On Sun, Dec 28, 2008 at 08:28:52PM +0100, Matthias Kilian spoke thusly: > On Sun, Dec 28, 2008 at 12:39:57PM -0600, Denny White wrote: > > I know from past experience and doing a lot of online searching > > that this problem crops up from time to time. Usually someone > > just did a new install and doesn't know the locate.datebase > > isn't created by default. > > There's an empty /var/db/locate.database in etcXX.tgz. Didn't you > install etcXX.tgz? > > Ciao, > Kili
Okay, maybe I worded my question wrong. Sure, I installed all sets including etcXX.tgz. Regardless, I'm still trying to solve the problem of the update failing now. As I said before, I've tried to check every conceivable thing I could come up with that might be causing it and am still coming up emtpy. Weird. And, IIRC, I did have to get it kicked off originally manually. Just for fun I just copied the original empty file to /var/db/ -r--r--r-- 1 root wheel 0 Dec 28 15:16 locate.database and ran weekly again. Did a listing from another term to show it had started and was trying: -rw------- 1 root wheel 0 Dec 28 15:19 locate.database.Hojrmm6565 After about 10 seconds it failed again. I ran top to watch how much load it was creating, not that it ever came anywhere near what it does while kernel or userland building, but just to see at what point it failed in case it would help someone else who was trying to help with the problem. 18714 root 57 0 4324K 5912K run - 0:10 31.49% perl That probably won't help but I'm at the 'shooting in the dark' point here. For the record, here's the locate portion of weekly to show that nothing's been altered. I decreased the indentation to keep within 72 characters: echo "" if [ -f /var/db/locate.database ]; then TMP=`mktemp /var/db/locate.database.XXXXXXXXXX` if [ $? -eq 0 ]; then trap 'rm -f $TMP; exit 1' 0 1 15 echo "Rebuilding locate database:" UPDATEDB="/usr/libexec/locate.updatedb" echo "${UPDATEDB} --fcodes=- --tmpdir=${TMPDIR:-/var/tmp}" | \ nice -5 su -m nobody 2>/dev/null 1>$TMP if [ -s "$TMP" ]; then chmod 444 $TMP chown root:wheel $TMP mv -f $TMP /var/db/locate.database else echo "Not installing locate database; zero size" fi else echo "Not rebuilding locate database; can't create temp file" fi else echo "Not rebuilding locate database; no /var/db/locate.database" fi I'm just wondering, since perl moved to a new number if something's changed there that's causing the problem. At that point I wouldn't know what to do. I did check ulimit to see what the settings were. As a user I get time(cpu-seconds) unlimited file(blocks) unlimited coredump(blocks) unlimited data(kbytes) 524288 stack(kbytes) 4096 lockedmem(kbytes) 166098 memory(kbytes) 497108 nofiles(descriptors) 128 processes 64 and as root time(cpu-seconds) unlimited file(blocks) unlimited coredump(blocks) unlimited data(kbytes) 1048576 stack(kbytes) 8192 lockedmem(kbytes) 166098 memory(kbytes) 497108 nofiles(descriptors) 128 processes 1310 Once again, shooting in the dark. Denny White -- /"\ ASCII Ribbon Campaign \ / Respect for low technology. X Keep e-mail messages readable by any computer system. / \ Keep it ASCII. =============================================================== GnuPG key : 0x1644E79A | http://wwwkeys.nl.pgp.net Fingerprint: D0A9 AD44 1F10 E09E 0E67 EC25 CB44 F2E5 1644 E79A ===============================================================