Please help me understand the below scenario:
My TempDB drive is 100 GB
I have split and assigned the TempDB into multiple files based on below point:
I have 24 core logical processor , so I have split the TempDB into maximum 8 data files (Tempdev .... Tempdev8) and 1 Templog file. I assigned maximum 9 GB to each data file with 512 MB of Autogrowth. For log file 13 GB with 256 GB of Autogrowth. Leaving enough buffer space for Autogrowth.
This is my TempDB current State.
TempDB properties
SP_HELPDB TempDB output
TempDB current utilization
**** My Questions ****
- Is my TempDB configuration correct ? Did I do it right way after going through all advices ?
- If a new transaction comes , will it use the already assigned size for the TempDB data files as there is lot of unused space internally or will it go for the buffer disk space which is the unassigned disk space , which is 15 GB left over out of 100 GB.
- Should I assign the least initial size or my initial setting is ok ? Is it a mistake to assign more initial size , please guide me more on this point , how it works if we assign more initial size and less initial size ? whats difference ? It is really confusing for me , please advice and help me to get rid of confusions.
- What are the disadvantages / fault of my configuration ?
- Will SQL Server pre-assign required space for a transaction based on the activity , for ex: my ETL is taking up more space and drive is going into RED and then not releasing space post transaction is completed , I need to manually shrink the tempDB data files . please highlight some inputs on this.
- With my current settings should I opt for enabling trace flag 1117 , please advice.
Best Regards,SQLBoy