Have an old MOSS2007 "mammonth" on SQL 2008, which have performance issues.
IIS often crashes and gives "Unknown error".
It is a two-tier Farm with Database on one server and WFE+App.Server on the other. Hardware is the same:
About 1'800 users registered, size of Content DB is about 470Gb. Databases are well separated according best practicies (databases are stored on different LUN's, MDF and log files are separated too).
There are 4 web applications:
- Central Admin,
- Personal Sites,
- Shared Services administration,
- Root web application, example http://sp.
This root site collection contains about 10 site-collections: http://sp/blabla/another_site_collection1,2,3
.. and most of them have few webs and workspaces, (in total not more than 150).
Search is configured in Shared Services Provider, and there are about 1'270'000 elements crawled.
I used
%stsadm.exe% -o enumsites -url "http://sp" >> c:\site_collections_size.txt
command and found that Root site collection and one other has more than 100Gb:
<Sites Count="43"> .. <Site Url="http://sp" ..
StorageUsedMB="112567,5" StorageWarningMB="0" StorageMaxMB="0" /> ..
<Site Url="http://sp/blabla/blabla" .. StorageUsedMB="124295,8"
StorageWarningMB="180500" StorageMaxMB="200500" /> ..
What I detected is two problems:
I think 8Gb RAM for WFE could be Ok, but 8Gb for database is extremely small.
And the second problem I think is Search - the additional crawl runs for 49 hours, and it is still running non-stop. And also there is configured Full Crawl every week. Configuration of additional Crawl is strange: "for 24 hours, every 8 hours"
So I think Search is configured wrong and it hangs the system. I think I need to turn of Full crawl (it is only necessary on first time, as I know), and I'll set additional crawl to night time. What do you think?