Don't use task manager to check SQL Server's memory. It lies (and I've posted this soooo many times recently)

Use perfmon, use the target and total server memory counters.

You're not going to see a difference in inserts or restores. See the article I referenced where Tony and I tested the difference between about 20 VLFs and 16000 VLFs...

Can't make specific, useful suggestions without seeing the definitions of FTcomparison and returnFlagForTemp, but in general you need to rip those cursors out, convert them and the function calls to...

The QO will chose the smallest index, so if there's a nonclustered index that's 1 byte smaller than the clustered, that'll be the one chosen. Whether it makes sense to...

You're not going to see much of a difference with a hundred or so VLFs. Try getting up to tens of thousands.

You see the effect mostly on recovery, when starting...

Sure. Run update statements to make the required data changes.

WhiteLotus (10/7/2015)

---

I assume the yellow highlighted is the Victim and the Red one is the culprit ?? But the information that I can read is the object name , No...

Can you post create table statements for all involved tables please?

Without seeing the code, impossible to give any specific suggestions.

In general, you probably need to:

Replace the cursors with set-based queries (and no, don't just replace with a while loop). This...

nm, ignore

You need SP 4 for an in-place upgrade.

Please can you get the deadlock graph and post it? Can't diagnose a deadlock on vague descriptions of processes.

On 2008?

Yes. The number in brackets is the number of characters that the column can hold. NVarchar(10) can hold 10 characters, "SQL Server" is 10 characters, so it'll fit.

Please post the deadlock graph and details of the two queries involved.