>>> On Mon, Mar 17, 2008 at 5:06 PM, in message <[EMAIL PROTECTED]>, "Collinson.Shannon" <[EMAIL PROTECTED]> wrote: -snip- > So either it's such a no-brainer that it doesn't bear mentioning for the > most part, or it's a bad idea (explained in some doc I haven't been able > to google). Which is it? Thanks!
The first hit on my Google search explained quite a bit how tmpfs works, and what the mount command looks like: http://www.funtoo.org/en/articles/linux/ffg/3/ I don't have a good feel for the performance implications. According to the article above, tmpfs uses both virtual memory and swap space to satisfy the space needed. And, you can put a limit on how big it can grow. I wouldn't necessarily think that tmpfs in and of itself would increase your working set size. If a page isn't referenced for quite a while, z/VM should page it out, regardless of what it's being used for by Linux. As others have talked about, real storage in a shared environment is something to be used wisely. If you don't have a need for a blazingly fast /tmp file system, you might be better off with either real DASD, or VDISK space (assuming you have z/VM). If you like the idea of /tmp being completely cleaned out with every system boot, there's nothing stopping you from doing that in a local startup script. The Filesystem Hierarchy Standard explicitly states that the contents of /tmp are not guaranteed to be preserved across a reboot. Mark Post ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
