UPDATE STATISTICS GETTING SLOWER

Question: The Statistics tasks in the SQL Server Maintenance plan is getting slower? When I first implemented the task it used to take 1 hr – recently it’s built up 4-5 hrs. I’ve checked underlying resource issues – and everything looks OK. Could you advise on how I can progress with troubleshooting. I’ve included a screeshot of Update Statistics Task. Answer: It is not unusual to find UPDATE STATISTICS getting slower over time. Looking at the way you have this configured All existing statistics – this includes all index and table statistics Sample size of 25 % It is very... Read more →

SQL Server , Meltdown and Spectre FAQ

I'm getting various questions about the Spectre and Meltdown impact on SQL Server. I'll start listing the most popular questions and the response. There is a lot of activity in progress across all the communities about mitigating the impact of Meltdown and Spectre . Typically there will be patching by different vendors and this needs to be managed properly. Where is the Microsoft SQL Server guideance documentation for Meltdown and Spectre ? The main link I reference is Protect SQL Server from attacks on Spectre and Meltdown side-channel vulnerabilities . The main advice is start planning for patching. There is... Read more →

Wait Stats

Backupthread,latch_ex,pageiolatch,cxpacket ,lck_m_is,logmgr Read more →

Type Conversion may affect cardinality estimate in query plan

Type Conversion may affect cardinality estimate in query plan is a warning found on Execution Plans. This may negatively impact the correct index usage. Cardinality is a critical part of how indexes are used. Data type matching has many benefits. It’s a practise often neglected in database design but which can have a number of long term consequences for query access and database performance One long term cost of neglecting data type matching is it can lead to more memory required. When data types are matched , this leads to less bytes requires , leading to less data pages in... Read more →

Chinese build the most powerful computer in the world

Cyber superstardom and the great super power race is on! 93 petaflop Sunway TaihuLight 93,000 trillion calculations per second. 1.31 PB of primary memory Advanced manufacturing, weather forecasting ,Life Sciences & big data analytics 10.5 million locally-made processing cores and 40,960 nodes , running on a Linux-based operating system. As a computer scientist , this must be an amazing challenge to write software which can exploit the capacity. Read a very interesting technical paper written by Jack Dongarra Related articles Find Table Scans in Query Plan Cache Memory grants and execution plans How to find CurrentClockSpeed and MaxClockSpeed using Powershell... Read more →