Michael W. Ryan on Mon, 12 Jul 1999 12:20:47 -0400 (EDT) |
On Mon, 12 Jul 1999, Jason S. wrote: > At one point I was the admin of an E3000. 6CPU's, 2GB of memory. It was > running netscape enterprise server (ssl), apache (http), and Oracle 7. > It was a matter of hosting fee's, limited space and funding that made > one big box a better solution for us than several smaller ones. Despite > how it sounds, it performed rather well. There was only one incedent > that led to unscheduled downtime and that was due to a bad processor. It > ran for about 2 years. By that time the company had laid out a plan to > expand its site and I'd moved on. Not to argue against your point, but I think the 6 CPUs helped alot. > The point I'm attempting to make is that despite it not being ideal, you > CAN do everything on one box and expect it to work. If its slow, then > you can deal with it. But it should at least work. I do agree with you. I think the problem with MS is that once the performance starts to drop below a certain point, the system starts getting "frustrated" and just throws up its hands. My company experience an example of this with an internal file-server style database app. It lived at one site and was used by people at our other site. Our connectivity was a T1 (well, a frame relay, actually, but we manage to get full bandwidth around 90% of the time) through our ISP. The remote site started getting errors trying to find DLLs. The DLLs were there, but the accesses were timing out. Incidentlly, our solution was to run the app through a Terminal Server. Michael W. Ryan, MCP, MCT | OTAKON 1999 mryan@netaxs.com | Convention of Otaku Generation http://www.netaxs.com/~mryan/ | http://www.otakon.com/ PGP fingerprint: 7B E5 75 7F 24 EE 19 35 A5 DF C3 45 27 B5 DB DF PGP public key available by fingering mryan@unix.netaxs.com (use -l opt) _______________________________________________ Plug maillist - Plug@lists.nothinbut.net http://lists.nothinbut.net/mail/listinfo/plug
|
|