I install SQL 2012 in cluster and cofigure tempdb: 1 file per core, initial size 1024Mb, but after SQL automatic change initial size to different value. Why SQL automatic change initial size for tempdb? How I can disable this options?

Hello,

When does SQL Server changes the "initial size", after restart of the server/service or during runtime?

After cluster group move to other server

When you set an initial size for TempDB, when SQL start's up, it will revert to the initial size that you provided.

I'm guessing that when SQL failed over, the disk containing TempDB did not failover to the other node and TempDB was recreated using default settings provided by the Model database.

default settings: 1 tempdb file, but sql create all files (60) and change initial size to more

What high availability are you using, Mirroring/AlwaysOn/Log Shipping ?

Windows cluster

Is the TempDB disk being failed over when the instance fails over to the other node?

Memory error

Memory error

??

Tempdb disk work without error and disk success move to other servers in cluster.

Just so I have this right:

You have 2 nodes, SQL has been installed as clustered on these nodes.

When you failover, SQL comes up and TempDB size changes?

Please confirm that you are not using 2 separate SQL instances

Yes, 3 nodes, SQL clustered installed.

No separete SQL instances.

There is no way that TempDB will take on different settings when SQL is failed over. It's one instance of SQL server

How I can diagnostic why initial size changed? Maybe audit?

Hallo inevg,

with respect - just a few phrases googled and found hundreds of solutions for tracking growth of tempdb :)

http://lmgtfy.com/?q=trace+growth+of+tempdb+with+extended+events

thank you, how can I determine what has changed from the initial size of the database tempdb?

Hello the change is you are seeing due to the tempdb expanding for the transactions or for any other activity that runs against on the tempdb.

so look out under the hood.

Can I get a log in the form of "system" user changed the initial size of the database or the "administrator" changed the initial size?

you can use the extended events

https://www.simple-talk.com/sql/database-administration/optimizing-tempdb-configuration-with-sql-server-2012-extended-events/

or you can schedule an job for  sql profiler server side trace but extended events is new form of profilerbut Iam still learning that extended events.

did you see anything from the default trace for temdpb.

There is no easier way to audit changes to system configuration SQL server? For example, such as an audit file system, logon events, changes to objects in Active Directory, and so on?