The things I've looked for regarding performance issues:
* Speed/duplex settings on your NICs - I don't have any hard numbers on this, but I've seen in the past (not necessarily with Remedy) where if the NIC's and switch port's speed/duplex settings don't match or if they're set to auto/auto, explicitly configuring these settings to match (e.g. 1Gbit/Full) on both the NIC and switch port have yielded HUGE throughput improvements. * Use the SQL profiler tool to set up a trace that logs all transactions that take more than N amount of time to complete. I generally use 1000ms when I do traces like this, as it's my belief that no single SELECT statement should take more than a second to complete. You can use the output of this tool to find queries (SELECTS are the easiest to analyze) that run long, and see if you're perhaps missing an index on an oft-searched column. * Use the Windows Performance monitor on the SQL server itself and get some stats on memory usage, and especially disk I/O (average disk queue length for individual disks is the most helpful). Slow disk I/O is second only to lack of available RAM in its performance-killing effectiveness. In past experience, I found that having even a handful of heavily used tables all on the same disk can cause an I/O bottleneck. I've also found that having even two or three heavily used tables and those tables' indexes on the same disk hurts too. * Of course, turning on API and SQL logging and running that through arloganalyzer doesn't hurt either if nothing else is telling you anything interesting. This tool would help you determine if you have enough available threads configured, for example (e.g. a high thread wait time would indicate you don't have enough threads, etc.). Mike From: Action Request System discussion list(ARSList) [mailto:arsl...@arslist.org] On Behalf Of Tim Rondeau Sent: Thursday, February 18, 2010 12:43 PM To: arslist@ARSLIST.ORG Subject: Performance issues ** Arsystem 7.1 patch 7 ITSM suite, using SLM, CMDB, Change Management, Problem and Incident We are running SQL 2005 on windows 2003 server, plenty of CPU/disk and RAM. Looking for some feed back on how some have handled performance issues. We process about 400 changes a day and about 800 incidents a day. Just seems like this system can't handle it. DB timeouts, white screens all the time. DBA of course states no issues. In March we are moving to Oracle RAC, which will be 2 robust db servers. APP will still be on windows. Thanks Tim _Platinum Sponsor: rmisoluti...@verizon.net ARSlist: "Where the Answers Are"_ _______________________________________________________________________________ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org Platinum Sponsor:rmisoluti...@verizon.net ARSlist: "Where the Answers Are"
