Good day, We use Haproxy version '1.5-dev19' and we want to report a bug for you.
We had one problem on our haproxy servers with 100% load on CPU cores. It seems like worker process has gone to the eternal cycle. Problem was resolved only with haproxy process restart. Haproxy process stack trace : -------------------------------------------------------------------------- #0 0x0000000000443b7b in chash_get_next_server (p=0x6d1810, srvtoavoid=0x0) at src/lb_chash.c:348 348 src/lb_chash.c: No such file or directory. in src/lb_chash.c (gdb) bt #0 0x0000000000443b7b in chash_get_next_server (p=0x6d1810, srvtoavoid=0x0) at src/lb_chash.c:348 #1 0x000000000044246e in assign_server (s=0x284d2d0) at src/backend.c:615 #2 0x0000000000442636 in assign_server_and_queue (s=0x1e) at src/backend.c:791 #3 0x0000000000442771 in srv_redispatch_connect (t=0x284d2d0) at src/backend.c:1030 #4 0x000000000045b278 in process_session (t=0x288e150) at src/session.c:1180 #5 0x000000000040ef55 in process_runnable_tasks (next=0x7fffd97d702c) at src/task.c:238 #6 0x0000000000405d28 in run_poll_loop () at src/haproxy.c:1261 #7 0x0000000000407e6c in main (argc=<value optimized out>, argv=0x7fffd97d7248) at src/haproxy.c:1592 -------------------------------------------------------------------------- In you archives http://marc.info/?l=haproxy we didn't find the similar problem with 100% CPU load. Possibly you know about this problem in '1.5-dev19' Thank you in advance ----------------------------- With best regards Roman Lazarenko Odnoklassniki.ru Mail.ru Group e-mail: [email protected]<mailto:[email protected]>
