Here are my posts on the MSDN blog SQLServerFAQ for the month of October.
A guide to troubleshooting tempdb growth due to Version Store usage:
Troubleshooting tempdb growth due to Version Store usage
An email chain with Jonathan (Blog | Twitter) resulted in this blog post:
Troubleshooting specific Login Failed error messages
A much needed VDI troubleshooting tool called SQL Server Backup Simulator:
September has been a content month for me both on the SQLServerFAQ blog on MSDN as well on this blog. Here is a round up of my posts on SQLServerFAQ for September:
I was intrigued by this when someone asked me what the Scan Count value meant w.r.t. SEEKs/SCANs:
SCAN COUNT meaning in SET STATISTICS IO output
I had worked on multiple issues for SQL Server 2005 regarding 511. Wanted to demonstrate how easy it is to troubleshoot this on SQL Server 2008 with XEvents:
Troubleshooting Error 511 using XEvents
Ran into a Compile Lock blocking issue because the Stored Procedure being used had an Open Symmetric Key command in it. Here is the post talking about the issue and how to workaround it:
Open Symmetric Key command prevents plan caching
I ended up doing a sequel for the Scan Count values w.r.t. Joins in the post below:
Scan Count meaning in SET STATISTICS IO output: Part 2
I was working on a root cause analysis for an OOM (Out-of-Memory) issue for SQL Server 2008 R2 and I needed to schedule notifications when the available memory on the server fell below a certain value. SQL Server 2008 and above has a nifty little DMV to do just this. Read about it’s usefulness here:
I recently worked on a soft-NUMA configuration issue for SQL Server 2008. I decided to document the steps to configure soft-NUMA in a blog post on SQLServerFAQ on MSDN. Here it is:
How to set Soft-NUMA for SQL Server 2008 R2
Happy NUMA configuration! 🙂 It can be quite daunting if you have multiple CPUs.
This has been an age long debate on whether to use GUIDs or INTs. Thought I would pen down some points based on some tests that I ran on my test machine.
SQL Backup softwares can cause Out Of Memory conditions if the BufferCount and MaxTransferSize is not taken into account on 32-bit systems:
Incorrect BufferCount data transfer option can lead to OOM condition
In the recent past, I worked on a few issues where the use of monotonically increasing clustered index keys were turning into hotspots in the database leading to latch blocking and causing performance degradation as an end result. I illustrate this with the help of an example in the blog post below.
Monotonically increasing clustered index keys can cause LATCH contention
