Installation procedure on second cluster node

Hi,

One of my clients has been trying to implement a bad SQL SP3A installation procedure to upgrade version on a two active passive node cluster.  As expected, only the active node has been upgraded while the passive node remained in initial version.  Did anybody meet this kind of situation before?  Is there a way to "reallign" the nodes in term of versions without having to break up the whole thing and start all over again...  Thank you for the input and experience...

Racimo
LVL 23
Racim BOUDJAKDJIDatabase Architect - Dba - Data ScientistAsked:
Who is Participating?
 
arbertCommented:

"Sp_db_upgrade"


Hmm, the way it sounds, you're way beyond this point.  I can only imagine how screwed up the dependencies and resources are.....

"All I know full-text index format is upgraded by SP3."  There were many more additions/changes to the system tables/procs than that....


I think your assumption is correct above in that you need to rebuild that cluster.
0
 
arbertCommented:
What version of Windows 2k or 2003?
0
 
Racim BOUDJAKDJIDatabase Architect - Dba - Data ScientistAuthor Commented:
Hi arbert,  What's up...

the system is on Win2K

Racimo
0
Cloud Class® Course: C++ 11 Fundamentals

This course will introduce you to C++ 11 and teach you about syntax fundamentals.

 
arbertCommented:
"As expected, only the active node has been upgraded while the passive node remained in initial version"

Sure you already saw this:  http://support.microsoft.com/?kbid=811168

How did they initiate the install?  I take it the failed over and attempted to install on the passive node as well?
0
 
ispalenyCommented:
Clustered configuration is just a theory for me.
When database formats are not upgraded, you can turn off an active node and upgrade the new active node. If they are not, it can be a real problem.
0
 
Racim BOUDJAKDJIDatabase Architect - Dba - Data ScientistAuthor Commented:
arbert,

<<How did they initiate the install?>> in the worst possible manner as they have implemented a 5 year old procedure working for SQL7 ;(.  They counted on the MSInstaller to shutdown the entire system and services before updating the libraries but as you know, it's not reliable when platform is not perfect.  To my knowledge the components should be updated on all nodes from one install.  So as a result, the passive node has not been updated.  Additionally  (god these guys are good) they simultaneously tried to implement both SP3A and 3 hotfixes (versionning troubleshooting is a mess)

<<I take it the failed over and attempted to install on the passive node as well?>> Not even that as they relied on a totally obsolete procedure described above...(and I told you only the most competent part).  They did not even know that the passive node was not updated until I pointed it out.  I just want to see if there's a getaround of breaking up the whole damn thing and start all over again...

ispaleny,
<<When database formats are not upgraded>>  What do you mean by upgraded format?
0
 
ispalenyCommented:
Sp_db_upgrade is a procedure upgrading physical database format to the latest format. I don't know if there are multiple version of format for SQL Server 2000. All I know full-text index format is upgraded by SP3. I also expect, that SQL Server 2000 is able to read any SQL Server 2000 format, because of full detach-attach database procedure support.
0
 
Racim BOUDJAKDJIDatabase Architect - Dba - Data ScientistAuthor Commented:
<<Sp_db_upgrade is a procedure upgrading physical database format to the latest format. I don't know if there are multiple version of format for SQL Server 2000. All I know full-text index format is upgraded by SP3.>>
SP3 not only does updates on files stored in the database but also on some binaries and components foe the overall systel to work but thanks for  your help.  matter here is above DBA approach if single detach attach procedure or code TSQL specific procedure ;)
0
 
Racim BOUDJAKDJIDatabase Architect - Dba - Data ScientistAuthor Commented:
<<I think your assumption is correct above in that you need to rebuild that cluster.>>That's what I was afraid of (I wished in secret there would be a tool around to fix this back or at least get back to some previous versionning).  The client exploiting team is really  bunch of unconpetent db killers !!! (They already lost about 200Mb almost critical data) (:((

Anyway I'll leave this thread for additional suggestion until next week then close it....Thank arbert :)

Racimo

0
 
arbertCommented:
hmmm, sounds like we're working with the same team :)  I just had a client lose 400gb because of poor backups and some how the blame came back on me--haven't quite figured that out!
0
 
Racim BOUDJAKDJIDatabase Architect - Dba - Data ScientistAuthor Commented:
What I have not figured out yet is this is one major company with a bunch of money....still they can't even hire a dba :)))
0
 
arbertCommented:
"still they can't even hire a dba :)))"

That's funny.  It's amazing how companies don't seem to think their data is an important asset.  Too many more incidents like this, and they'll probably get a clue huh?
0
 
Racim BOUDJAKDJIDatabase Architect - Dba - Data ScientistAuthor Commented:
yep some heads have already began to spin (not mine fortunately)  LOL
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.