I've installed a SQL Reporting server (2008 R2), with some reports. But I've some performances issues.
The first call of the day to the server(going on the report interface by example), is VERY slow(something like 30-45seconds at best).
The report generation is then "fast"(1-2 seconds).
The next calls to the server are always fasts until the next day. I've the impression that it loads a lot of thing in the memory. But what can takes 30-45 seconds to be loaded in memory??? And how to load it only once?
The server is good enough(quad core, 8GB of ram, never near its capacity for now).
What is the problem? How can I resolve this ?
Thoses reports will be launched only 4-5 times in a week, so they will always be slow if I can't change this. And since it's available for customer, I just can't make them understand this(and the report is called through a website, so I risk to have timeout).
Thank you very much
It seems to be an SSRS issue. There is nothing wrong with your report.
It's "normal" that SSRS takes more time to load the first time you access it after a long time of inactivity. The issue is caused by the way how SSRS works and SSRS regularly restarts application domain after a specific time period. After the application domain is restarted, then upon first request to the SSRS it needs to load all the settings and it takes quite a long time.
This blog show's a workaround for the situation
Here is the powershell script that I wrote to fix the problem. It is setup as a task to run every 1:00am:
Stop-Service "SQL Server Reporting Services (MSSQLSERVER)"
Start-Service "SQL Server Reporting Services (MSSQLSERVER)"
$wc = New-Object system.net.webClient
$cred = [System.Net.CredentialCache]::DefaultNetworkCredentials
$wc.Credentials = $cred
$src = $wc.DownloadString("http://localhost/Reports/Pages/Report.aspx?ItemPath=***NAME OF HOME PAGE***")
The best solution I could come up with was to issue a 'curl' command to the http page of the report using via windows batch command in a windows service. This ran up the page(s) every morning before the users came in.
Don't have access to the code anymore (job was a while ago) but this question shows how to use curl:
http://blogs.plexibus.com/2009/01/15/rest-esting-with-curl/
As Diego said, SSRS has some issues. The first call it is slow, regarding your server configuration. I recommend you the following config to add in rsreportserver.config (located if you don't know in C:\Program Files\Microsoft SQL Server\MSRS10.MSSQLSERVER\Reporting Services\ReportServer\ )
If you want to increase the max memory used by SSRS: (which means 7 GB)
<WorkingSetMaximum>7000000</WorkingSetMaximum>
If you want to improve the first call, you can set (in minutes)
<RecycleTime>4320</RecycleTime>
Actually SSRS has a reset (recycle) time in which cleans its buffer. By default its setted at 720 min (12h) so thats why if you open a report every morning it actually load very slow.
As you need you can set the recycle time higher (2-3 days). I don't recommend a higher time because the buffer will fill up and you will get only blank pages, so you will have to manually restart Reporting Services.
It may be completely unrelated to SQL Server. Try to see if is not the code sign revocation list check issue, see Fix slow application startup due to code sign validation
I converted the main query and the dropdown controls on the page to load from stored procedures, it made a difference of say 5 sec in the loading process.Avoid using any inline queries.
Related
We recently upgraded our SQL server to 2016 and I turned on QueryStore to do the analysis that it provides. I'm encountering a problem where, even if the time period of the report is Last hour, it will generate a message that says "Couldn't connect to database" even when running it on the database server itself. Sometimes if I keep refreshing the report it will eventually display some data, but it's intermittent at best. I'm running SSMS 17.5 on a sql server 2016 server.
We are having a somewhat similar issue with another program that connects to the database where it will sometimes not be able to connect, but every time I run my queries in SSMS, run reports in SSRS, or even use activity monitor, I never see any connection drops, so I'm not sure if it is related.
Thank you in advance for any help!
I find it works fine with the statistic set to Avg, StdDev, or Total. Max and Min give the error.
I found this happens when the query store runs out of space and gets into cleanup mode.
In database properties in SSMS try playing around with Query Store settings: for how many days it stores the query stats and does it get into "size cleanup" mode. More info on how to keep it adjusted: https://learn.microsoft.com/en-us/sql/relational-databases/performance/best-practice-with-the-query-store?view=sql-server-ver15#Configure
Two month ago I deployed a new VM in Azure. I used the pre-configured "SQL Server 2016 SP1 Standard on Windows Server 2016" with 7 GB of RAM, and I chose the offered option to make backups automatically. Only other things I changed is add it to AD and put some databases (largest of ~2 GB size of backup file)
Now the server is running a service called SqlIaaSExtension.Service which I understand is for doing these backups as well as automated patching. You can find the services description here: MS service description
The problem is, it keeps on building up memory until after some weeks the SQL Server itself fails to execute larger queries. A restart of the SqlIaaSExtension.Service fixes the problem, but this is not at all a sustainable solution.
Does anybody know a working solution other then disabling the service and loosing the functionality altogether?
My setup (german):
I have meanwhile got some Information from Microsoft:
There seems to be an error in the SqlIaaSExtension.Service which is known to MS and will eventually be fixed.
Workaround is:
A: If you don´t need the functionality - remove this service, as indicated in the service description.
B: If you want to keep the functionality - restart the service periodically. Possibly automate via Task-planner.
Updated info from MS 19/07/2017: Error is identified and should be fixed in the next 7-10 Days. A mitigation is restarting the service if necessary.
Updated info from MS 31/07/2017: Error should be fixed in Version 1.2.19.0. This can be checked from the Azure Portal under "extensions" in the VM-Menu.
I know similar questions have been asked before...
I am using SQL server 2005, with SSRS 2005 installed on the same box. (aka. production DB, Report DB/TempDB, Database engine, and SSRS all in the same box).
We have about 200 reports deployed in the box.
SSRS/DB is running on a W2k3 64-bit VM.
Now the problem...
Occasionally almost on a daily basis our users get the 'operation timeout' error (error in XML document....). At first I thought it was a report size problem, but then when I try the Report Manager URL (http://<>/reports), nothing appears on the browser. The only thing I can do is to recycle the Report server IIS pool and it will work again. Everytime when the 'operation timeout' happens, the Report Manager URL will not work, and I can't find any logs in IIS to indicate there's a problem.
I researched on the net and found that some people have put a dummy report as part of the SQL server agent job which runs every 10 minutes from 9-5 to 'warm up' the SSRS. The dummy report made a small connection to the DB on one row from a very small table. The operation timeout problem seems to have disappeared for 95% of time, but it still happens. Strange enough, when the operation timeout problem happens, I notice the dummy report job has also stopped working. In this case, I had to recycle the IIS pool, and start the SQL server job again, and then SSRS will work again (until the same problem happens next time)
The error I got from the SQL server job is:
System.IO.IOException: Unable to read data from the transport connection: An existing connection was forcibly closed by the remote host
However I am totally confused by how the IIS issue on the report server somehow affects the SSRS job. Maybe I am on the wrong track but that's bizzare.
My observation so far is if it takes forever for the Report Manager URL (http://<>/reports) to appear it is a bad sign that something has gone terribly wrong on SSRS.
I have also added a new task which call the SSRS Report Manager http://<>/reports URL using PowerShell in order to 'warm-up' the IIS but it does not seem to make much difference.
Can someone point me to the right direction? Thanks. WM
In the past, after much research, I've found memory allocation for SSRS to be the root of many issues. You can try this.
Add the following into the <Service> node in the rsreportserver.config file
<WorkingSetMaximum>4000000</WorkingSetMaximum>
The file is typically in c:\program files\Microsoft SQL Server\MSRS11.iMIS\Reporting Services\ReportServer
This sets the maximum memory available for the report which also set the minimum memory to 60% of the maximum.
https://msdn.microsoft.com/en-us/library/ms159206(v=sql.110).aspx
I am using SQL Server Reporting Services 2012 and received this error without any known cause: The report execution eqaiekfzmk2snc55y0zrow55 has expired or cannot be found. (rsExecutionNotFound).
While I have found other posts describing problem through Google searches, the resolutions did not help me:
Restarting SQL Server, SQL Server Agent, and SQL Server Reporting services
Increasing the Execution Timeout through SQL Server Management Studio when connected to the Reporting server
Adding rs:ClearSession to the URL querystring (and trying IE, Chrome, and Firefox)
Redeploying after each troubleshooting step and retesting
I looked in the Reporting Services log file folder C:\Program Files\Microsoft SQL Server\MSRS11.MSSQLSERVER\Reporting Services\LogFiles but I see the datestamp is over two months old and I could see nothing related to the symptom.
I looked looked in ExecutionLog3 and did not see anything related to the symptom. use ReportServer; select * from ExecutionLog3;
To find out what did work, I verified that:
The query and results are sound, as seen in Management Studio
I can preview the report in Data Tools on the server
I can view the report when remoting into the server
I only see the error when viewing the page from outside the server. This is a relatively lightweight query and result set, so I cannot believe that this problem has anything to do with execution timeouts.
I changed the name of the file and redeployed. I am able to see that report now, but this isn't a true resolution because I still don't know what is truly causing the problem and how to fix it. If the symptom appears again, I can't keep changing the filename and redeploy.
Is there a way to get a better idea of what is happening? A specific log file or a property I need to change?
Update:
I thought I had this problem worked out, but apparently not. I found nothing useful in the error logs: only a restatement of the same error message visible in the browser. When I redeploy (using SQL Server Data Tools), the error goes away... for a few hours or until the next day, when I need to redeploy to make the error go away.
I know this is an old question but I had this problem recently and it turned out to be a bad session cookie. The cookies session-id matched the guid in the error message and once I deleted the cookie all worked fine after that. The report at one point had been configured to cache a temporary copy
but that had since been turned off (however, the problem existed before that had been turned off so it may not be relevant).
Hopefully this answer will help someone else save the hour I spent figuring it out in my environment :)
This might help someone.
In my case, The report url had trailing spaces (a silly mistake) which caused this.
I've added &rs:Command=ClearSession to the end of my url and works fine with me.
As stated in a different answer you can clear the session which usually resolves this issue.
If you have a question mark in your URL already then add the following to the end.
&rs:Command=ClearSession
If you do not have one then you need to add the following to the end.
?rs:Command=ClearSession
I just had this problem, it was for an existing report that had been working correctly. However, the Report Builder had been open for some time in another window while I was working on something else, and I hadn't saved my work (I was applying a filter, and didn't want to save my changes with my test filter). It occurred to me that since the report HAD been working, but it had been sitting idle, it might have gone stale. I opened the Dataset Properties, clicked Query Designer, then "Run Query". The Query Designer then got a fresh request from the data source. I closed the Dataset Properties window and clicked "Run", and my report was again displayed.
For me, I had no trailing space.
Some people had luck with clearing Session.Keys of "Microsoft.Reporting.WebForms.ReportHierarchy"
I solved it by Session.Clear in the global.asax
For us, the error appeared trying to run a report on an SSRS 2016 server using Internet Explorer 11. The user had created a bookmark that linked directly to the report. What may have happened: IE preserves cookies and temporary internet files for favorites to "help them load faster". The user may have initially ran the report, then created the bookmark to the report which contained session information.
To fix: Delete the bookmark, then cleared browser history in IE (CTL+SHIFT+DEL) being sure to uncheck "Preserve Favorites website data".
Provide Used: Microsoft OLE DB Provider for SQL Server. Can anyone help me with this..
I was trying to connect with LLBLgen
This MSDN blog page has some useful on this...
http://blogs.msdn.com/sql_protocols/archive/2006/12/02/understanding-kerberos-and-ntlm-authentication-in-sql-server-connections.aspx
In my case, I found the account was locked.
Reason was I previously, on another machine more than 3 times tried to login.
It did not recognise me - and tthen finally it locked my account.
Reopening account made all work fine.
br
Jan
The error you get is almost always caused by a problem with using Windows Authentication. Please try switching to a SQL server login (username/password), or make sure your current Windows login has access to the SQL server and database you're trying to connect to.
-Edoode
I fixed this by mapping a drive to the server running MSSQL. This seemed to generate some kind of trust that allows MSSQL to connect without this error even after a reboot.
I used to get this error sometimes when connecting to my local SQL Server with Windows Authentication. I never fixed it unfortunately - it went away when I reinstalled windows.
I think a reboot used to fix it - have you tried that? Not exactly the best solution, I know :P
Try to synchronize your date and time with the your domain's. The SSPI issue may be related to Active Directory authentication problems, some of them related to date and time changes. This is very simple to check and fix. Try it out!
There is a Microsoft KB article that addresses many of the reasons for this area (KB811889) at the following URL: http://support.microsoft.com/kb/811889.
A lot of Googling shows that one of the diagnostic steps helped most people who encountered the issue.
I recently had this exact issue where I'd get this error only when authenticating with certain accounts, but not others. Ultimately what was causing my problem was not mentioned in any KB or article I found on the net, but through trial and error I discovered that when the account used through SSPI authentication to SQL Server (2k8) happened to be in a large number of groups (in my case over 250) you would get the "Cannot Generate SSPI context" error. I suspect it has something to do with overflowing the security token that Kerberos uses and have seen similar strange authentication problems for user accounts in a large number of groups.
I get the problem when I have the time set differently on my client machine than either the server or the AD machine ( I was trying to test into the future).
Short Answer: Have you recently change the user the service is running as? Was there a system crash?
Long Answer:
I know this is old, but I want to post my experience that I just had.
We had spent hours Googling and found nothing that worked.
Eventually we ran across a set of actions that could cause this:
If you change the user that the Sql Server runs as (e.g. from Local System to a domain usr) and do certain updates and the server doesn't safely reboot -- you get this.
So, we set things back to Local System and bam it worked. Swapped it to the domain user, no worky worky. Ok. Swapped it to Local System, rebooted, swapped it to domain user, rebooted, bam -- worky worky. All was good in our world. Later that morning it crapped out again... still working on that now but the priority is changing and I'm not sure we're going to continue work on this problem so I wanted to post something in case this happens to someone else.
What caused ours was we did an update and, apparently, we learned that it's bad practice to let Sql Server run as Local System so we changed it to a domain user. We never rebooted, but restart the service. A month later, we do updates. We don't reboot. A month goes by and a power strip fries causing the server to have an unexpected shutdown. Yet another month later we find out problem because we rarely connect to this particular database (Interestingly, Sql Server 2008 worked fine... it was only 2005). Or... at least this is the best we've come across.
Our admin guy doesn't like Vista and likes to blame everything on Vista (refuses to let us test Windows 7)... so he Googled "sspi vista" or something like (I know it had sspi and vista, but it might have had another one... in case you need to Google it was well) that and ran across an article that pretty explained our scenario after we had a meeting we all remember these pieces and placed this picture together.
In my case, the time synchronization issue in the Windows 2003 domain environment was actually the issue.
This was quite easy to overlook as the two had been on two different time zones, whilst showing the same times on their clocks; which in effect was about 1 hour apart.
So other than the time on their watches, check the time zones as well.