Hi all,
Currently i am testing ejabberd on AWS. But, Sometimes segfault is happening
In a 16 core environment(c4.4xlarge type), one instance runs well, but after 2-3 days only one core is 100%. (other cores are 0%)
Also all sessions are disconnected. At the same time, segfault occurs in ejabberd.(16.01)
This usually happens when there are many sessions. I'm also using Redis(PaaS) for sessions management and mssql for user information & Roster.
*OS
ubuntu 14.04.3 LTS ( kernel : 3.13.0-74-generic)
*Erlang Option
+K true +P 4000000 +A 10 +S 16:16 +sfwi 500 +scl false +sub true +stbt db +sbwt very_long +swt very_low +Mulmbcs 32767 +Mumbcgs 1 +Musmbcs 2047
*Erlang version
Erlang/OTP 18.2
*syslog
Nov 1 04:49:40 ip-192-168-7-151 kernel: [ 9319.786176] show_signal_msg: 27 callbacks suppressed
Nov 1 04:49:40 ip-192-168-7-151 kernel: [ 9319.786181] 11_scheduler[1955]: segfault at 7fac02e16c68 ip 0000000000521fbd sp 00007facf67f49b0 error 4 in beam.smp[400000+22b000]
Nov 1 04:50:03 ip-192-168-7-151 kernel: [ 9342.931284] TCP: TCP: Possible SYN flooding on port 5280. Sending cookies. Check SNMP counters.
Here is a contents of core dump
(gdb) bt
#0 0x00007f99e99d2095 in ?? () from /lib/x86_64-linux-gnu/libexpat.so.1
#1 0x00007f99e99d23fd in ?? () from /lib/x86_64-linux-gnu/libexpat.so.1
#2 0x00007f99e99d46b5 in ?? () from /lib/x86_64-linux-gnu/libexpat.so.1
#3 0x00007f99e99d7924 in ?? () from /lib/x86_64-linux-gnu/libexpat.so.1
#4 0x00007f99e99d8641 in ?? () from /lib/x86_64-linux-gnu/libexpat.so.1
#5 0x00007f99e99da5ef in XML_ParseBuffer () from /lib/x86_64-linux-gnu/libexpat.so.1
#6 0x00007f99e9bfae64 in parse_nif (env=0x7f9a06959df0, argc=2, argv=0x7f9a1744c2c0) at c_src/xml_stream.c:804
#7 0x0000000000444045 in process_main ()
#8 0x00000000004d051a in ?? ()
#9 0x0000000000621d4b in ?? ()
#10 0x00007f9a19555184 in start_thread (arg=0x7f9a0695a700) at pthread_create.c:312
#11 0x00007f9a1907a37d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:111
Thank You in advance! :)
Joseph
