For the last few weeks we've been experiencing challenges with our scheduled full crawl of our Sharepoint 2013 site.
Crawl is set to begin at 1 AM on Saturday morning. It will run for about 3 hours and then the server will just lock. We are running Sharepoint on a 2012 R2 server with 24 GB of memory with 8 cores assigned to it. While the crawl is running, memory is usually pegged at 99/100% and memory at 50/55%.
SQL is also 2012 with similar specs just more memory. SQL processor rarely goes above 65% during this process.
A reboot of the Sharepoint server will bring the server back and it resumes the crawl. After about 3 or 4 crashes it's fine for the week. Incrementals are working great.
Looking at the crawl history, I see latency running around 120-130 for incrementals and balloons to 300-400 on fulls.
We have roughly 840,000 files in the crawl store. As I inherited this setup from someone else, there are some updates that need to be done to Sharepoint and SQL as well as some health analyzer issues (i.e. content databases too large).
I dont see any *clear* problems other than possibly the SQL server is bogged down with drive speeds but I would presume to see other problems if that were the case.
All this said: help. What should I be checking here? Really tired of spending my saturdays babysitting this server!