Sunday, December 28, 2014

fixmlindex.exe Process Problems in FAST Search SharePoint 2010

As you can tell from my blog, I have had many issues with my FAST search service. This time I ran into an issue where the fixmlindex.exe process was runaway. This memory leak caused issues with my SSRS service. It doesn't directly impact SSRS, it's just that for me I have these services on the same app server in my SharePoint 2010 farm. So when this process ran away, it affected my SSRS instance which also uses a ton of memory. FAST and SSRS were competing for whatever memory they could get.

Fixing this issue isn't too bad. All you need to do is stop and start FAST. Use the process I outlined in my post Properly Stop and Start FAST Search for SharePoint 2010.

Sunday, December 21, 2014

Large Sprel Log FAST Search SharePoint 2010

I had an issue where my FAST server was constantly running with high processor utilization for sprel.exe. Additionally, I had run out of space on my C: drive on the server where my FAST index was. I used the tool called SpaceMonger to determine what was taking up so much space on my drive. Turns out my sprel log file was enormous.

I performed the following steps to get the sprel file and process back in order:
  1. Stop the FAST search service (see Properly Stop and Start FAST Search for SharePoint 2010).
  2. Delete the sprel file at C:\FASTSearch\var\log\sprel.
  3. Started the FAST search service
  4. Run the following in FAST Search Management Shell:
  5. nctrl stop sprel
    nctrl start sprel
  6. Check the log files in: C:\FASTSearch\var\log\sprel and verify the new file created.
This solved my space and runaway service issue.

Sunday, December 14, 2014

Clear People Search Index in FAST SharePoint 2010

When you install FAST Search in SharePoint 2010, you get two main connectors: FAST Search Connector SSA and FAST Search Query SSA. The first is used for indexing your site contents. The latter is for your people search.

Every so often you may need to reset these indexes. If you need to clear the index for the FAST Search Connector SSA, you can find the directions for that on my post Stuck Index for FAST Search in SharePoint 2010. The directions just for resetting the index are at the bottom of the post.

Stop Crawls

Before you do this, make sure you stop any crawls on your People Content Source. You can find this by going to Central Administration > Manage Service Applications > click on FAST Search Query SSA. Click on Content Sources in the left navigation under the Crawling section. Then use the dropdown menu to stop a crawl if it is running. If a crawl is scheduled to run soon, you may want to edit the content source and set the crawl schedule to none so it doesn't interrupt your index reset.

Reset People Index

Resetting the index for People is very easy. Go to Central Administration > Manage Service Applications > and click on FAST Search Query SSA. In the left navigation under Crawling, look for the link Reset Index. Click on this link and then click the "Reset Now" button. When you click this, you will reset the People index. There will be no search results for People search so make sure you do this during an off-hours time in a production environment.

Crawl Content Source

Then click on Content Sources and choose to do a Full Crawl on your People content source. If you have a lot of profiles, it may take a while to crawl. Search results will show up as they are crawled.

Sunday, December 7, 2014

The number of Distributed Cache hosts in the farm exceeds the recommended value (SharePoint 2013)

After I set up my SharePoint 2013 farm, I received an error from the SharePoint Health Analyzer: "The number of Distributed Cache hosts in the farm exceeds the recommended value." When you set up a new farm, the default is that the distributed cache will deploy to all the servers in the farm. You should remove the distributed cache from all app servers. This leaves the distributed cache on the Web front ends (WFEs).

The user that is going to remove the distributed cache needs a lot of privileges on the server and the database, so I performed these actions under my SharePoint install account. The account needs the following rights:
  • dbowner and securityadmin on the database
  • member of local administrators on the server
  • farm administrator
Now remove the distributed cache from your app servers:

Start the SharePoint 2013 Management Shell (run as administrator if you haven't turned off UAC) and run the cmdlet:
Remove-SPDistributedCacheServiceInstance

Finally, return to Central Administration, go back and choose to reanalyze the Health Analyzer rule again and this will remove the alert.

References
http://technet.microsoft.com/en-us/library/jj891124(v=office.15).aspx

Saturday, December 6, 2014

Properly Stop and Start FAST Search for SharePoint 2010

You may have come across the command "nctrl stop" and "nctrl start" to stop and start FAST search. This command is generally better used for small environments. For larger, multi-server environments a more sequential shut down should be used to prevent errors. Here are the proper steps to stop and start the FAST search indexer. Use FAST Search Powershell and run as administrator.

To stop, follow the following process:
  1. Stop all crawls and ensure none will try to start.
  2. nctrl stop contentdistributor
  3. nctrl stop procserver_1
  4. nctrl stop qrserver
  5. indexeradmin -a suspendindexing
  6. Watch the fixmlindex process under Services for each partition. Give it about 5 minutes to settle into an idle. It won't fully stop.
  7. nctrl stop indexer
  8. nctrl stop search-1
  9. nctrl stop
To start, run the process in reverse:
  1. nctrl start
  2. nctrl start contentdistributor
  3. nctrl start procserver_1
  4. nctrl start qrserver
  5. nctrl start indexer
  6. nctrl start search-1
You can check the status periodically by using the command:
nctrl status

References:
http://support.microsoft.com/kb/2529030