Hi John, On Mon, Jun 30, 2025 at 04:20:15PM -0400, John Lauro wrote: > Had a couple servers running 3.2.x core dump in the last 2 days (both > built from source). I think they were running for over a week. > > Here is a backtrace of one: > (gdb) bt > #0 0x00007f06c15015ef in raise () from /lib64/libc.so.6 > #1 0x00007f06c14d4e65 in abort () from /lib64/libc.so.6 > #2 0x00000000005aa45c in ha_panic () at src/debug.c:826 > #3 ha_panic () at src/debug.c:757 > #4 0x00000000006a9fd5 in wdt_handler (sig=<optimized out>, > si=<optimized out>, arg=<optimized out>) at src/wdt.c:190 > #5 <signal handler called> > #6 0x0000000000698e25 in fwlc_get_next_server (p=<optimized out>, > srvtoavoid=srvtoavoid@entry=0x2ad148f0) at src/lb_fwlc.c:772 > #7 0x000000000056e851 in assign_server (s=s@entry=0x7f06bc06add0) at > src/backend.c:706 > #8 0x000000000056ecbb in assign_server_and_queue > (s=s@entry=0x7f06bc06add0) at src/backend.c:985 > #9 0x00000000005728e8 in assign_server_and_queue (s=0x7f06bc06add0) > at include/haproxy/queue.h:89 > #10 srv_redispatch_connect (s=s@entry=0x7f06bc06add0) at src/backend.c:2256 > #11 0x0000000000572c47 in back_handle_st_req (s=0x7f06bc06add0) at > src/backend.c:2581 > #12 0x00000000004f277f in process_stream (t=<optimized out>, > context=context@entry=0x7f06bc06add0, state=<optimized out>) at > src/stream.c:2376 > #13 0x000000000065703e in run_tasks_from_lists > (budgets=budgets@entry=0x7f06c14a7880) at src/task.c:655 > #14 0x0000000000657b24 in process_runnable_tasks () at src/task.c:889 > #15 0x00000000005be988 in run_poll_loop () at src/haproxy.c:2864 > #16 0x00000000005beff7 in run_thread_poll_loop (data=<optimized out>) > at src/haproxy.c:3080 > #17 0x00007f06c18921da in start_thread () from /lib64/libpthread.so.0 > #18 0x00007f06c14ec8d3 in clone () from /lib64/libc.so.6 > > Not sure how to reproduce, but it's been a long time since I last had > a crash. and I think these might be the only servers I am running 3.x, > most running 2.8.
Thanks for reporting! Something very similar has been recently fixed, so hopefully it is the same problem, and 3.2.2 should be fine. If you feel like applying the patch on top of 3.2.1 and recompiling, you can just apply commit e28e647fef43e5865c87f328832fec7794a423e5. Regards, Olivier
