I have a procedure in a history database that does insert into 3 tables inside a transaction. users complaint that the proc sometimes takes too long during heavy usage. I did some traces to see what is taking up the time, I found that the rpc duration was averaging > 500 ms (should only take 50ms). I checked to see if one of that statements were taking too much time, but only see the commit transaction statement taking around 500 ms). I check the avg disk queue to be around 30. ( this is on a single local disk) .
So is this definitely a disk issue, or is there something else I need to check
I need some help knowing what to look for in Profiler to troubleshoot an issue.
I've got an application that accesses a SQL Server database that has suddenly started timing out when users launch and attempt to log in, and I'm trying to find out where and why the application might be timing out (whether it's a server issue, a stored procedure or SQL query from the application that could be optimized, a table that could be truncated or archived, etc.). All I have to work with from troubleshooting the database side are a series of trace files from Profiler that were run for a total of about 5 minutes while the application was launched and then timed out. Of course, there are a whole lot of statements being issued, hundreds of tables being accessed, lots of stored procedures and even more ad-hoc queries coming straight from application source code.
So my question is, what do I need to look for in these trace files that might be a red flag to an issue? I'm no DBA, but I know that really long durations might be a tip-off. I'm only seeing these on the occasional Event:Audit Logout (which I read in another thread could potentially be very normal). Anything else that I might want to filter for?
I am debating whether to go to all the trouble of setting up on-demand Profiler traces on some test servers for the developers here. Really just tracing RPC:Completed and SQL:BatchCompleted, so the developers can at least try to catch a performance problem before going to production. The question I have, though, is just how useful is this sort of information to mid- to low-level (i.e. experience) developers. One of the bigger concerns is over Java applications, which like to hide their queries behind a lot of "sp_cursorfetch" calls.
My question to the forum is if you are a developer, have you ever dreamed of having this sort of information available? How useful is it?
I am going to try to post a poll along with this, but I am not sure it will work..
If monitoring for duration with sql profiler, what does the number represent ie 2733906 is it milliseconds, thousandths, looked in BOL no clear definition
Sarfaraz writes "I have captured SQL Profiler data. I was reviewing top running CPU intensive SQL statements. The Duration (in seconds) 1.39, 1.09, 0.16 and CPU (in seconds) 0.97, 0.95, 0.16 respectively for some SQL statements. How do I know what is the normal baseline for duration and CPU in order to determine the CPU intensive SQL statements.
Secondly same question for long running procedure duration 0.14, 0.11. What is the normal baseline here. Is this normal or too long.
We have a VLDB ( few table with above 200 million records ). This database is used for performance testing by simulating for 150 users and executing all necessary functional flows.
When I examined the profiler results , I could see some very high values as shown below in the duration column for many events.
1521729 3462142 1624325 3211255 1248276 3903998
Does it mean that that SP or the T-Sql statement is taking this much time in milliseconds to execute and give the output ?
I have a very stranger problem that I need to understand... In last days I executed a plan SQL 2005 Profiler to review TSQL Duration. When reviewing the results encounter that a SP displays a value of 4037312 in field DURATION which are not normal. Could to help me to identify why passes this?
I've set the Duration of my trace to "Greater than or Equal to: 1000". However when I start my trace the Duration column is now empty. Prior to the setting, there were values showing in this column. Any ideas on how to fix this?
Profiler was run against a database looking for "long running" queries. I used the Duration column to filter out the queries that I didn't want. When reviewing the output, I noticed that for some queries the StartTime was equal to the EndTime even though the Duration was set higher. My question is, "What can account for this discrepancy and what inferences should I draw about the difference?" Does the difference represent a resouce being locked or some other type of blocking (Duration) and once the query was allowed to run, it completed quickly? TIA
======================================= If Tyranny and Oppression come to this land, it will be in the guise of fighting a foreign enemy. -James Madison, fourth US president (1751-1836)
Which works fine, but what I need to calculate the total duration of a request based on the duration of the tasks completed in the request based on Req_ID. I would like to use the CASE statement I have to determine the SLA_Mins for each task and add them together to get total request SLA_Mins.
Below is the create table schema and data
GO /****** Object: Table [dbo].[MidrangeOtherSourceControl] Script Date: 06/03/2015 18:13:15 ******/ SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO CREATE TABLE [dbo].[MidrangeOtherSourceControl]( [Req_ID] [float] NULL, [Service_Name] [nvarchar](255) NULL,
Any suggestions on how I can monitor the following without using traces? I am a dba/developer working as a developer on a contract, and I'm supposed to be tuning. However, I can't run traces. I've got my own procs that monitor locking, etc. But I would like to get at least i/o and cpu throughout the day. It would also be nice to get the query executed. Basically, the type of stuff you'd normally use traces for.
I know about @@cpu, @@io etc., but these are basically useless (no?) since they only record since the server was started. There is a stored proc but it only monitors these things since the last time it was run.
Does anyone know how I could utilize the above? I tried to write a script but I couldn't get it to work. :(
I realize that in general this is a ridiculous request, but I thought I would ask anyway.
This one has stumped me. Hopefully somebody can help. A while ago, I setup a trace that posted the log to the desktop. I needed to stop the trace this morning, so I went into the profiler and deleted the traces. There was a private and shared trace. Now every time I start up something that has to do with sql server, the log pops up on the desktop. I'm not sure why the trace wasn't deleted or stopped. The trace includes what program accessed sql. Whether it is EM or Query analyzer or ISQL. It gets posted in the log. Any suggestions? I need to remove this because the log fills up the drive and causes the server to crash.
I want to use ShowPlanXML in my traces. Information witch is given by this event is quite interesting but I dunno how to bound it with other events. For example I have SqlBatchComplete event and want to get XML plan of it but I have no idea how to do it.
When I restart the server that hosts the database engine SQL Server 2005 Standard Edition SP4, the trace gets up mytrace-5.trc with a size of 100 MB and id is 2, leaving the hard disk space, then the SQL Server stops execution of the trace due to lack of space. I do not know how to erase the trace, because I do not know where is it.
The archive log looks like : SQL Trace ID 2 was started by login "sa"
Trace ID '2' was stopped because of an error. Cause: 0x80070070(There is not enough space on the disk.). Restart the trace after correcting the problem.
We are planning to convert or change all existing Traces to Extended Events in SQL server 2012. What is the procedure to convert custom traces. We have already created some below custom traces: like this we are planning to convert for all servers.
I found this nifty code on stackoverflow that works well but I'm trying to send the results to a text file and the column lengths are huge. I used CAST for the first line and it worked great but I can't seem to make it work with duration. Here's the original code:
I am trying to get a query that will allow me to report the time taken to complete a certain training module.
The database itself does not have a duration field so I am tring to get the duration by using MIN and MAX. I can get the timing for when the module was opened and the time for the last mouse click on it, from this I need to be able to calculate the time taken to complete.
Query I am using to get the basic info comes from 3 tables so I have only attached the relevent output. Query used is as follow:
SELECT * FROM PPS_SCOS, PPS_TRANSCRIPTS, PPS_TRANSCRIPT_DETAILS, PPS_PRINCIPALS WHERE PPS_SCOS.SCO_ID = PPS_TRANSCRIPTS.SCO_ID AND PPS_TRANSCRIPTS.TRANSCRIPT_ID = PPS_TRANSCRIPT_DETAILS.TRANSCRIPT_ID AND PPS_TRANSCRIPTS.PRINCIPAL_ID = PPS_PRINCIPALS.PRINCIPAL_ID AND PPS_SCOS.NAME LIKE 'MTM-106 The Dangers of Smoking' AND PPS_PRINCIPALS.NAME LIKE 'Nigel Cordiner' AND PPS_TRANSCRIPTS.TICKET NOT LIKE 'l-%' ORDER BY PPS_TRANSCRIPT_DETAILS.DATE_CREATED
Output:
pps_scospps_scospps_transcript_detailspps_principalspps_principals SCO_IDNAME DATE_CREATED PRINCIPAL_ID NAME 136850MTM-106 The Dangers of Smoking08:17:2516287Nigel Cordiner 136850MTM-106 The Dangers of Smoking08:17:2516287Nigel Cordiner 136850MTM-106 The Dangers of Smoking08:17:4016287Nigel Cordiner 136850MTM-106 The Dangers of Smoking08:18:2516287Nigel Cordiner 136850MTM-106 The Dangers of Smoking08:18:5716287Nigel Cordiner 136850MTM-106 The Dangers of Smoking08:19:1416287Nigel Cordiner 136850MTM-106 The Dangers of Smoking08:19:4716287Nigel Cordiner 136850MTM-106 The Dangers of Smoking08:20:2116287Nigel Cordiner 136850MTM-106 The Dangers of Smoking08:20:4416287Nigel Cordiner 136850MTM-106 The Dangers of Smoking08:21:2616287Nigel Cordiner 136850MTM-106 The Dangers of Smoking08:22:1316287Nigel Cordiner 136850MTM-106 The Dangers of Smoking08:24:5516287Nigel Cordiner 136850MTM-106 The Dangers of Smoking08:25:1216287Nigel Cordiner 136850MTM-106 The Dangers of Smoking08:25:2916287Nigel Cordiner 136850MTM-106 The Dangers of Smoking08:26:4916287Nigel Cordiner 136850MTM-106 The Dangers of Smoking08:27:0216287Nigel Cordiner 136850MTM-106 The Dangers of Smoking08:27:2916287Nigel Cordiner 136850MTM-106 The Dangers of Smoking08:27:4316287Nigel Cordiner
Have added the column heading and the tables the output comes from.
Relatively new to SQL so any help would be greatly received.
SELECT h.JobNum, (CASE WHEN MONTH(h.JobCompletionDate) = 1 THEN datediff(day,MIN(l.ClockInDate),MAX(l.clockInDate)) ELSE 0 END) AS JAN, (CASE WHEN MONTH(h.JobCompletionDate) = 2 THEN datediff(day,MIN(l.ClockInDate),MAX(l.clockInDate)) ELSE 0 END) AS FEB, ... FROM JobHead h INNER JOIN LaborDtl l ON h.JobNum=l.JobNum WHERE JobCompletionDate>='20070101' AND JobCompletionDate <'20080101' AND l.ClockInTime<>0 GROUP BYh.JobNum ,h.JobCompletionDate
The query shows, for each job, the month in which the job completed, and the number of hours it took to complete. I€™m calculating the number of days€™ duration by doing a datediff between the oldest and newest clockindates. I need to ignore adjustment transactions in the labordtl table €“ these rows are easily identified as they have clockintime values of 0. So far, so good. Now here€™s my problem.
There are some jobs which have only one €śreal€? labor transaction €“ this could happen if the job only took one day to complete. Other labor transactions may exist for that job, but let's say they are adjustments which we can ignore -- the date they were entered should not extend the duration of the job. In this situation, my datediff between the oldest valid transaction and the newest, returns 0. I don€™t have to count hours between clockintime and clockouttime. The rule is simply that if there is only one "real" labor transaction, I need to count this as a 1 day job.
I thought a nested CASE statement or expression might be the way to go but I didn't make any real progress.
Any ideas to solve this problem would be appreciated.
I'm developing a web app that displays the running packages and the total elapsed time. I'm calling GetRunningPackages() method and using the ExecutionDuration property of the returned package. The duration seems to be only for the currently executing container and not the entire package. Is there a way to get the duration of the entire package? Thanks.
Hi all, I want to find working duration between two datetimes in c#.i'm using following code... DateTime starttime = Convert.ToDateTime(Session["StartTime"]); DateTime endtime = DateTime.Now; TimeSpan duration = endtime - starttime; DateTime period = new DateTime(duration.Ticks); i want to store this duration in database through stored procedure, i've give datetime datatype to duration but it is giving error in conversion of TimeSpan to DateTime..Please help... Thanks
SQL 6.5 - run duration 6-7 hours SQL 7.0 - run duration 12-13 hours 175+ columns with total record size=570 4.2M records with tablesize 2.5G
It's a simple 'select into...' with some embedded logic from a work table with all char fields into the actual table converting char fields into various data types (int, datetime, real, etc.) Here is a sample of the code:
SELECT LoanNum=CASE WHEN ISNUMERIC(ACCT#)=1 THEN CONVERT(int,ACCT#) ELSE NULL END, PaidToDt=CASE WHEN PAIDDT = '0001-01-01' THEN NULL WHEN ISDATE(PAIDDT)=1 AND SUBSTRING(PAIDDT,1,2) = '19' THEN CONVERT(smalldatetime,PAIDDT) WHEN ISDATE(PAIDDT)=1 AND SUBSTRING(PAIDDT,1,2) = '20' AND SUBSTRING(PAIDDT,3,2) < '79' THEN CONVERT(smalldatetime,PAIDDT) ELSE NULL END, . . . INTO db.owner.tablename FROM db.owner.wrktablename (NOLOCK)
Hearing complaints from users about speed on db server (I have almost no control on design) it just has to work. Ran profiler looking for all sql statements over 4000 millsec and in one hour returned over 715 tsql statements. Over 300 of these were over 10000 milliseconds. THis is on an 8 way Dell with 8 gig of RAM. Looking for opinions, how bad does this look compared to other servers you are taking care of? Cache hit ratio is at 99 % and system queue length still under 1, but this does not look good.
There is a trigger to monitor the modification on a table, and it turn on. For a special duration, I need to turn off this trigger to modify the table. And then turn on the trigger again.
I have a table called Tickets which contains ticket information for a machine. Each machine can have more than one ticket number opened at the same time. The ticket number contains start date/time and end date/time of the ticket. Thereefore the table looks something like this:
I want to be able to calculate total duration time(in hours) that EACH MACHINE had a ticket open...but here is the tricky part. The total duration time that a machine had ticket open has to encompas any tickets that may fall in the same time period. For example: If Machine A has a ticket open at 8:30 and the ticket is closed at 10:00. Meanwhile, Machine A had another separate ticket open at 9:30 which was closed at 10:30. In this case, the total duration time for this machine would be from 8:30 to 10:30 for a total of 2 hrs duration time.
Can anyone help me get started in tackling this problem or provide any examples?
Is there anyway to tell how long this will run for -- or how far it has got? I have a large database that has just had most of the data removed. The command has been running for 8 hours and I have just stopped it to let something else run quickly. Any way of telling how much longer it will take?
/* This SP has 2 functions. a) if @method='duration' gives the average run duration in minutes for successful jobs b) if @method='failures' displays failures/cancels/still executing jobs It defaults to today's date. Specify @xdate for a different date -- Louis Nguyen */
CREATE PROCEDURE UtilityJobsHistory ( @method varchar(100)='duration' ,@xdate datetime=null ) AS set nocount on set transaction isolation level read uncommitted
if @method='duration' begin
select @xdate=isnull(@xdate,getdate())
/*run_duration is in HHMMSS format; drop SS*/ /*run_staus: 1 complete 2 retry*/ /*step_id: 0 is final job outcome*/ /*run_date: yyyymmdd format*/
/*today's performance*/ select a.name,minutes=avg((b.run_duration / 100)/100*60 + (b.run_duration / 100)%100) into #today from msdb..sysjobs as a join msdb..sysjobhistory as b on a.job_id=b.job_id where run_status in ('1','2') and step_id=0 and run_date =convert(varchar,@xdate,112) group by a.name
/*7 day average performance*/ /*populate #D with dates in yyyymmdd format*/ create table #D (run_date varchar(50)) declare @idate datetime set @idate=@xdate while @idate>dateadd(day,-7,@xdate) begin insert into #D select run_date=convert(varchar,@idate,112) select @idate=dateadd(day,-1,@idate) end
/*Avg7Days*/ select a.name,minutes=avg((b.run_duration / 100)/100*60 + (b.run_duration / 100)%100) into #avg7Days from msdb..sysjobs as a join msdb..sysjobhistory as b on a.job_id=b.job_id join #D as c on b.run_date = c.run_date where run_status in ('1','2') and step_id=0 group by a.name
/*output*/ select name=cast(a.name as varchar(35)),OneDayAvg=a.minutes,SevenDayAvg=b.minutes from #today as a join #avg7days as b on a.name=b.name order by a.name
return end
if @method='failures' begin
select @xdate=isnull(@xdate,getdate())
select status=case run_status when 0 then 'FAILED' when 3 then 'CANCELED' when 4 then 'EXECUTING' end ,name=cast(a.name as varchar(35)),step_name ,time=replace(convert(varchar,@xdate,107),' ','')+' '+right('0000'+cast(b.run_time/100 as varchar),4) ,b.message from msdb..sysjobs as a join msdb..sysjobhistory as b on a.job_id=b.job_id where run_status in ('0','3','4') and run_date=convert(varchar,@xdate,112) order by run_status,a.name
Hi guys, I am having difficulty calculating the time duration between receiving process to shipping process. I have a table that consists of: Order#, Processes, Time_In, Time_Out. Order# can be 1, 2, 3, 4, 5. While at the same time Order# 1 can go through more than one process, i.e.: Receiving, VisualTest, MechanicalTest, ..., Shipping. Every Order# does not necessarily goes through all processes, but surely they will go through receiving process and shipping process. For each process we will have recorded time when the order# comes in and when it finishes with each process. I need to calculate the length of time from Time_In from Receiving to Time_Out in Shipping.
I'm using RDA (Remote Data Access) to pull 20 tables to my Pocket PC. It took quite a long time so I ran a trace to see what happened. Everything looks fine except for when it runs:
exec [mydb]..sp_primary_keys_rowset N'Person',NULL The duration is: 18446744073!!!
A couple of more tables has this enomous duration others have about 5000 which seems more normal.
