Date: Mon, 15 Apr 1996 17:39:14 -0400 (EDT) From: Charles Owens <owensc@enc.edu> To: questions list FreeBSD <freebsd-questions@freebsd.org> Subject: HELP!: page fault (revisited) Message-ID: <Pine.FBS.3.92.960415170608.17160A-100000@dingo.enc.edu>
next in thread | raw e-mail | index | archive | help
Folks:
I'm still having trouble with page faults, and it seems to occur when I
create users.
Some more background:
I set up a msqld database to contain user info. User's fill in a Web form
account application that dumps into the database. An operator can then
pull up the application data (using a different Web interface) and process
it if desired. The whole thing is built using Apache, mSQL and perl
scripts. (The mSQL database engine itself is running on a separate
FreeBSD box which doesn't ever crash.) The actual unix account creation
is handled by a hacked version of the stock adduser perl script.
The modified adduser (I call it ypadduser) works with NIS and the quota
system. This is what it does:
Add new user entry to /var/yp/master.passwd
from /var/yp - make -DNOPUSH
create new user directory, etc.
quotaoff -a
edquota -Pprotouser -u newuser
quoton -a
After last user is created:
from /var/yp - make (pushing to NIS sub servers)
quotaoff -a
quotacheck -a
quoton -a
That's basicly it. It seems that the system is rock solid most of the
time, but if I use the above scripts to create users I seem to always have
a crash.
Here's the lastest syslog output from a crash:
Apr 15 16:19:27 itsdsv1 /kernel: Fatal trap 12: page fault while in kernel mode
Apr 15 16:19:27 itsdsv1 /kernel: fault virtual address = 0x0
Apr 15 16:19:27 itsdsv1 /kernel: fault code = supervisor write, page not present
Apr 15 16:19:27 itsdsv1 /kernel: instruction pointer = 0x8:0xf0180745
Apr 15 16:19:27 itsdsv1 /kernel: code segment = base 0x0, limit 0xfffff, type 0x1b
Apr 15 16:19:27 itsdsv1 /kernel: = DPL 0, pres 1, def32 1, gran 1
Apr 15 16:19:27 itsdsv1 /kernel: processor eflags = interrupt enabled, resume, IOPL = 0
Apr 15 16:19:27 itsdsv1 /kernel: current process = 12698 (mail.local)
Apr 15 16:19:27 itsdsv1 /kernel: interrupt mask =
Apr 15 16:19:27 itsdsv1 /kernel: panic: page fault
Apr 15 16:19:28 itsdsv1 /kernel:
Apr 15 16:19:28 itsdsv1 /kernel: syncing disks... 22 22 20 14 7
My system (ASUS P100) has 64 megs RAM and ~160 megs swap space spread
accross two drives. Here's swapinfo output:
Device 1K-blocks Used Avail Capacity Type
/dev/sd0s1b 32768 0 32704 0% Interleaved
/dev/sd1b 131072 0 131008 0% Interleaved
Total 163712 0 163712 0%
The smaller chunk is on a 1 gig SCSI-2 drive ("DEC DSP3105S 392A") while
the balance is on a 4 gig SCSI-2 WIDE drive ("SEAGATE ST15230W 0298").
I'm currently running FreeBSD-2.1-stable (a la 3/16/96) but I had the same
crashing (and other problems) when I was running 2.0.5-RELEASE.
I could really use some help figuring this out! If anyone replied to what
I posted about this on Friday, entitled "unravelling page fault
occurances", thanks, but please repost. I let my inbox disk quota fill up
so I could have missed it. (oops :)
I'm trying to position FreeBSD as _the_ server platform for our entire
instituion, and I really need to get this problem resolved if I'm to
succeed.
Thanks very much...
---
-------------------------------------------------------------------------
Charles Owens Email: owensc@enc.edu
"I read somewhere to learn is to
Information Technology Services remember... and I've learned that
Eastern Nazarene College we've all forgot..." - King's X
-------------------------------------------------------------------------
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.FBS.3.92.960415170608.17160A-100000>