On our LAN we have a Metaframe XP 1.0 FR3 farm consisting of 15 servers of which one is the dedicated datacollector. 10 servers are Windows 2003 standard and are configured identical (containing applications like the Office 2003 suite, SAP and Agresso). Then we have 4 servers which are running Windows 2000 SP4. These servers contain either Hyperion or i-CRM. It's the Windows 2003 servers which are giving us problems at the moment. All hardware is from IBM (x345), containing dual 3 GHz CPU's and 4 Gb RAM. Since we installed SP 4 for Metaframe on the 10 generic servers, the IMA service randomly stoppes on several, if not all servers. The only eventviewer entry we can find is a 7036 entry which states that NT AUTHORITY\SYSTEM has sent the IMA service a stop control. The IMA service most of the time stops somewhere between 11:00 and 11:20 AM, but isn't consistent in that behaviour. All servers were up to date with pre SP4 Metaframe patches and are up to date regarding MS patches. I did notice that all servers are quite busy the last couple of days. An average user load of between 60 and 70 isn't out of the ordinary and peaks of 90+ aren't as well. Lately we copied the advanced load evaluator several times and added schedule rules, so that we were able to implement a reboot schedule.
Anyone have other suggestions?
Robert Elsinga =8-) Sintjohannesga, the Netherlands
e-mail: robert at elsinga.org i-net: www.elsinga.org
Phone: +31 871 901111, fax +31 842 22 76 82