I'm getting a kernel panic every time I start getting traffic (~ 20Mb/s
total) to two squid processes. This was not happening with -RELEASE. I
upgraded on the advice that STABLE would decrease the %cpu spent in
system, which was much higher than that spent in user. The machine's
only fucntion is as a reverse proxy (2 separate squid processes to take
advante of both cpus).
Let me know if a dump would be helpful, or if there's anything I can do
to help.
Dual AMD Opteron system..
Jeff
ps. I don't believe the problem is with a zebra/ospfd kernel
interaction. the open port RST responses are expected; i can explain
why if it would provide insight into the problem...
www2.cdn.sjc#uname -a
FreeBSD www 5.3-STABLE FreeBSD 5.3-STABLE #2: Sun Dec 5 21:06:14 PST
2004 root@www2.cdn.sjc:/usr/obj/usr/src/sys/SMP amd64
FreeBSD/amd64 (www2.cdn.sjc) (ttyd0)
login: Dec 6 13:03:37 www2 ospfd[18772]: old umask 23 127
Limiting open port RST response from 518 to 200 packets/sec
Limiting open port RST response from 318 to 200 packets/sec
Limiting open port RST response from 264 to 200 packets/sec
Limiting open port RST response from 206 to 200 packets/sec
Fatal trap 12: page fault while in kernel mode
cpuid = 1; apic id = 01
fault virtual address = 0x18
fault code = supervisor read, page not present
instruction pointer = 0x8:0xffffffff803a14b3
stack pointer = 0x10:0xffffffffb1bd9800
frame pointer = 0x10:0x0
code segment = base 0x0, limit 0xfffff, type 0x1b
= DPL 0, pres 1, long 1, def32 0, gran 1
processor eflags = interrupt enabled, resume, IOPL = 0
current process = 44 (swi1: net)
trap number = 12
panic: page fault
cpuid = 1
boot() called on cpu#0
Uptime: 23m44s
Cannot dump. No dump device defined.
Automatic reboot in 15 seconds - press a key on the console to abort
Rebooting...
cpu_reset called on cpu#0
cpu_reset: Stopping other CPUs
On Mon, Dec 06, 2004 at 01:19:54PM -0800, Jeff Behl wrote:> I'm getting a kernel panic every time I start getting traffic (~ 20Mb/s > total) to two squid processes. This was not happening with -RELEASE. I > upgraded on the advice that STABLE would decrease the %cpu spent in > system, which was much higher than that spent in user. The machine's > only fucntion is as a reverse proxy (2 separate squid processes to take > advante of both cpus). > > Let me know if a dump would be helpful, or if there's anything I can do > to help.You really need to get a debugging traceback - see the chapter on kernel debugging in the developers' handbook. Kris -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 187 bytes Desc: not available Url : http://lists.freebsd.org/pipermail/freebsd-stable/attachments/20041206/b7785fbd/attachment.bin
On Mon, 06 Dec 2004 13:19:54 -0800, Jeff Behl <jbehl@fastclick.com> wrote:> I'm getting a kernel panic every time I start getting traffic (~ 20Mb/s > total) to two squid processes. This was not happening with -RELEASE. I > upgraded on the advice that STABLE would decrease the %cpu spent in > system, which was much higher than that spent in user. The machine's > only fucntion is as a reverse proxy (2 separate squid processes to take > advante of both cpus). > > Let me know if a dump would be helpful, or if there's anything I can do > to help. > > Dual AMD Opteron system.. > > Jeff > ps. I don't believe the problem is with a zebra/ospfd kernel > interaction. the open port RST responses are expected; i can explain > why if it would provide insight into the problem... >I had a panic that looked a lot like this (panic in swi, came up when I had some network traffic), and it's also been reported by other people than me. Turning of SACK seems to work for most people (at least it worked for me). Put net.inet.tcp.sack.enable=0 in /etc/sysctl.conf. Arjan