Like I already said!!!

CTE, ROW_NUMBER(), CLR and Try...Catch *ARE* available even in 80 compatibility mode.

There are some DMV that are *NOT* usable in 80 but other than that language related...

On the Connect Site :

https://connect.microsoft.com/SQLServer/feedback/ViewFeedback.aspx?FeedbackID=175549

MS stated that this behaviour is *by design* and that setting a log file MAXSIZE to UNLIMITED will make it 2TB which is the Max Log...

>>.. it sure seems like it'll make a hotspot and it sure seems like you'd be in a world of hurt if any GetNextID attempt done in a long winded...

Sergiy

>>Why not use mirroring with multiple databases<<

It could potentially failover a single database out of a group and that is a problem!!!

*NOT* recomemeded.

Replication, for your information can run a...

Jeff,

The First problem with mirroring is that the destination DB is *Not* usable ( unless you create snapshots which is another kludge )

Mirroring *in my opinion* is a good choice...

apparently your update storedprocedure is not being called with the correct calling convention. I don't know exactly what the reason is... something/someone must have change the article definition etc... but...

try:

update d set intid= data.rnumb

(SELECT ROW_NUMBER() OVER (ORDER BY rno,rdata ASC) AS rnumb , cm, postamp

FROM po poo (nolock)

where poo.cm = 115 and poo.cmdesc like 'F40%') data

inner...

Your "guess" is wrong again! We are *not* using mirroring. If our Primary site goes down we move our clients to the replicas! Simple as that. Meanwhile we use replicas...

SSMS is a great example of "horrible bugs".

Well not really.

The compatibility level affects several "internal" features but not all SQL 2005 features are lost you can still use CTE row_number etc ...

If you need to know...

Is that number represented as a varchar ?

If yes then

update set col = '1' + col

where left(col,3) = '451'

if not then cast it:

update set col =...

Jeff, I am sorry if I missinterpreted your intentions.

There are many cases in which may things that seemed "normal" in non-replicated databases must be changed when replication is in...

David was spot on. When you fail over the software that you use must be able to function exactly as it was when connected to the primary.

The "not for replication"...

Once again IDENTITY columns *are* problematic in replicated environment with failover scenarios and there are *a lot* of tricks, kludges and workarounds to deal with the problem. Sequence TABLES is...

EXACTLY