squid v5 crashes with SIGABRT when ipv6 is disabled at kernel level but it is asked to connect to an ipv6 address by a client

Issue

• Disabling IPV6 causes squid to crash dump and restart

  1. Disable ipv6 using sysctl

   # vi /etc/sysctl.d/disable_v6.conf
       net.ipv6.conf.all.disable_ipv6=1
       net.ipv6.conf.default.disable_ipv6=1
  

  2. Restart server with squid enabled

  3. Wait for an IPV6 resource request
  
  

• In /var/log/messages log

  $ cat messages | grep systemd-coredump | sed s'/#012/\n/g'
  systemd-coredump[XXXXX]: Process XXXXX (squid) of user 23 dumped core.
  Stack trace of thread 238044:
  #0 0x00007f28d68a154c __pthread_kill_implementation (libc.so.6 + 0xa154c)
  #1 0x00007f28d6854d06 raise (libc.so.6 + 0x54d06)
  #2 0x00007f28d68287f3 abort (libc.so.6 + 0x287f3)
  #3 0x00007f28d682871b __assert_fail_base.cold (libc.so.6 + 0x2871b)
  #4 0x00007f28d684dca6 __assert_fail (libc.so.6 + 0x4dca6)
  #5 0x0000558ab96650b9 _ZNK2Ip7Address11getAddrInfoERP8addrinfoi (squid + 0x3cf0b9)
  #6 0x0000558ab966225a _ZL11comm_openexiiRN2Ip7AddressEiPKc.lto_priv.0 (squid + 0x3cc25a)
  #7 0x0000558ab96ae9a1 _ZN4Comm10ConnOpener8createFdEv (squid + 0x4189a1)
  #8 0x0000558ab96afed8 _ZN4Comm10ConnOpener5startEv (squid + 0x419ed8)
  #9 0x0000558ab9655eae _ZN9JobDialerI8AsyncJobE4dialER9AsyncCall (squid + 0x3bfeae)
  #10 0x0000558ab96545db _ZN14AsyncCallQueue8fireNextEv (squid + 0x3be5db)
  #11 0x0000558ab965496a _ZN14AsyncCallQueue4fireEv (squid + 0x3be96a)
  #12 0x0000558ab943b961 _ZN9EventLoop7runOnceEv (squid + 0x1a5961)
  #13 0x0000558ab953da08 _Z9SquidMainiPPc (squid + 0x2a7a08)
  #14 0x0000558ab93e0656 main (squid + 0x14a656)
  #15 0x00007f28d683feb0 __libc_start_call_main (libc.so.6 + 0x3feb0)
  #16 0x00007f28d683ff60 __libc_start_main@@GLIBC_2.34 (libc.so.6 + 0x3ff60)
  #17 0x0000558ab93ead05 _start (squid + 0x154d05)
  ELF object binary architecture: AMD x86-64