I'm having a lot of trouble tracking down the cause of the strange page file activity I am seeing on my server. In performance monitor, I get huge pages/sec peaks every second or 2 (not precisely regular but on average). The average pages/sec value is around 500, the peaks are in the thousands with little or no activity in between. Processor load is mostly low during all this, although processor 1A (multithreaded) is showing similar peaks and trough patterns to the page file and disk write activity. Available memory is high - of 1.5GB I have 650 MB available. PF usage is under 50% (898MB of 2302MB allocated) When major file access operations occur, paging also goes very high, with disk queues of up to around 8, and the whole thing slows down massively. But this regular paging is almost all writes, and according to sysinternals diskmon it's all to the same sectors of the disk (each peak writes to the same range), a series of 128-byte writes. I have stopped every service that I can without bringing the server down completely. I don't know if I'm focusing on this weird paging when I should be worrying more about the horrible slowdowns and massive paging that happens when you actually access the server in any way, but the symptoms would be easier to see when they stopped... but they haven't yet. This server is a lone domain controller and also has Exchange 2003 installed (but not fully implemented - it's not gone live yet, just has 2 test mailboxes on it and no mail flows yet). I'm also running Sophos enterprise console, WSUS, a helpdesk application (uses java), print services, and NetVault backup software... it sounds like a lot but it's only for 50 users and in my previous job we had a similar server doing everything for more stations and users (but not exchange...). It's all on 1 physical disk (SATA). It is intended that another disk be added which will obvioiusly alleviate some of the symptoms but the root cause will not have been solved. I don't know where to look next, and wondered if at the very least someone might know what I could look at next or who else i might ask! I expected a lot more from this server, to be honest, since it's not apparently 'under load' under normal circumstances unless one person copies a large batch of files from one place to another and it seizes up - it appears to have more problem doing that simple thing than my 800MHz workstations do. Please help! Thanks in advance... -Dev Lunsford

That is too much for a single disk in my opinion. At the very least you should be mirrored which will double the write abiltity. Better yet would be raid10 [not to be confused with raid 0+1 which is also better than a single disk. With a single disk you have file read/writes system read/writes pagefile read/writes AV reads mostly with some writes Exchange just hanging out WSUS read/writes and all the rest with all of these using memory. I will bet this sata is a large drive also which can increase access times due to stacked up requests. There is a simple experiment you can do to test if its disk access competition. Put in another smaller spare ide drive and move the pagefile to it. Then monitor pagefile usage. On another note, 898meg is a high to me for pagefile usage. But then I max out [4gig] all of my server mainboards. Cost of ram is cheap these days and more is better. Golly gee wilerkers everyone. Learn to Internet Search

You really should put Exchange on a separate server, especially when it gets going, and you have this current server providing so many services. Please help survivors of Hurricane Katrina. www.redcross.org