hardware is an HP DL380 G4, Dual HyperThreaded 3.0 CPU's and 4 GB RAM. OS is Windows Server 2003 Standard, SP2.

[asutosh]

Can you try deleting the Jobs in your data base before upgrade?

[RParker]

I had this problem, until I realized how long it would actually take to simply rebuild the cluster....

It doesn't really take that long, unless you have a zillion permissions then it might be a real pain. If you don't, you can keep your OLD database, and create a new one, and just switch between them.

We have a VM for the VC, so if you clone the VC to a new, reconfigure to point to the new database, you can keep both running at the same time, and just look at the different settings and configurations.

If this is on SQL, it's even easier, SQL is very fast, and you can be up and running in a half an hour. The database upgrade takes what, 45 minutes? At least for us it did, I don't know about you, but 90 minutes of my time just isn't worth it, besides, starting fresh isn't all bad....

Food for thought ;0

[J_Wood]

Yes, it was less time to rebuild the two clusters I had with a fresh install then spend amy more time troubleshooting the failed install.

Thanks anyway to all.

[bedobash]

I wasted many hours getting the DB upgrade to work.

And once the DB was upgraded, I STILL had to recreate the clusters and reconnect to the hosts. I should have just created a new DB considering I had to recreated everything anyways.

[kimono]

This had me going for a while, until I saw this conundrum of a statement buried in the upgrade guide:

"After a successful database upgrade, the system tries to start the vpxd service. If the service times out because the database is large, you see an error message noting that the upgrade has failed. The upgrade has not failed, however. The installer timed out waiting for the service to start."

http://pubs.vmware.com/vi35/upgrade/wwhelp/wwhimpl/common/html/wwhelp.htm?context=upgrade&file=upgra...

Classic error, up there with "the object does not exist, or has never existed"

/kimono/