SQL Server Process Processor Time (CPU Utilization) Over 100%
Mar 13, 2008
We are seeing that the %Processor Time for the sqlservr process in Perfmon is over 100%. I am trying to understand how can the percentage of use be over 100%, and why it is over 100%. Someone told me that if the machine has multiple processors, that it will be over 100%. If that is the case, how can I determine what the maximum and normal values are? If I have 4 processors, does that mean 400% is the max? Does not make sense since it is suppose to be a percentage value...
Could someone explain to me how the CPU Utilization value is being measured, and if it is going over 100%, why that is and how I can determine what the threshold should be for monitoring?
I have noticed that when using SQL Query Analyzer some of my queries will use 100% CPU on my PC and next to nothing on the SQL server, while other Queries require 100% CPU on SQL server and do next to nothing on my PC. Does anyone know what determines this??
Right now I can produce this by executing two very similar T-SQL selects. The one that runs on the server only has one additional join - a very simply join at that. If I can change my SQL to make it run client side in some situations, that would be VERY HELPFUL!
I have a data mining app that does a series of select statements (no inserts). I'm noticing an odd occurance where if I start up 4 copies of the app on a quad core machine - sql takes full advantage of the 4 cores for a few minutes and then drops to 75% utilization - the other 25% is on the idle process. Two of the apps appear to be sharing a single proc of sql as each of their throughputs is cut by 50%. If I then start a 5th copy of the app, the machine is brought to full 100% utilization - two of the apps continue to appear to share a proc. SQL is set up to use all procs and I have even tried select the priority boost to no effect.
Any ideas how to ensure full sql utilization with the same number of apps as cores?
We've had a problem for a few months now that has completely stumped us. We are running a heavily cursored massive data manipulation process on a 32 bit SQL Server instance running on a virtual machine, running ontop of VMWare, with the following specs
Processors: 2x2674MHz processors Memory: 4GB RAID 10 disk config
When we run our process on this machine, in total it runs in 30 hours.
When this process is run on another 32 bit server with the following specs
Processors: 8x3658MHx processors Memory: 8 GB SAN w/ RAID 5 disk config
It runs 25% slower
But here is the real kicker. When this process is run on a 64 bit server with the following specs
Processors: 8x3658MHz processors Memory: 8 GB SAN w/ RAID 5 disk config
It runs 75% slower.
This process consists solely of stored procedures written in TSQL. The weird thing is that on our smaller server, the CPUs' % utilization are evenly balanced (at 20-30%) when this large data manipulation process is running. However on the bigger servers, SQL Server latches onto a single processor and doesn't load balance across other processors. Such that what we're seeing is that only one processor out of the eight will be utilized and it will be throttled at 90% while the other 7 are at zero.
The default configuration settings in all three places.
Has anyone ever seen any behavior like this, where only one processor gets used by SQL Server during processing? Granted our processes are single threaded b/c they are using cursors but, it seems that the single thread shouldn't be restricted to one processor.
I was tuning a query testing out SARG with these two queries:
select col1 from table1 (nolock) where col1 like '#,%ABC%' or col1 like 'BC,%ABC%' select col1 from table1 (nolock) where col1 like '%ABC%'
I flushed out the cache, added an index to col1, then ran those two together. Provided below are the actual query plan and stat time:
Query 1: Query cost (relative to the batch): 0% select col1 from table1 (nolock) where col1 like '#,%ABC%' or col1 like 'BC,%ABC%' ---------------------------------------------------------------------------------------------------------------------------- SELECT Index Seek Cost:0% <------- [DB1].[dbo].[table1].[Idx..]
Cost: 100%
Query 2: Query cost (relative to the batch): 100% select col1 from table1 (nolock) where col1 like '%ABC%' ---------------------------------------------------------------------------------------------------------------------------- SELECT Index Scan Cost:0% <------- [DB1].[dbo].[table1].[Idx..]
Cost: 100%
------------------------------------------------STAT TIME----------------------------------------------------------- SQL Server parse and compile time: CPU time = 0 ms, elapsed time = 1 ms.
SQL Server Execution Times: CPU time = 0 ms, elapsed time = 1 ms. SQL Server parse and compile time: CPU time = 0 ms, elapsed time = 7 ms.
As expected, SARGable Query 1 did a nonclustered index seek and nonSARGable Query 2 did an index scan instead. According to the query plan, Query 1 consumed 0% relative to the batch whereas Query 2 is 100%. When I checked the CPU time, I was a bit confused because Query 1 showed CPU time of 938ms whereas Query 2 showed 515ms. I triple checked and every time I got similar results. I am sure I'm missing something, could someone please tell me what I'm missing? Thanks a bunch!
I have a server with little control over most of the codeset and db design. Recently I have seen both the Processor - %Processor time and Processor - % User time go fom about 6.3 to about 24.3. The system queue length has also gone from about .2 to 1.1. In my humble opinion both of the are signs of a problem coming (luckily the cache hit ratio is still sitting at about 99%). I have been running profiler to catch the things that take more that 4500 MS, and I can probably tie the 2 together. Any opinions, or real world comparisons appreciated
Having a SQL Server 2012 Enterprise (x64) on a Windows 2012 R2. We need to know, a reliable way, the number of processor sql server is using at a give time. We already know how many total processor are available to sql by getting info from sys.dm_os_sys_info.
For instance, a server has 40 processors, we want to know how many of those are being used at a given time. Since the load on the server may not be that high, we would like to know how many processors we can eliminate and the load will still be unaffected.
After watching the server performance for a while, we are predicting we may only need 16. But we would like to get some statistics before we reduce it to this number.
I have MS Time Seeries model using a database of over a thousand products each of which has hundreds of cases. It amazingly takes only a few minutes to finish processing the model, but when I click Mining Model Viewer to view the models, it takes many hours to show up. Once the window is open, I can choose model for different products almost instantly. Is this normal?
I am assigning package string variable to the StandardOuputVariable of a Process Task, and setting the TimeOut to 3, TerminateProcessAfterTimeOut to True, plus the Executable name. The program is executing about 40 seconds, and it seems not to terminate after the 3 seconds has passed.
When I execute the same package but without the StandardOuputVariable the task terminates after the 3 seconds.
Is this a bug or I shouldn't expect the task to terminate when the StandardOuputVariable is set???
I'm having a problem with one of my SQL servers (2000 Build 8.00.2140) where it is always reading CPU utilization of 70-100% (more often pegged at 100%).
I have an exact same SQL server running at a different location on a much less powerful (hardware wise) server that gets more traffic but only shows 7-21% CPU utilization.
Taskmgr shows Process sqlservr.exe as consuming all the resources. This is a dual 2 core 3.66Ghz (4 real CPUs) with 4GB RAM and 5 x 146 15K SCSI drives hooked up to a $1200 SCSI contoller (Dell server). RAM usage is pretty low, most I've seen is 1GB.
Is there any way to determine what specific connection/thread is causing this? Any diagnostic tools or anything that can show me specifically what is consuming this SQL server? Connection thread or anything that points back to a specific IP?
I made a cube with time dimension with hieracly year/month/date/hour the problem is that dimension is growin to fast. In older version of MSSQL (2000) the same dimension doesn't grew so much. Any ideas? The table is big (may be around 1 500 000 rows per month) now it contains around 4 500 000 rows.
I'm using a OLE DB COMMAND component to perform an update (SQL statement) but the procces takes about 9 hours, so I changed it to a stored procedure but it was the same, I need to update about a a million of rows and the package is very simple.
How can I improve the time, Can I use another component or startegy?
HI I am facing a strange problem with SQL Server 2005 . The CPU utilization with SQL Server 2005 is higer by about 70% comapared to SQL 2000.
In the same kind of Hardware and with the DB server up , I performed the following tests Declare @i int Set @i = 10 While @i < 100000 Begin Insert into arup_emp values(@i,'M',0)
Set @i = @i + 1 end
The CPU utilization average on SQL 2005 was 45% and on SQL 2K it was just 25% , I am seeing a lot of people who seems to be facing this problem but unfortunately not seeing any solution to this.
Can anyone through some light . Please note that I have tried to also see the MAXDOP options, but get the same results.
Our company recently combined our DBs into one SQL 2005 Server.
Dell Power Edge 1800 with 3.00 GHz Xeon Processor 800 FSB, 1 GB of RAM Dell Power Edge 1600 with 2.80 GHz Xeon Processor 533 FSB, 1 GB of RAM
Combined into one: Dell Power Edge 2950 Dual Core 1.6 GHz Xeon Woodcrest Processor, 4 GB of RAM
However, the CPU utilization on this new server is maintaining at about 90% with 3.82 GB of RAM used as well. It's a Windows Server 2003 R2 x64 edition running SQL Server 2005 SP2 x64. I have searched around Microsoft's website for any information that could be of help to me, but I was unable to locate anything. I was hoping that someone could provide some insight as to why this might be occuring. Or if this is a known issue.
We are using an IBM Xeon server with 4 GB RAM with windows 2000 server and MS Sql server 2005.
More frequently our server responce time is very slow, although the cpu utilization is between 7 - 10 %. we can not able to run even notepad on the server. we observed that the memory occupaid by the sqlserver prog is high. If we restart the server then it will return to normal level. But we do not want restart the server frequently.
We are using an IBM Xeon server with Sevrer 2000 and sql server 2005 with 4B RAM.
We noticed that the responce time of the server is very slow. During this time the memory occupancy by the sqlserver is very high. Althouh the cpu utilization is very low (7-10%) we are unable to run even notepad on the server. In some other situations the cpu utilization will be 100% (for more than one hour) during this time also the sql server occupaies more memory. If we reatarts the server then the problem will be fixed, but we do not want to restart the server very frequently.
We are having a big performance issue at our site. Here is the configuration of the box running SQL Server 2005:
64 bit Windows Enterprise Edition + SP1
Dual CPU, 16GB RAM
RAID 1 and RAID 5 - internal
SQL Server 2005 64-bit Enterprise Edition
With SP2 (CTP from December)
The "Lock Pages in Memory" is set and is being run under the same account that is being used to run SQL Server Services.
We are noticing that under load, the CPU utilization becomes nearly 100%. I have researched this and have come across a couple of posts that indicate that this issue was fixed in SP2 - example: One post talked about the hotfix #716 which is also a part of SP2 but even after the application of that service pack, we are still having this issue. I haven't tried setting the parameterization option to forced for the database yet.
Is this a known issue with SP2? If not, what can we look for and fix in our environment? Please let me know if I can provide more information.
Good afternoon everyone, I have written a view that pulls customer demographic infomration as well as pulling data from multiple scalar-valued functions. I am using this view to pull and send data from one database to another in the same SQL server. The problem that I am having is that I am running this import as a scheduled job in windows. The job is taking almost 24 hours to complete this task. The total number of records that are being pulled is around 21,000+. I have tried removing the functions from the view and it only takes the view 20 seconds to pull the demographic information from the same 21,000+ records but when I add the function calls this is where the time to complete goes through the roof. Has anyone encountered this before if so what would you suggest doing? Any help would be appreciated. Here is the syntax for my view: SELECT TOP 100 PERCENT CUS_EMAIL AS Email, CUS_CUSTNUM AS MemberID, CUS_PREFIX AS Prefix, CUS_FNAME AS FirstName, CUS_LNAME AS LastName, CUS_SUFFIX AS Suffix, CUS_TITLE AS Title, CUS_STATE AS State, CUS_COUNTRY AS Country, CUS_ZIP AS ZipCode, CUS_SEX AS Gender, CAST(CUS_DEMCODEA AS nvarchar(20)) + ',' + CAST(CUS_DEMCODEB AS nvarchar(20)) + ',' + CAST(CUS_DEMCODEC AS nvarchar(20)) + ',' + CAST(CUS_DEMCODED AS nvarchar(20)) AS DemoCodes, dbo.GetSubScribedDateMLA(CUS_CUSTNUM, CUS_EMAIL) AS MLASubscribedDate, dbo.GetSubScribedDateMLP(CUS_CUSTNUM, CUS_EMAIL) AS MLPSubscribedDate, dbo.GetSubScribedDateLDC(CUS_CUSTNUM, CUS_EMAIL) AS LDCSubscribedDate, dbo.GetMLAExpiration(CUS_CUSTNUM, CUS_EMAIL) AS MLASubExpireDate, dbo.GetMLPExpiration(CUS_CUSTNUM, CUS_EMAIL) AS MLPSubExpireDate, dbo.GetLDCExpiration(CUS_CUSTNUM, CUS_EMAIL) AS LDCSubExpireDate, dbo.IsProspect(CUS_CUSTNUM, CUS_EMAIL) AS AGMProspect, dbo.IsCurrentCustomer(CUS_CUSTNUM, CUS_EMAIL) AS AGMCurrentCustomer, dbo.IsMLAMember(CUS_CUSTNUM, CUS_EMAIL) AS MLAMember, dbo.IsMLPMember(CUS_CUSTNUM, CUS_EMAIL) AS MLPMember, dbo.IsLDCMember(CUS_CUSTNUM, CUS_EMAIL) AS LDCMember, dbo.CalculateTotalRevenue(CUS_CUSTNUM, CUS_EMAIL) AS AGMTotalRevenue, dbo.GetPubCodes(CUS_CUSTNUM, CUS_EMAIL) AS ProductsPurchased, dbo.GetEmailType(CUS_CUSTNUM, CUS_EMAIL, CUS_RENT_EMAIL) AS EmailType, CUS_COMPANY AS Company, CUS_CITY AS City FROM dbo.CUS WHERE (CUS_EMAIL IS NOT NULL) AND (CUS_EMAIL <> '') AND (CUS_EMAIL_VALID = 'Y') AND (CUS_EMAIL LIKE '%@%.%') AND (CUS_RENT_EMAIL = 'Y' OR CUS_RENT_EMAIL = 'R' OR CUS_RENT_EMAIL = 'I') AND (CHARINDEX(' ', CUS_EMAIL) = 0) AND (CUS_EMAIL NOT LIKE '@%') Thanks in advance Michael Reyeros
Hello:I run one process that calls the following the store procedure andworks fine.create PROCEDURE sp_GetHostSequenceNumASBEGINSELECT int_parameter_dbf + 1FROM system_parameter_dbtWHERE parameter_name_dbf = 'seqNum'UPDATE system_parameter_dbtSET int_parameter_dbf = int_parameter_dbf + 1WHERE parameter_name_dbf = 'seqNum'ENDGOIf I run two processes that call the above store procedure, I mightoccasionally get the dirty data of int_parameter_dbt. I guess that iscaused by two processes accessing to the same resource simultaneously.Is there any way to lock the store procedure call from MSSQL Serverand allow only one process to access it at a time?Thanks for help.Best Jin
I've just updated my databases to V7.0 from 6.5. I am having some issues with CPU utilization after rebuilding the databases. I rebuilt the databases from scripts after modifying the scripts to deal with any SQL issues between the versions.
Symptoms:
When a bunch of stored procedures are executed the % Processor Time his 100% on a single processor system, the Processor Queue Length hits 13, the Batch Request/Sec starts at 20 and rises to 47 before dropping off, Context Switches go from 370/sec to about 833/sec before dropping off and there is an increase in page faulting.
My testing was done on a single processor but my real system is dual-300+ with 128Mb RAM. The database is only 40Mb and after running the index wizard there were no changes advised.
After reading a lot in help areas for MS and here I am not sure if this has to do with how SQL Server runs now (i.e. system configuration) or if it is a query issue with parallelism. I'm assuming query parallelism since the performance spikes occur when stored procedures are run. On a single processor is there a way to address this. Then, hopefully, I can deal with the real multiprocessor system.
If I install an instance of SQL Server on a box with 4 processors but specify in the licensing part of the installation procedure that it's a '2 processor license', does that really mean that SQL Server will only use 2 of the 4 processors? I ask because, having done this, if I check the SQL Server instance's properties, on the Processors tab, there are 4 processors ticked - giving the impression that 4 processors are being used by this SQL Server instance. Can anyone throw some light on this?
Another question... If I wish to reconfigure the same SQL Server instance from a 2 to a 4 processor 'per processor' license, how can this be done?
Every once in a while (every couple weeks or so) one of our SQL Server 2005 servers will have its performance take a dive. Response time on applications will slow way down. In checking €˜Task Manager€™ the CPU utilization will be up to 50 to 70 percent for the SQL server service. We have tried looking at €˜Perf Mon€™ there doesn€™t seem to be any consistent performance item that is out of whack. On occasion, it may appear to be a lot of paging going on, but I don€™t know what application is causing it. We€™ve looked at SP_who2 to see if there are any blocks, but there do not appear to be any. Though on occasion we will see a €˜Killed/roll-back€™ entry on €˜SP_who2€™ that has a lot of CPU cycles that, at least at one point in time, had a lot of counts against it.
In looking at a trace that we have set up, we don€™t see anything glaring that is held up or even a recurring theme (application) on the list that would point to a problem.
We find that by restarting the SQL Server service, this resolves the issue for another couple of weeks.
Is there something that I can look at that would tell me at least, what PC name is doing all of the IO against the database. I could then at least call the user and say €˜Hey, what are you doing€™ so that we can start to narrow down what is going on.
Regarding Standard vs Enterprise processor usage, I have read the SQL2K5 edition comparison articles that MS puts out, but I recently viewed a webcast that gave me cause for concern.
There was a webcast called "Selecting the Right SQL Server 2005 Edition", hosted by Michael Raheem, where he stated that although Standard edition allowed 4 processors (based on sockets, not cores) some index operations only leveraged one core at a time. So even if you have, let's say, 4 dual core CPUs (= 8 processor cores), only one core at a time will be used, and the rest will be idle during that operation.
Can anyone give a better idea of which operations are full multi-core in SQL Standard. Michael mentions indexing - does that mean strictly explicit reindexing operations which are part of a maintenance routine, or does that apply to index operations during normal inserts and/or updates?
We will be using this to run a backend for a commercial website, and we need to be sure whether we can get by with Standard or we need to make the investment in Enterprise.
Hi everyone, Primary platform is 2005 under 64-bit.
I'm trying to figure out what does 'Processor Affinity' and 'I/O Affinity' means when you're viewing server properties. We've got eight processors. Thanks for your comments,
I was browsing Microsoft's SQL Server site, looking forsome details about SQL Server 2005. Didn't find whatI was looking for...I'm thinking about moving an existing SQL Server 2000workload to a new box, using SQL Server 2005, andmaybe the 64-bit version.My questions are:1. What is the current target date for release of SQL Server 2005?Will 64-bit ship when 32-bit ships?2. Will 64-bit SQL Server 2005 require a special versionof Windows Server 2003 (e.g. Windows Server 2003 Enterprise x64)?Will it work with both Intel and AMD processors?3. How many CPUs, and how much memory, will be supported bySQL Server 2005, 32-bit and 64-bit, on each OS that can runSQL Server 2005.I'm looking for a chart here, something like the chart onpage 117 of Kalen Delaney's "Inside SQL Server 2000" book.SQL Server 2005 SQL Server 2005Feature Enterprise 32-bit Enterprise 64-bit------------------- ----------------- -----------------CPUs supportedWin Srvr 2003:Win Srvr 2003 Adv:Win Srvr 2003 Ent x64:Physical memorysupportedWin Srvr 2003:Win Srvr 2003 Adv:Win Srvr 2003 Ent x64:Has Microsoft published this info, and I just can find it?
We have recently tested upgrading our web service from sql 2000 to 2005 sp1. The upgrade went smoothly enough, however we now have the problem of the sqlserver.exe process taking 90-100 % of the processors time, but using only 100 MB of memory.
We have 6GB available and we are running the enterprise editions of Windows 2003 and SQL 2005.
Machine specs,
DL380 G2, 2 X 2.8 Ghz Zeon, 6GB ram, Raid 5, database partition of 140 GB, Log partition of 35 GB.
Db is 25 GB, Log is 12 GB. Largest table has 32 million rows.
