Hello,
I'll try to put as much as info in the first place:
I had a SPF 2010 with searchexpress on a physical server. There were 3 web app (CA, MainWebApp, DEVWebApp) and 2-3 site collections in each of the last 2 ones.
We got here anew infrastructure based on Vmware Esxi 5.1 and NetApp for the storage. I choose to create 2 VM for a new SPF2013 farm:
- Server 1 : App, WFE, Search, ... 4 vCPU @ 2.50 Ghz, 8 GB RAM, 50 GB C: (pagefile.sys 12GB). (I know about the 80 GB recommended, I'll extend the partition later).
- Server 2 : SQL Server : 8 vCPU @ 2.50 Ghz, 8 GB RAM.
The upgrade process went fine; I follow the PDF provided by MS in 3 main steps.
Now the problem: the delay in each sites for navigating between the pages is very slow, around 6 seconds each time (except the first time but this is the way .net works with aspx). I did this test:
Going to the home site of a site collection clicking on the sub site containing a doc lib with 3 web parts displaying docs from the doc lib (less than 50 items)
It took 6 seconds.
I have run a SQL Profiler in the mean time :
edit : copy paste the image into paint for being able to see it correctly :)
The content DB associated to the site I was navigating in isWSS_Content_IT. The selected lines are the ones that seem to retrieve the data (I've executed few of them and I got the expected results in SQL management studio). We can see that the first event is starting at 14:55:47:280 and the last one at 14:55:47:440. That is 160 ms for retrieving the data (I hope I'm not wrong with my profiler trace...)
My second observation is withfirebug in FireFox:
edit : copy paste the image into paint for being able to see it correctly :)
Here I can see that my browser wait 6 seconds for thedefault.aspx of the site.
Now I'll give you more info about myenvironment to filter out the hints you might have for me :)
My server 1 is about 50-70% ofRAM utilization.
My search settings:
The five noderunner.exe are running quite fine.
I have 2 content sources for my sites withcrawling continuous activated. The lasts crawl duration are about 3 minutes. The full crawl is enabled each day in the middle of the night.
I have alreadytest to stop the search services, I still have the delay.
SQL DBs:
The biggest DB I have is the WSS_Content_IT +- 900 MB. The recovery model is set to full for all DB's.
I don't have custom code onto the farm.
I think this is all I can explain here. I hope you will have some hints for me :) thanks in advance!
Greg :)