Date: Thu, 7 Jan 1999 17:20:24 +0000 From: Geoff Buckingham <geoffb@gti.noc.demon.net> To: "Richard Seaman, Jr." <dick@tar.com>, current@FreeBSD.ORG Subject: Re: Squid + linuxthreads in action wrt crash Message-ID: <19990107172024.A25899@gti.noc.demon.net> In-Reply-To: <19990106151732.A752@tar.com>; from Richard Seaman, Jr. on Wed, Jan 06, 1999 at 03:17:32PM -0600 References: <E0zxzBQ-0007Tt-00@fanf.noc.demon.net> <E0zxzBQ-0007Tt-00@fanf.noc.demon.net> <19990106142240.A375@tar.com> <E0zy08Z-0007Wm-00@fanf.noc.demon.net> <19990106151732.A752@tar.com>
next in thread | previous in thread | raw e-mail | index | archive | help
More on the crash we had:
We reproduces it again today and captured some core.
Now I'm way out of my depth here but...
(kgdb) bt
#0 0xf0131e9f in boot ()
#1 0xf01321c8 in at_shutdown ()
#2 0xf01ba9ad in trap_fatal ()
#3 0xf01ba443 in trap_pfault ()
#4 0xf01ba0d6 in trap ()
#5 0xf01b933f in generic_bzero ()
#6 0xf01a7344 in vm_object_allocate ()
#7 0xf01ac824 in vnode_pager_alloc ()
#8 0xf01587e7 in vfs_object_create ()
#9 0xf015d2b2 in vn_open ()
#10 0xf0159d1d in open ()
#11 0xf01bac3f in syscall ()
#12 0xf01b124c in Xint0x80_syscall ()
#13 0x804ebd6 in ?? ()
#14 0x280df990 in ?? ()
#15 0x280dff49 in ?? ()
#16 0x280dff09 in ?? ()
#17 0x280dfb47 in ?? ()
#18 0x400 in ?? ()
#19 0x83ffffd0 in ?? ()
(kgdb)
This is current plus linux threads softupdates and ccd, are any of them
particularly implicated by the above?
Also I had trouble geting the above as the kernel is a.out and has no
debugging symbols. I got to the above by recompiling the kernel that
crashed after a config -g and using that as 'symbol-file' for kgdb and
the actual kernel as 'exec-file' on a second, older machine with an
a.out gdb. How unsafe was this?
For anyone wishing to try and reproduce this or just interested in the
potential of kthreads and squid, the following is the output of systat
at the time of the crash, the system had been running under this load for
about 10 mins:
3 users Load 1.02 1.00 0.76 Thu Jan 7 12:50
Mem:KB REAL VIRTUAL VN PAGER SWAP PAGER
Tot Share Tot Share Free in out in out
Act 113856 2072 156124 4392 95996 count
All 935636 6712 1640992 10392 pages
cow Interrupts
Proc:r p d s w Csw Trp Sys Int Sof Flt zfod 1658 total
1 6 23 1330 928 5639 1659 679 9 75484 wire 100 clk0 irq0
118608 act 128 rtc0 irq8
50.1%Sys 6.8%Intr 35.7%User 0.0%Nice 7.4%Idl 739080 inact 110 pci irq10
| | | | | | | | | | cache 1285 pci irq11
=========================+++>>>>>>>>>>>>>>>>>> 95996 free fdc0 irq6
daefr 35 sio0 irq4
Namei Name-cache Dir-cache prcfr sio1 irq3
Calls hits % hits % react
1191 920 77 24 2 pdwake
pdpgs
Discs da0 da1 da2 da3 da4 da5 ccd0 intrn
KB/t 21.11 8.66 6.91 6.81 8.40 8.90 7.49 8345 buf
tps 2 16 37 35 10 12 111 72689 desiredvnodes
MB/s 0.04 0.14 0.25 0.23 0.08 0.11 0.81 69461 numvnodes
28 freevnodes
I will attempt to do this again in the next few days without ccd and
softupdates. Is debugging going to be made easier or immposible by switching
to an elf kernel?
--
GeoffB
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990107172024.A25899>