Tony, I'm in a similar situation right now. I am testing a squid cache that will be deployed for a simliar number of users and similar hardware (eerily similar...). I'll take a stab at answering your questions based on my research over the past few weeks. If anyone has any corrections, please feel free to make them as I would like to know for myself as well.
> The hardware: HP ProLiant DL360, 3 GB RAM, dual 2.something > gig Xeon CPU's, dual 72 GB 10k SCSI drives. I can hardware > RAID the disks, but I'm not sure I want to given the massive > amount of disk activity this box is destined for. (Or do I?) > I only have this one box to work with at the moment. Consider RAID 0 for the cache partition for maximum performance. Apparently this is where you will get the most performance bang for your buck (the disks). > The people: anywhere from 500 to 2,000 concurrent users, with > the potential for up to 5,000+ in the event of a news event like 9/11. > > I'm planning to use SuSE 9 with squid transparently. I think > I can handle setting up squid and the other little packages > that we intend to mix in with it (already tested on a smaller What other packages mixed in are you talking about? I am currently evaluating 2 redirectors (squirm for domain redirection and adzap just for fun) and am using 2 reporting packages (SARG & Squid Graph). > scale), but I'm not sure about sizing the partitions. Is one > file system better than another for caching? How many > partitions? How big? Should I mirror the drives? I need the > best performance with just a dash of fault tolerance. :) The > config of the box will be backed up frequently in case it > needs to be rebuilt. I'm thinking a partition scheme like this: > It sounds like your on the right track but if you want to check your numbers, re-read the user guide chapter tha addresses disk/ram/cpu. http://squid-docs.sourceforge.net/latest/html/c21.html I found this to be an excellent resource for scaling the box. > /boot 100MB reiser > / 10 GB reiser > /var/log 20 GB reiser > /var/cache 30 GB aufs (or reiser? this is the cache_dir) > > Are there any squid configuration parms that I should be > aware for a deployment of this size? Any "gotchas" to look > out for? Any on-going administrative bummers? Cool tools for > administration? I'd like to run the package that comes with > SuSE and can be updated with the provided tools, but I can > compile and install from source if necessary. Any arguments > in favor of one over the other? I was lazy and used the precompiled bins on a redhat 9 box that was already configured for development. The one feature that did not precompile was the NTLM auth features. Check the ./configure --help and really make sure you have everything you want to use. > > Any recommendations or pointers in the right direction are > appreciated. Thanks! > > Tony Gettig > Kalamazoo Public Schools >
