On 17/05/2006 7:38 a.m., Henrik Nordstrom wrote:
tis 2006-05-16 klockan 12:46 -0400 skrev Jeremy Hall:
I noticed that those zeroized lists would sometimes have a router id of
0.0.0.0 and then either wccp would try to send to 0.0.0.0:2048, it would
abort wccp2 processing when it encountered it, or it would throw an
error. Did you try it after zeroizing the list?
The 0.0.0.0 entry was always there as the first entry in the list.. now
it's not there anymore.
I don't have any WCCPv2 equipment to run a full test with, but the code
is quite simple..
Regards
Henrik
Speaking of WCCPv2, I'm seeing this logged when I do a squid -k reconfigure:
2006/05/17 16:48:00| Beginning Validation Procedure
2006/05/17 16:48:00| Completed Validation Procedure
2006/05/17 16:48:00| Validated 35418 Entries
2006/05/17 16:48:00| store_swap_size = 1555560k
2006/05/17 16:48:01| storeLateRelease: released 0 objects
2006/05/17 16:55:30| Reconfiguring Squid Cache (version 2.6.DEVEL-CVS)...
2006/05/17 16:55:30| FD 23 Closing HTTP connection
2006/05/17 16:55:30| FD 24 Closing WCCP socket
2006/05/17 16:55:30| FD 24 Closing WCCP socket
2006/05/17 16:55:30| assertion failed: comm.c:647: "F->flags.open"
2006/05/17 16:57:34| Starting Squid Cache version 2.6.DEVEL-CVS for x86_64-unkno
wn-linux-gnu...
2006/05/17 16:57:34| Process ID 23052
2006/05/17 16:57:34| With 1024 file descriptors available
2006/05/17 16:57:34| Performing DNS Tests...
2006/05/17 16:57:34| Successful DNS name lookup tests...
2006/05/17 16:57:34| DNS Socket created at 0.0.0.0, port 32983, FD 8
2006/05/17 16:57:34| Adding nameserver 192.168.0.5 from /etc/resolv.conf
Under gdb I'm seeing this:
Program received signal SIGABRT, Aborted.
[Switching to Thread 47026893756432 (LWP 22594)]
0x00002ac54cbd75c5 in raise () from /lib64/libc.so.6
(gdb) bt full
#0 0x00002ac54cbd75c5 in raise () from /lib64/libc.so.6
No symbol table info available.
#1 0x00002ac54cbd8f10 in abort () from /lib64/libc.so.6
No symbol table info available.
#2 0x000055555557f4fb in xassert (msg=0x5842 <Address 0x5842 out of bounds>,
file=0x5842 <Address 0x5842 out of bounds>, line=6) at debug.c:415
No locals.
#3 0x000055555557ca32 in comm_close (fd=24) at comm.c:647
F = (fde *) 0x2ac54d61abd0
#4 0x00005555555c1baf in wccp2ConnectionShutdown () at wccp2.c:725
No locals.
#5 0x00005555555c1c05 in wccp2ConnectionClose () at wccp2.c:744
service_list_ptr = Variable "service_list_ptr" is not available.
(gdb) frame 1
#1 0x00002ac54cbd8f10 in abort () from /lib64/libc.so.6
(gdb) frame 2
#2 0x000055555557f4fb in xassert (msg=0x5842 <Address 0x5842 out of bounds>,
file=0x5842 <Address 0x5842 out of bounds>, line=6) at debug.c:415
415 abort();
(gdb) frame 3
#3 0x000055555557ca32 in comm_close (fd=24) at comm.c:647
647 assert(F->flags.open);
(gdb) frame 4
#4 0x00005555555c1baf in wccp2ConnectionShutdown () at wccp2.c:725
725 comm_close(theInWccp2Connection);
(gdb) frame 5
#5 0x00005555555c1c05 in wccp2ConnectionClose () at wccp2.c:744
744 wccp2ConnectionShutdown();
(gdb) info locals
service_list_ptr = Variable "service_list_ptr" is not available.
(gdb)
My WCCPv2 doesn't seem to work at all actually, config looks like this:
wccp2_router 192.168.0.1
wccp2_version 4
wccp2_forwarding_method 1
wccp2_return_method 1
wccp2_service standard 0
wccp2_outgoing_address 192.168.0.5
Seems that the server doesn't even attempt to talk to the router at all for some
reason :(
reuben
