SharePoint.H@ck

Went to a customer who had setup Kerberos but the RSS Viewer Web Part was still throwing the “does not support authenticated lists” message, even though Kerberos was setup 100% correctly.

Turns out the issue was with the FQDN of the sites, they had http://portal.company.co.za as their default mapping which RSS does not seem to like, once I added an internal default mapping to point to http://portal the RSS Viewer worked 100%. So the customer still access SharePoint using http://portal.company.co.za but SharePoint communicates with http://portal.

Not sure what the exact issue is with the FQDN, suspect if might be something to do with communication between servers; will let you know when I find out.

Regards,

Doug “bobTheBuilder” McCusker

*insert spooky music*

Ok so here is one that truly shows that I have no social life: a customer once asked me “why can we specify 1 – 30 usage log files, when configuring usage analysis processing?”, the answer to which at that time was “not entirely sure, but I know if you hit 30 it will stop recording usage data" – at least that is what the docs say :)

So, I left it for a while but then a mate asked the same question and days of working with WSS 2.0 started itching in my brain, this is what I found and it is still relevant:

Taken from the following document: http://office.microsoft.com/download/afile.aspx?AssetID=AM102437481033 (modified to be relevant to WSS 3.0 / MOSS 2007):

“

Inside this folder is a folder for every virtual server named using the Windows SharePoint Services virtual server globally unique identifier (GUID), and under those folders, a different folder for each day will be generated. The path of these folders cannot be modified.

You can configure the limit of logs to be created on a daily basis, with a maximum number of 30 log files. If you set a maximum number of n log files, this applies to each virtual server. This means that the log folders of each of the virtual servers (web applications) will contain at most n log files. In other words, if this number is set to 1 there will be 1 log file for each virtual server. Having many virtual servers (web applications) and having many log files might reduce performance during logging. You should consider increasing the number of log files if a front-end Web server (standalone or member of a Web farm) has log files with more than a million entries.

The front-end Web server might not have enough memory to memory-map a really large log file, which leads to a situation in which the log file might not get processed. Each hit that a front-end Web server receives uses approximately 200 bytes (B) in a log file. As a result, approximately 200 megabytes (MB) of RAM are used to memory-map a log file that contains a million hits. Memory mapping occurs only for several minutes during usage processing. Because log files are processed serially, when you have several log files, a smaller memory footprint results when a log file is processed. 

Windows SharePoint Services logs HTTP 2.x information to the log files and does not log HTTP 3.x or HTTP 4.x information to the log file. When you use multiple log files, the log files are created at the same time, and all hits from one website are contained in the same log file. Usage data for a website is updated one time each day.

The number of requests that are sent to the back-end server during usage processing is proportional to the number of websites on the server. However, the memory footprint on the back-end server is not affected by the number of websites. The additional load that usage processing generates does not significantly affect the performance of the back-end servers on the server farm.

”

Regards,

Doug “bobTheBuilder” McCusker

The other day I ran into an issue where the admins were seeing an Event ID in the application log that indicated that a content deployment job had failed. Well the actual details were something like this: content was copied successfully to the destination server, but could not find the web app to import into.

Now, content deployment jobs in this scenario had been stopped and deleted a few months back and the admins did not know why this was happening, what I found was the following:

• There was still a content deployment file in the content deployment folder of the destination server.
• Even though content deployment is turned off, if there is a file in the folder SharePoint will try and try to import the contents.
• The error was always after the default admin app pool recycle period.

Just though you would be interested. Remember to delete this file when cancelling content deployment jobs!

A couple of things I have been noticing is that a lot of content ends up in the TEMP directory of the MOSS service accounts (think content deployment, workflow etc.)

We all know that the default TEMP directory (if unchanged) is in the %userprofile%\AppData\Local\Temp directory which lies on the %system% directory, what I have been doing during setup of MOSS is to ensure that this variable is changed so that any TEMP data is stored on a separate disk.

You will be surprised at how big the TEMP directory can get!

Regards,

Doug “bobTheBuilder” McCusker

Currently at the 2009 SharePoint Conference, watch this blog as I will be posting new content about SharePoint 2010, mostly around how to do things I did in 2007 in 2010 :)