DEngle's SharePoint Blog

IIS 7.5 32-bit App Pool

dengle — Thu, 05 May 2011 19:07:00 +0000

When running an app as a 32-bit app in IIs 7.5 you may experience the app pool stopping every time you try to browse the site. We kept getting a 503. I added the app pool identity to the framework directory with read permissions and this did not help. It turns out that the Framwork64 machine.config had the Users group with read rights. The Framework\v4.xxxx machine.config had explicit permissions only and was not inheriting anything. I set it to inherit which gave read rights to the framework directory. This fixed the issue and the app pool remained started.

SQL Server Resource Throttling

dengle — Tue, 29 Mar 2011 16:00:49 +0000

I ran across a SQL 2008 server recently that from day to day was having issues with memory allocation to several SQL server instances. This is a development server so the load is varied depending on the work being done that day but we do have several other SQL servers and clusters that this information will also pertain to. Here is a summary of how I fixed the memory allocation to these instances to stabilize the servers:

If the SQL server has multiple instances check the ‘memory’ setting inside management studio for each instance. Default is 2147483647 MB. So if you have multiple instances they will each be set to this. The problem comes in with how SQL allocates a buffer pool of memory as its resource consumption increases, or perhaps better stated, the lack of it releasing those resources once it’s done. What we see is when memory allocation increases due to SQL processing transactions, the database engine will increase the buffer pool, but once the load decreases the buffer pool holds onto that memory. Typically when you look at the resource consumption on a single instance server you’ll see that single instance using almost all the memory on the server, typically 99% of it. So when you have a multi-instance SQL server those instances fight for that memory. If instance A happens to grab 80% before instance B can get what it needs then you only have 20% left for instance B. Until the SQL service is restarted you will never release that memory for other processes. I found this to be the case on this particular server where I have an instance for MOSS 2007 and an instance for SharePoint 2010. What was happening was since the utilization is currently higher on MOSS 2007 that instance was allocating almost all the memory on the server for its sqlserver process and leaving none for the SharePoint 2010 instance. The SharePoint 2010 instance was left running with very little memory, whatever scraps were left, and when it really needed to ramp up it could not, and was choking. Furthermore it was choking other processes on the server to the point you could hardly log in let alone run management studio (I had to run it remotely). So what I did was increase the memory on the server to 64 GB from 32 GB. But I also subtracted 8 GB from that total to leave for the OS, and then set the max memory on each instance to be 28 GB. I checked it the next morning and the MOSS 2007 instance is using all 28 GB I allocated for it, as I suspected, and the SharePoint 2010 instance is using about 6 GB. Now should the SharePoint 2010 instance need to run some heavy queries it will have the memory needed to ramp up its buffer pool and run the transactions, whereas if I had not limited the MOSS 2007 instance it would have ran its buffer pool all the way to the limit of the server RAM total, 64GB. Meanwhile you can log onto the server with ease and run other utilities including management studio without issues. The best part is these values can be changed on the fly in SQL without a reboot. My plan now is to monitor these instances and the server itself and throttle down the memory allocation on each instance to a manageable level and decrease the memory allocated to the server to reclaim some of those resources.

Just something to keep in mind when the call goes out next time for “MORE RAM!”. Contrary to popular belief, we have to pay for that stuff.

Beware of -stsadm -o deleteconfigdb

dengle — Tue, 29 Mar 2011 15:51:46 +0000

Our scenario was this: we moved our farm to new hardware. We followed this TechNet article on how to move the databases and then use cliconfg.exe to set up a SQL server alias on the farm machines. What we noticed was that the config db was still referencing the old SQL server when we looked in central admin. So we followed another article on running the deleteconfigdb operation and then the setconfigdb operation, but did not test it on a development environment first. The deleteconfigdb operation completed successfully, however, the setconfigdb operation did not. We could not get our renamed server’s config database to attach. We ended up having to rebuild our farm and installing all features again. This was a long arduous task just to get the config db server name updated in Central Admin. And in the end, the server name was never updated. It still references the old server. I do not recommend running these commands to accomplish this and before running this command at all I recommend you run it in a test environment to make sure it will work for you.

Submitted Email Records not flowing through Records Center in MOSS 2007

dengle — Tue, 29 Mar 2011 15:41:53 +0000

We experienced an issue where submitted email records were being sent to the records center but not being picked up by the timer job and passed over to the various libraries we have configured with routing rules to sort the records. These email are being sent from Exchange through managed folder rules and are tagged with a Label via the managed folders. We were getting the records in the email enabled doucment library and the meta data was attached but the records were hung there. We noted that the records center processing job was also not started and we could not get it to run.

RESOLUTION: We ran the following command- stsadm -o activatefeature -name RecordsManagement -force . This started the records center processing job and the submitted emails records started processing and dropping into their correct libraries in the records center.

401 Unauthorized when editing data source in SharePoint 2010

dengle — Tue, 29 Mar 2011 15:37:12 +0000

When we try to edit a data source on a SharePoint 2010 farm with SSRS running in integrated mode we receive a 401 unauthorized error. In the ULS logs we see this:

Exception encountered for SOAP method GetSystemProperties: System.Net.WebException: The request failed with HTTP status 401: Unauthorized.

at Microsoft.SqlServer.ReportingServices2010.RSConnection2010.SetConnectionProtocol()

at Microsoft.SqlServer.ReportingServices2010.RSConnection2010.SoapMethodWrapper`1.ExecuteMethod(Boolean setConnectionProtocol)

RESOLUTION: In central admin we changed the reporting services to run as trusted account from windows account. This causes SSRS to pass the service account for credentials instead of the currently logged in user.

Thousands of .en-US.resx errors in SharePoint trace logs

dengle — Tue, 29 Mar 2011 15:33:45 +0000

We have been plagued by thousands of entries in our SharePoint MOSS 2007 trace logs that look like this:

#20015: Cannot open “Proj.en-US.resx”: no such file or folder.
(#2: Cannot open “Proj.en-US.resx”: no such file or folder.)
Failed to open the language resource keyfile Proj.
#20015: Cannot open “Proj.en-US.resx”: no such file or folder.

This is an operation looking for a language file that does not exist. The problem lies with installing only the english language version of SharePoint. Not that this is a problem, but some features are written to access language-specific versions of files. Try running this from a command prompt to fix it. It cleaned up most of mine:

cd “C:\Program Files\Common Files\Microsoft Shared\Web Server Extensions\12\Resources”
copy avm.resx avm.en-US.resx
copy compproc.resx compproc.en-US.resx
copy ep.resx ep.en-US.resx
copy er.resx er.en-US.resx
copy hlpdsk.resx hlpdsk.en-US.resx
copy itteam.resx itteam.en-US.resx
copy proj.resx proj.en-US.resx
copy pws.resx pws.en-US.resx
copy rer.resx rer.en-US.resx
copy tsa.resx tsa.en-US.resx

hiberfil.sys on my SharePoint Servers

dengle — Tue, 08 Jun 2010 15:36:31 +0000

I have several Vmware virtual servers that my QA farm runs on. I noticed today that there was a system file on the root of the C: drive titled hiberfil.sys, and the size exactly matched my pysical RAM. I suspected this was a hibernation file but why would a server need to hibernate. These are all 2008 servers and I have never enabled this. I checked the advanced power options and nothing was checked to enable this. I tried deleting the file and it would not. I did some searching and came across an article about using powercfg.exe /h off to get rid of the file. I ran this command on all 3 servers and it instantly removed the hiberation file.

Exception from HRESULT: 0×80041054

dengle — Tue, 25 May 2010 17:03:50 +0000

We were seeing this on one of our web applications on our external SharePoint 2007 farm. The internal URL was ok. Only the external anonymous access mapping was affected. I found another blog that noted that the antivirus could be effecting the content DB of this web app so I accessed Central Admin for this farm and went into operations > anti-virus. I unchecked all boxed and clicked OK and then went back in and rechecked the boxes that were checked originally. The site was now OK.

Unknown SQL Exception 8117 occured

dengle — Thu, 20 May 2010 19:26:53 +0000

UPDATE: This exception turned out to be from a application pool recycle happening too often. We had the max memory and fixed number of requests recycle paramaters set to a low enough point where the application pool was recycling so often it caused these SQL exceptions. Once we removed those parameters the exceptions stopped.

Original Post below:

We started getting this message on our front-end on the 9th. I don’t think its effecting anything but I will keep this post updated as I find out any new info. Below is the complete details of the error from the application event log:

Log Name:      Application
Source:        Windows SharePoint Services 3
Date:          5/9/2010 4:19:06 AM
Event ID:      5586
Task Category: Database
Level:         Error
Keywords:      Classic
User:          N/A
Computer:      SMMOSSFE02.smcorp.steinmart.com
Description:
Unknown SQL Exception 8117 occured. Additional error information from SQL Server is included below.

Operand data type nvarchar is invalid for sum operator.
Operand data type nvarchar is invalid for sum operator.
Event Xml:
http://schemas.microsoft.com/win/2004/08/events/event“>


    5586
    2
    484
    0×80000000000000

    114367
    Application
    SMMOSSFE02.smcorp.steinmart.com


    8117
    Operand data type nvarchar is invalid for sum operator.
Operand data type nvarchar is invalid for sum operator.

Hello SharePoint world!

dengle — Thu, 22 Apr 2010 18:28:07 +0000

After working with SharePoint since the early 2000′s (SharePoint Team Services, SharePoint Portal Services), and coming across numerous break/fix issues, cool tricks, and weeding through page after page of other blogs, I figured it was time for me to contribute to the SharePoint blogging world with one of my own. These are my real world experiences using and administering multiple SharePoint farms in my career. These include SharePoint Team Services, SharePoint Portal Services, Windows SharePoint Services WSS 3.0 together with Microsoft Office SharePoint Server 2007, and the long awaited SharePoint 2010, Office 14 suite. You of course will find tidbits about SQL Server Reporting Services, Performance Point Server, Internet Information Services, Windows Server 2003 and 2008, and more. I will try to be up to date and accurate with my posts, but feel free to correct me if I overlook something, or speak out of turn. Please comment as you see fit regarding my posts. I look forward to this becoming my own little SharePoint knowledgebase as well.

Welcome!