SharePoint Search Crawls : Incremental Cawls for Day to Day living

I was at Microsoft the other day talking to some of their PFEs regarding all things bright and SharePoint. The topic of search crawls came up. I received some information which I wasn't aware of which I thought I would log.
As most of us are aware SharePoint Search can conduct a "Full crawl" and an "Incremental crawl". What most people do (myself included in the past) is schedule a full crawl late evening / early morning and schedule incremental crawls at regular intervals throughout the day. To be honest I'm not sure where this habit came from. I guess it just seems to make sense......except it doesn't.

A scheduled Incremental crawl is sufficient for day to day crawling. There is no need and as far as MS have informed me, no advantage in running Full crawls on a regular basis. Full crawls are only required if a crawl rule or crawl property has been changed (there might be another couple of situations - but these are config changes).

All changes are stored in a change database and this is what the incremental crawl looks at to see what it needs to update. The changes database automatically clears down so you don't need to worry about it getting too big. There is one caveat and that is that the change database does clear itself after a certain amount of time so you have to make sure you do your incremental crawl within this time. This period is in the region of 30 days so assuming that you are going to schedule incremental crawls more frequently than this you are ok.

So, for day to day search indexing, only schedule incremental crawls.

"Access denied" to Site Collection administrator

This post is to remind me of something which is a "school boy error" in my world of SharePoint. I had been handed a farm which was in a mess and I was attempting to get it up and running. One of the problems I had was I kept getting "Access denied".

• I got access denied when I accessed that root of the site. http://site/
• I got access denied when I accessed the settings page http://site/_layouts/settings.aspx
• I got access denied when I tried to modify any file in SharePoint designer. 

I was defiantly a site collection administrator.

Then the sun started to shine. It turns out that the site collection had been locked when I checked in "Site collection Quotas and Locks". Now why didn't I think of this sooner. I should have known better.

Replacing Central Administration on a corrupt server

I was handed an issue whereby one of the SharePoint servers in a farm had fallen apart. For reasons I won’t go into now, this application server had been rolled back so that all that remained was a clean windows 2008 build. There was not a hint of SharePoint on this poor thing. The task was to reintroduce this server to the farm.

This sounded straight forward at first but was made a little more complicated by the fact that central administration was on the machine which was lost so I was unable to make any farm alterations this way.
The farm in had two web front ends and cluster sql boxes. The dead application server waited to join the farm.

My approach was as follows:

1. Back up existing farm settings and content - Log on one of the web server and Back up IIS websites, 12 hive and main site collections.
2. Run SharePoint products and technologies configuration wizard – Run wizard on web server and configure to install central administration on this web server.
3. Remove redundant server from farm – Navigate to central administration and remove the application server from the farm (via operations -> Servers in farm -> Remove server). This updates farm so that there is an accurate representation of server architecture. 
4. Install SharePoint on Application server and connect to farm - Log on to application server and Install SharePoint and connect to existing farm.
5. Run SharePoint products and technologies configuration wizard – Run wizard on Application server and configure to install central administration on this application server.

Hey presto, you have a central administration up and running again on the correct server.

RSS feed stopped providing content

I was passed an issue whereby RSS feeds had "stopped working" for an Internet facing website. On closer inspection I found that the web site was using feedburner to provide their feeds. After getting the details from feedburner I was able to work out that the SharePoint feed was coming from:

http://www.website.com/_layouts/feed.aspx?xsl=1&web=/news&page=xxxxxxxx-yyyy-zzzz-qqqq-wwwwwwwwwwww&wp=xxxxxxxx-yyyy-zzzz-qqqq-wwwwwwwwwwww

I called this page and surprise surprise it returned blank - feeds not working.

Looking at the query string for the requested page suggested the following:

• The feed related to the '/news' site.
• The feed related to a specific page denoted by the guid.
• The feed related to a specific web part denoted by the guid.

After navigating to the '/news' sub site, I was able to see that there was only one page and on this page there was a content Query web part which was RSS enabled. This web part was not returning any results within the page. When I clicked on the RSS link I was taken to the same URL referenced by the feedburner so I had worked out where the RSS content was coming from.

I edited the web part and found that it was pointing that a sub site which no longer existed. Ergo, no results were displayed and therefore there was no rss content for feedburner to index.

After the CQWP was reconfigured to pull content from a site which existed, the RSS feed populated itself and everyone was able to get on with their lives. :)

Very simple really.

Backing up live site collections - STSADM backup without lock

I required a back up of a production site collection today. I usually take backups OOH but for the task I was conducting I wanted a quick backup so that I could do a fast restore and run a few tests.

The default stsadm backup operation (after service packs installed) will set the database to read only which is obviously not a good idea for production environments. Fortunately there is a switch which doesn't look the database which was exactly what was required in this case.

stsadm -o backup -url http://mainsite/ -filename d:\backup.bak -nositelock