On Wed, Jan 28, 2004 at 11:34:19AM -0500, Kent Borg wrote: > On Wed, Jan 28, 2004 at 03:44:14PM +0000, Alan Dunkley wrote: > > On Wednesday 28 Jan 2004 3:01 pm, Ron Herardian wrote: > > > When installing everything and allowing for future updates and packages I > > > am using the following disk layout: > > > > > > Mount Point Size > > > /boot 100MB > > > / 500MB > > > /usr 4GB > > > /var 2GB > > > swap 2x physical RAM, e.g., 1GB > > > /home TBD, e.g., 1GB per user > > > /opt TBD/catchall [3rd party servers will be installed here] > > A couple suggestions: > > Swap: Just use 1 GB. It is big enough to be significant from a RAM > perspective, but it is still small and cheap in terms of disk sizes. This is a good place to start unless you have experience otherwise. > Put a lot of physical RAM in your machine, certainly. Yes, Yes, Yes... DRAM.... Remember that Linux supports swap to files. If you need more swap space for a project you can add swap files at a later time. You do want enough DRAM + swap to permit your largest processes to do common things like "fork(); exec()" and not have the kernel return errors for want of virtual resources. Today DRAM is inexpensive and "disk-speed to DRAM-speed" ratios so large that you do not want much if any swap IO. With modern VM systems and concepts like shared pages, copy on write etc... the key value for a 'big' swap is kernel book keeping. The kernel must believe that it has has enough backing store in case you "touch" the pages the process requested. I happen to like testing various bounds and limits so my swap space is big. Big enough for two run away big processes (~6GB). In practice this is way too large and when swapping the box is worthless for interactive work. The classic 2xDRAM, 4xDRAM, 8xRAM swap rules had a lot to do with the balance of DRAM cost (expensive) to disk cost to system usefulness. Today time slices are very short. When time slices are compared to disk access time extensive swap IO does not make sense. When the effective runtime for a set of pages is larger than the disk IO time and overhead it can make sense. See stuff like job control, Ctl-Z stop, checkpoint and restart. On an interactive workstation swap IO does not make much sense. -- T o m M i t c h e l l mitch48-at-sbcglobal-dot-net
