Crash in HEAD - cache digests

[GoogleCodeExporter]

What steps will reproduce the problem?
1. tproxy/ICP/WCCP (not sure if it's caused by these) and new COSS rebuild
scripts

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb7cf06c0 (LWP 31191)]
0x080b72f8 in storeLog (tag=3, e=0x9df9e38) at store_log.c:68
68              logfilePrintf(storelog, "%9ld.%03d %-7s %02d %08X %s %4d
%9ld %9ld %9ld %.*s %" PRINTF_OFF_T "/%" PRINTF_OFF_T " %s %s\n",
(gdb) bt
#0  0x080b72f8 in storeLog (tag=3, e=0x9df9e38) at store_log.c:68
#1  0x080afac8 in storeRelease (e=0x9df9e38) at store.c:1628
#2  0x080affa3 in storeUnlockObjectDebug (e=0x9df9e38, file=0x8109045
"peer_digest.c", line=702) at store.c:334
#3  0x0809b835 in peerDigestReqFinish (fetch=0xf8efc50, fcb_valid=1,
pdcb_valid=1, pcb_valid=1, reason=0x811cc49 "Gateway Time-out", err=1) at
peer_digest.c:702
#4  0x0809c01c in peerDigestFetchAbort (fetch=0xf8efc50, reason=0x811cc49
"Gateway Time-out") at peer_digest.c:612
#5  0x0809c347 in peerDigestFetchReply (data=0xf8efc50, nr={node =
0x3c3ccba8, offset = 0}, size=1314) at peer_digest.c:398
#6  0x080b30cc in storeClientCallback (sc=0x9ec23c0, sz=-1) at
store_client.c:267
#7  0x080b34c4 in storeClientCopy2 (e=0x1bfe36e8, sc=0x9ec23c0) at
store_client.c:520
#8  0x080b37b2 in InvokeHandlers (e=0x1bfe36e8) at store_client.c:870
#9  0x080ae1a7 in storeBufferFlush (e=0x1bfe36e8) at store.c:2049
#10 0x08076970 in errorAppendEntry (entry=0x1bfe36e8, err=0x37406010) at
errorpage.c:331
#11 0x080b08d2 in storeRequestFailed (e=0x1bfe36e8, err=0x37406010) at
store.c:1469
#12 0x0807a3b7 in fwdStateFree (fwdState=0x9ec2480) at forward.c:104
#13 0x080e35ce in commCallCloseHandlers (fd=18) at comm.c:637
#14 0x080e4243 in comm_close (fd=18) at comm.c:814
#15 0x080e8196 in comm_select (msec=10) at comm_generic.c:344
#16 0x080944a4 in main (argc=2, argv=0xbff08764) at main.c:933
(gdb) 

Original issue reported on code.google.com by Fraw...@gmail.com on 27 Jul 2009 at 11:21

[GoogleCodeExporter]

Another bt

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb7ec36c0 (LWP 6770)]
0x080b72f8 in storeLog (tag=3, e=0x8c39578) at store_log.c:68
68              logfilePrintf(storelog, "%9ld.%03d %-7s %02d %08X %s %4d %9ld 
%9ld
%9ld %.*s %" PRINTF_OFF_T "/%" PRINTF_OFF_T " %s %s\n",
(gdb) bt
#0  0x080b72f8 in storeLog (tag=3, e=0x8c39578) at store_log.c:68
#1  0x080afac8 in storeRelease (e=0x8c39578) at store.c:1628
#2  0x080affa3 in storeUnlockObjectDebug (e=0x8c39578, file=0x8109045
"peer_digest.c", line=702) at store.c:334
#3  0x0809b835 in peerDigestReqFinish (fetch=0x8b72bd8, fcb_valid=1, 
pdcb_valid=1,
pcb_valid=1, reason=0x81060fa "Not Found", err=1) at peer_digest.c:702
#4  0x0809c01c in peerDigestFetchAbort (fetch=0x8b72bd8, reason=0x81060fa "Not
Found") at peer_digest.c:612
#5  0x0809c347 in peerDigestFetchReply (data=0x8b72bd8, nr={node = 0xb03e008, 
offset
= 0}, size=341) at peer_digest.c:398
#6  0x080b30cc in storeClientCallback (sc=0xafca2f0, sz=-1) at 
store_client.c:267
#7  0x080b34c4 in storeClientCopy2 (e=0x8b72de8, sc=0xafca2f0) at 
store_client.c:520
#8  0x080b37b2 in InvokeHandlers (e=0x8b72de8) at store_client.c:870
#9  0x0807bfd0 in fwdComplete (fwdState=0xafca350) at forward.c:1148
#10 0x08088efd in httpAppendBody (httpState=0xafca3a0, buf=0xb036155 "", len=0,
buffer_filled=0) at http.c:995
#11 0x0808b746 in httpReadReply (fd=19, data=0xafca3a0) at http.c:1254
#12 0x080e8432 in comm_select (msec=2) at comm_generic.c:264
#13 0x080944a4 in main (argc=2, argv=0xbfe68674) at main.c:933
(gdb) 

Original comment by Fraw...@gmail.com on 27 Jul 2009 at 11:21

[GoogleCodeExporter]

Disable using peer digests for now. This particular bug seems very much related 
to
cache peer digests.

Original comment by adrian.c...@gmail.com on 28 Jul 2009 at 1:51

[GoogleCodeExporter]

Original comment by adrian.c...@gmail.com on 28 Jul 2009 at 1:52

• Changed title: Crash in HEAD - cache digests
• Changed state: Accepted
• Added labels: Priority-High, Version-Head
• Removed labels: Priority-Medium, Version-1.0

[GoogleCodeExporter]

I've started fleshing this out for some further testing. Stay tuned.

Original comment by adrian.c...@gmail.com on 27 Oct 2009 at 6:49

[GoogleCodeExporter]

Would you mind upgrading to the latest LUSCA_HEAD and trying it out? I've 
committed something which should 
work around the bug. I'm trying to figure out why the specific conditions exist 
that cause the bug but I won't be 
able to without further debugging.

Original comment by adrian.c...@gmail.com on 27 Oct 2009 at 7:40