[squid-users] SIGABRT (coredump) in Ip::Address::getAddrInfo(addrinfo*&, int)
Alex Rousskov
rousskov at measurement-factory.com
Wed Sep 27 21:49:09 UTC 2023
On 2023-09-27 15:43, Dmitry Katsubo wrote:
> On 2023-09-27 19:15, Alex Rousskov wrote:
>> On 2023-09-27 11:08, Dmitry Katsubo wrote:
>>
>>> After upgrading Squid from v4.13-10+deb11u2 (bullseye) to v5.7-2
>>> (bookworm) I started to get about 5 core dumps per day like below, provided .
>>> How can I find out the root of the problem and eliminate it?
>>
>>
>> Your Squid is most likely suffering (among other v5 bugs) from Squid Bug
>> 5154: https://bugs.squid-cache.org/show_bug.cgi?id=5154
>>
>> There is currently no fix for that bug (that I can recommend without serious caveats). There has been significant progress towards a fix, but it may have stalled at
>> https://github.com/squid-cache/squid/pull/1421
>>
>> The best known way to prevent bug 5154 is to enable IPv6 support, but that is not feasible in many environments.
>>
>> Another Squid admin has reported success with an unofficial (and probably incomplete) workaround at https://lists.squid-cache.org/pipermail/squid-users/2023-September/026103.html
> Debain sid has v6.1-2, which I believe is also affected, hence no reason to try it 🤔
Your call. I believe Squid v6 has fixed several bugs that are still
present in v5.
> After reading the bug#5154 description I see that the issue is because of disabled IPv6, in my case, via sysctl:
>
> net.ipv6.conf.default.disable_ipv6=1
> net.ipv6.conf.all.disable_ipv6=1
>
> I believe re-enabling IPV6 should help the issue?
If Squid successfully detects IPv6 support, then bug 5154 crashes should
go away. If Squid does not, it should log a message to cache.log that
starts with "WARNING: BCP 177 violation"
> I have checked why I have disabled it ... and it turned out that was an advise to fix the following error message in /var/log/squid/cache.log:
>
> SendEcho ERROR: sending to ICMPv6 packet to [2606:4700:e0::ac40:6225]: (101) Network is unreachable
>
> 😱 Boomerang has returned.
I have not studied that ICMP problem, but it might have a solution that
does not involve disabling IPv6 support. The latter, as it turns out,
opens a rather large Pandora box!
HTH,
Alex.
>>> # coredumpctl gdb -1 --no-pager
>>> PID: 357187 (squid)
>>> UID: 13 (proxy)
>>> GID: 13 (proxy)
>>> Signal: 6 (ABRT)
>>> Timestamp: Wed 2023-09-27 11:27:08 CEST (5h 16min ago)
>>> Command Line: $'(squid-1)' --kid squid-1 --foreground -sYC
>>> Executable: /usr/sbin/squid
>>> Control Group: /system.slice/squid.service
>>> Unit: squid.service
>>> Slice: system.slice
>>> Storage: /var/lib/systemd/coredump/core.squid.13.f7e894167d564f788825cae8827bd2df.357187.1695806828000000.zst (present)
>>> Size on Disk: 2.5M
>>> Message: Process 357187 (squid) of user 13 dumped core.
>>>
>>> Module libsystemd.so.0 from deb systemd-252.12-1~deb12u1.amd64
>>> Stack trace of thread 357187:
>>> #0 0x00007f752b2a9d3c n/a (libc.so.6 + 0x8ad3c)
>>> #1 0x00007f752b25af32 raise (libc.so.6 + 0x3bf32)
>>> #2 0x00007f752b245472 abort (libc.so.6 + 0x26472)
>>> #3 0x00007f752b245395 n/a (libc.so.6 + 0x26395)
>>> #4 0x00007f752b253e32 __assert_fail (libc.so.6 + 0x34e32)
>>> #5 0x00005576cf7c0ee4 _ZNK2Ip7Address11getAddrInfoERP8addrinfoi (squid + 0x466ee4)
>>> #6 0x00005576cf7bc3c5 _Z11comm_openexiiRN2Ip7AddressEiPKc (squid + 0x4623c5)
>>> #7 0x00005576cf80fdfe _ZN4Comm10ConnOpener8createFdEv (squid + 0x4b5dfe)
>>> #8 0x00005576cf8109c1 _ZN4Comm10ConnOpener5startEv (squid + 0x4b69c1)
>>> #9 0x00005576cf7ab09c _ZN9JobDialerI8AsyncJobE4dialER9AsyncCall (squid + 0x45109c)
>>> #10 0x00005576cf7a779f _ZN14AsyncCallQueue8fireNextEv (squid + 0x44d79f)
>>> #11 0x00005576cf7a7b52 _ZN14AsyncCallQueue4fireEv (squid + 0x44db52)
>>> #12 0x00005576cf55f689 _ZN9EventLoop7runOnceEv (squid + 0x205689)
>>> #13 0x00005576cf55f778 _ZN9EventLoop3runEv (squid + 0x205778)
>>> #14 0x00005576cf665a50 _Z9SquidMainiPPc (squid + 0x30ba50)
>>> #15 0x00005576cf4fdae1 main (squid + 0x1a3ae1)
>>> #16 0x00007f752b2461ca n/a (libc.so.6 + 0x271ca)
>>> #17 0x00007f752b246285 __libc_start_main (libc.so.6 + 0x27285)
>>> #18 0x00005576cf505cc1 _start (squid + 0x1abcc1)
>>> ELF object binary architecture: AMD x86-64
>>>
>>> GNU gdb (Debian 13.1-3) 13.1
>>> Reading symbols from /usr/sbin/squid...
>>> (No debugging symbols found in /usr/sbin/squid)
>>> [New LWP 357187]
>>> [Thread debugging using libthread_db enabled]
>>> Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
>>> Core was generated by `(squid-1) --kid squid-1 --foreground -sYC'.
>>> Program terminated with signal SIGABRT, Aborted.
>>> (gdb) bt
>>> #0 0x00007f752b2a9d3c in ?? () from /lib/x86_64-linux-gnu/libc.so.6
>>> #1 0x00007f752b25af32 in raise () from /lib/x86_64-linux-gnu/libc.so.6
>>> #2 0x00007f752b245472 in abort () from /lib/x86_64-linux-gnu/libc.so.6
>>> #3 0x00007f752b245395 in ?? () from /lib/x86_64-linux-gnu/libc.so.6
>>> #4 0x00007f752b253e32 in __assert_fail () from /lib/x86_64-linux-gnu/libc.so.6
>>> #5 0x00005576cf7c0ee4 in Ip::Address::getAddrInfo (this=this at entry=0x5576d2407a98, dst=@0x7ffea5b87f90: 0x5576d2a8af00, force=force at entry=0) at ./src/ip/Address.cc:663
>>> #6 0x00005576cf7bc3c5 in comm_openex (sock_type=sock_type at entry=1, proto=proto at entry=6, addr=..., flags=1, note=0x5576d2ab0cd0 "[2a03:2880:f273:c8:face:b00c:0:167]") at ./src/comm.cc:347
>>> #7 0x00005576cf80fdfe in Comm::ConnOpener::createFd (this=this at entry=0x5576d2a86bc8) at ./src/comm/ConnOpener.cc:288
>>> #8 0x00005576cf8109c1 in Comm::ConnOpener::start (this=0x5576d2a86bc8) at ./src/comm/ConnOpener.cc:261
>>> #9 0x00005576cf7ab09c in JobDialer<AsyncJob>::dial (this=0x5576d1e16558, call=...) at ../../src/base/AsyncJobCalls.h:175
>>> #10 0x00005576cf7a779f in AsyncCallQueue::fireNext (this=this at entry=0x5576d1abbae0) at ./src/base/AsyncCallQueue.cc:60
>>> #11 0x00005576cf7a7b52 in AsyncCallQueue::fire (this=0x5576d1abbae0) at ./src/base/AsyncCallQueue.cc:43
>>> #12 0x00005576cf55f689 in EventLoop::dispatchCalls (this=0x7ffea5b885a0) at ./src/EventLoop.cc:144
>>> #13 EventLoop::runOnce (this=this at entry=0x7ffea5b885a0) at ./src/EventLoop.cc:121
>>> #14 0x00005576cf55f778 in EventLoop::run (this=this at entry=0x7ffea5b885a0) at ./src/EventLoop.cc:83
>>> #15 0x00005576cf665a50 in SquidMain (argc=argc at entry=5, argv=argv at entry=0x7ffea5b88888) at ./src/main.cc:1719
>>> #16 0x00005576cf4fdae1 in SquidMainSafe (argv=0x7ffea5b88888, argc=5) at ./src/main.cc:1406
>>> #17 main (argc=5, argv=0x7ffea5b88888) at ./src/main.cc:1394
>>>
More information about the squid-users
mailing list