So I heard Zibby yammering on about 'tmpfs' for the /tmp partition, on this list, and said 'hey, that sounds like a really cool idea'. So I decided to try it. problem is, there's *NO* doco about it yet. so here's what I gleaned from sorting through kernel-traffic archives, assembled into a micro-HOWTO. what is it? ----------- tmpfs is a RAM-based filesystem, that can be swapped out. it also has size-limitation options (which keeps it from filling up all your memory and bricking up the system). what do you need? ----------------- Linux 2.4.x (I think >2.4.2, but I could be wrong) mount v2.10h (?) (I have 2.11b. someone reported bugs in 2.11e, regarding tmpfs). procedure: ---------- - make sure you have a recent version of mount. - get a recent kernel - build the kernel with '#define CONFIG_TMPFS 1' in the .config file. if you do a 'make menuconfig', it's under Filesystems as 'Virtual memory file system support (former shm fs)' - try it out. here's my sample run: root at steel:/mnt# mkdir /mnt/temp root at steel:/mnt# dpkg -l mount Desired=Unknown/Install/Remove/Purge/Hold | Status=Not/Installed/Config-files/Unpacked/Failed-config/Half-installed |/ Err?=(none)/Hold/Reinst-required/X=both-problems (Status,Err: uppercase=bad) ||/ Name Version Description +++-==============-==============-============================================ ii mount 2.11b-4 Tools for mounting and manipulating filesyst root at steel:/mnt# mount -t tmpfs tmpfs /mnt/temp/ root at steel:/mnt# mount /dev/hda1 on / type ext2 (rw,errors=remount-ro,errors=remount-ro) proc on /proc type proc (rw) devpts on /dev/pts type devpts (rw,gid=5,mode=620) /dev/hda4 on /home type ext2 (rw) /dev/hda5 on /usr type ext2 (rw) /dev/hda6 on /var type ext2 (rw) tmpfs on /mnt/temp type tmpfs (rw) root at steel:/mnt# df -h Filesystem Size Used Avail Use% Mounted on /dev/hda1 471M 34M 412M 8% / /dev/hda4 25G 1.9G 22G 8% /home /dev/hda5 4.6G 546M 3.8G 13% /usr /dev/hda6 1.8G 375M 1.3G 22% /var tmpfs 953M 0 953M 0% /mnt/temp root at steel:/mnt# umount /mnt/temp/ root at steel:/mnt# mount -t tmpfs -o size=128M tmpfs /mnt/temp/ root at steel:/mnt# mount /dev/hda1 on / type ext2 (rw,errors=remount-ro,errors=remount-ro) proc on /proc type proc (rw) devpts on /dev/pts type devpts (rw,gid=5,mode=620) /dev/hda4 on /home type ext2 (rw) /dev/hda5 on /usr type ext2 (rw) /dev/hda6 on /var type ext2 (rw) tmpfs on /mnt/temp type tmpfs (rw,size=128M) root at steel:/mnt# df -h Filesystem Size Used Avail Use% Mounted on /dev/hda1 471M 34M 412M 8% / /dev/hda4 25G 1.9G 22G 8% /home /dev/hda5 4.6G 546M 3.8G 13% /usr /dev/hda6 1.8G 375M 1.3G 22% /var tmpfs 128M 0 128M 0% /mnt/temp as we can see from this: - yes, I have nearly 1GB of swap. doesn't hurt, and I've got to use this huge disk *somehow*. :) - tmpfs, if not given a size= option, will assume it can use *all* the space. this is a *BAD THING* because runaway processes filling up your filesystem, will fill up all of your memory, giving an Out Of Memory condition that's awfully hard to recover from. and here's the entry I put in my /etc/fstab: tmpfs /tmp tmpfs defaults,size=500M 0 0 be warned, I haven't yet actually tried rebooting with this setup; but I don't see why it wouldn't work. Carl Soderstrom -- Network Engineer Real-Time Enterprises (952) 943-8700