MASTER database unrecoverable - Script Upgrade Mode

  • Fun! IT was applying a Windows upgrade to clustered SQL 2008 R2 environment when script error occurred (Event Viewer) apparently during changes to MASTER database. Now SQL is down, and when I bring it up in single user mode, I am unable to connect and get message about system being in "script upgrade mode" and only allowing administrator to connect. Tried logging on as "sa" with same results.

    I was planning on connecting and running RESTORE DATABASE from our MASTER full backup taken earlier today.

    System has been down over an hour, and is distributor for all our replication (among other things). Contacting Microsoft support, but welcome any insight anyone may have.

    Have Fun!

  • Do the error logs indicate which script was causing the error?  Because those scripts are all available to run in the SQLServer\MSSQL\Install folder.

    "But!" I hear you cry, "I can't even launch SQL Server"

    True, but you can launch it from the command line with Trace Flag 902, which will allow for some limited functionality - enough to run the erring script, anyway.

    Give it a go, see what happens.

    Thomas Rushton

  • Thanks! Will check and update on progress. So far I know Microsoft said "troll the internet" because SQL 2008 R2 is no longer supported. We may end up just re-installing, but your suggestion gives another possible avenue. Thanks again for the suggestion.

    Have Fun!

  • The question becomes what Windows patch were they applying that was running a SQL Server upgrade? A SQL Server security patch in a Cumulative Update?....

    As far as I know SQL Server 2008 is still under Microsoft support until July. odd that Microsoft told you to troll the Internet for assistance.

  • Yup, I wasn't on the call (IT handled it) but the support response may be because the instance is still on SP1, not SP3 (just guessing on reason).

    In any case, the on-site DBA was finally able to restore the Master, supposedly after dealing with a compressed drive (not sure what genius compressed the drive with the databases on it) and cluster issues. Good news is system is up, and we were able to recover all the databases. Got rid of existing replication setup and ran scripted setup successfully. Most things are working fine now, although some residual issues with a remote subscriber not getting added correctly (add seems fine, but subscriber has no record and remote distributor agent job does not list subscriber). Now that I "say this out loud", I'm thinking of more things I can check.

    Thanks for suggestions, it is always nice to know there is someone out there. 🙂

    Have Fun!

Viewing 5 posts - 1 through 4 (of 4 total)

You must be logged in to reply to this topic. Login to reply