F5 LTM freqüentemente mata processos com SIGKILL

Question

Temos um dispositivo BIP-IP 6400 LTM que está eliminando processos com uma frequência alarmante. A CPU é consistentemente em torno de 23% de utilização, portanto, isso não é um problema.

Aqui está uma amostra de /var/log/ltm :

Oct  7 08:21:55 local/pri-4600 info bigd[3471]: reap_child: child process PID = 25338 exited with signal = 9
Oct  7 08:22:15 local/pri-4600 info bigd[3471]: reap_child: child process PID = 25587 exited with signal = 9
Oct  7 08:22:34 local/pri-4600 info bigd[3471]: reap_child: child process PID = 25793 exited with signal = 9
Oct  7 08:23:10 local/pri-4600 info bigd[3471]: reap_child: child process PID = 26260 exited with signal = 9
Oct  7 08:23:36 local/pri-4600 info bigd[3471]: reap_child: child process PID = 26584 exited with signal = 9
Oct  7 08:23:40 local/pri-4600 info bigd[3471]: reap_child: child process PID = 26647 exited with signal = 9
Oct  7 08:23:45 local/pri-4600 info bigd[3471]: reap_child: child process PID = 26699 exited with signal = 9
Oct  7 08:23:55 local/pri-4600 info bigd[3471]: reap_child: child process PID = 26805 exited with signal = 9
Oct  7 08:25:36 local/pri-4600 info bigd[3471]: reap_child: child process PID = 28079 exited with signal = 9
Oct  7 08:27:15 local/pri-4600 info bigd[3471]: reap_child: child process PID = 29286 exited with signal = 9
Oct  7 08:27:16 local/pri-4600 info bigd[3471]: reap_child: child process PID = 29307 exited with signal = 9
Oct  7 08:27:56 local/pri-4600 info bigd[3471]: reap_child: child process PID = 29793 exited with signal = 9
Oct  7 08:29:20 local/pri-4600 info bigd[3471]: reap_child: child process PID = 30851 exited with signal = 9
Oct  7 08:33:00 local/pri-4600 info bigd[3471]: reap_child: child process PID = 1122 exited with signal = 9
Oct  7 08:33:16 local/pri-4600 info bigd[3471]: reap_child: child process PID = 1299 exited with signal = 9
Oct  7 08:34:15 local/pri-4600 info bigd[3471]: reap_child: child process PID = 2054 exited with signal = 9
Oct  7 08:35:16 local/pri-4600 info bigd[3471]: reap_child: child process PID = 2784 exited with signal = 9
Oct  7 08:35:16 local/pri-4600 info bigd[3471]: reap_child: child process PID = 2807 exited with signal = 9
Oct  7 08:35:35 local/pri-4600 info bigd[3471]: reap_child: child process PID = 3015 exited with signal = 9
Oct  7 08:36:15 local/pri-4600 info bigd[3471]: reap_child: child process PID = 3601 exited with signal = 9

Isso é normal? Se não, o que poderia estar causando isso?

    

Answer 1

Este era um bug conhecido no software 10.2.4 BIG-IP que estávamos rodando.

Do suporte da F5:

...you hit a known issue tracked internally as: bug ID539130 "bigd can deadlock while processing SIGCHLD causing bigd heartbeat failure and SIGABRT" -=Condition=- External monitors that run for a long time and are killed by the next iteration of the monitor, may cause bigd to crash and core, this causes a temporary lapse in health monitoring.

A correção foi atualizar o software com Hotfix-BIGIP-10.2.4-HF12-866.11-ENG .

    

Answer 2

O bigd é o daemon de monitoramento no BIG-IP e, portanto, parece que um monitor que está em uso está falhando. Você deve abrir um caso com suporte e enviar seu qkview para ihealth.f5.com. Aqui está uma solução relacionada a essa mensagem de erro:

link