Re: [squid-users] Squid 3.x and wccp2_router crash

On Dec 18, 2013, at 9:18 PM, Don Brearley <donbrearley_at_hibbing.edu> wrote:
> [. . .]
> 2013/12/18 21:14:59| storeLateRelease: released 0 objects
> 2013/12/18 21:14:59| Ip::Address::GetSockAddr : Cannot convert non-IPv4 to IPv4. from [::]
> Assertion failed: (false), function GetSockAddr, file Address.cc, line 966.
>
>
> Program received signal SIGABRT, Aborted.
> [Switching to Thread 803407400 (LWP 100109/squid)]
> 0x0000000802cb923c in thr_kill () from /lib/libc.so.7
>
>
> (gdb) backtrace
> #0 0x0000000802cb923c in thr_kill () from /lib/libc.so.7
> #1 0x0000000802d5d63b in abort () from /lib/libc.so.7
> #2 0x0000000802d40d35 in __assert () from /lib/libc.so.7
> #3 0x0000000000679bc7 in Ip::Address::GetSockAddr (this=0x7ffffffe7e80,
> buf=@0x7ffffffe7eb0) at Address.cc:966
> #4 0x00000000005ff0c7 in wccp2HandleUdp (sock=13,
> not_used=) at wccp2.cc:1191
> #5 0x00000000006d93c8 in Comm::DoSelect (msec=1000) at ModPoll.cc:504
> #6 0x0000000000673c7d in CommSelectEngine::checkEvents (
> this=, timeout=1000) at comm.cc:2068
> #7 0x000000000051f50c in EventLoop::checkEngine (this=0x7fffffffda70,
> engine=0x7fffffffdaf0, primary=) at EventLoop.cc:51
> #8 0x000000000051f725 in EventLoop::runOnce (this=0x7fffffffda70)
> at EventLoop.cc:125
> #9 0x000000000051f8e8 in EventLoop::run (this=0x7fffffffda70)
> at EventLoop.cc:95
> #10 0x0000000000594273 in SquidMain (argc=2, argv=)
> at main.cc:1501
> #11 0x00000000005945d8 in SquidMainSafe (argc=,
> argv=) at main.cc:1216
> #12 0x00000000004b4601 in _start ()
> #13 0x00000008009ca000 in ?? ()
> #14 0x0000000000000000 in ?? ()
> (gdb)

I can’t find the squid core from the 3.3.9 crashes I saw, but frames #4 and #3 do look familiar.

FWIW, I run FreeBSD kernels without IPV6 support, which may be a factor in why wccp2_address 0.0.0.0 works around the crash.

Guy