by Joe DeBis » Thu, 12 Sep 1996 04:00:00
HELP!!!!!
I have a client who is running out of inodes. The root file system of
their system only have the OS on it. It is v3.2.4.2. Using df -t shows
about 300 inodes and falling! I don;t know where all the inodes have
gone!!!!! The max is 65000, so 64700 are in use!! How can I track this
down?????
Please help as soon as possible!
SCO support was useless on this problem!!!
Joe DeBiso
Plantrol Systems, Ltd.
by Jean-Pierre Radl » Thu, 12 Sep 1996 04:00:00
I assume that /etc/celantmp is running once a week, so that /usr/tmp and /tmpQuote:Joe DeBiso writes:
> HELP!!!!!
> I have a client who is running out of inodes. The root file system of
> their system only have the OS on it. It is v3.2.4.2. Using df -t shows
> about 300 inodes and falling! I don;t know where all the inodes have
> gone!!!!! The max is 65000, so 64700 are in use!! How can I track this
> down?????
> Please help as soon as possible!
> SCO support was useless on this problem!!!
If you're making LoneTar or BackupEdge backups, then look at the last master
listing.
If not, you might have to run ncheck and just look for odd files.
--
by James R. Sulliva » Fri, 13 Sep 1996 04:00:00
Places I'd look are:
/tmp
/usr/tmp
/usr/mmdf/...
/usr/spool
/usr/adm/sa
I once had a system that was slower than molasses in January. The problem turned
out to be a previous consultant, who in an effort to reduce disk space, took an
axe to the /usr/mmdf structure. The result was hourly e-mail from cron detailing
the problems with /usr/mmdf, but the problems prevented the e-mail from being
delivered, so it was sitting in the queue directories. The directory was huge,
consuming disk space, inodes, and massive amounts of CPU every hour as it tried
to read a 10,000+ directory to find a spare slot to add the next message.
cleanque -w cleared the problem (although it took a long time :-)
1. Big Time Server for a Big Time NG
I'm working at putting together my RH5.2 system, and have been pretty
successful getting everything installed to this point, but now I'm
kind of lost as to the direction I need to go in. I'm currently doing
the dial-up ISP thing, but will move to cable as soon as I get things
set up properly and am sure about security. Right now, I'm having a
few problems in getting started.
1) I can't seem to telnet into my linux box unless I start up a
server manually. I usually do this on port 5001 if that matters.
That works fine. So why won't it work for good ol' port 23? I
connect, but I never get a login prompt... just a lost connection.
2) Samba doesn't want to accept my login. I can see the machine in
my Net 'Hood, and it will ask for a password for resource
"\\LINUX\IPC$". I type in my PW and it always says that it's
incorrect. I can't type it incorrectly THAT many times. :-)
3) This might be a Win98 thing or might not, but I configured TCP/IP
to look at my linux box for DNS first. Whenever I try to name my
linux box, the dial-up crud comes up. Does this mean that DNS isn't
working on my linux machine?
Thanks for all of your help,
Will
3. I have a BIG, BIG,BIG problem with DOSEMU 0.98.5.
4. Simple C++ wrapper for Xlib?
5. big disk problems - or: big-disk problems?
7. Lots of inodes on a big disk on SVR4
8. procmail
9. Display Problem big time...Help needed...!!!!!!!!
10. On-disk inode layout on big-endian boxen (A/UX, ISC, m68k Xenix)
12. Display Problem big time...Help needed...!!!!!!!!
13. I'm Screwed Big-Time(mouse problem)