[OpenSIPS-Users] memory leak?

Gary Nyquist gn62 at gmx.us
Mon Nov 3 21:43:06 CET 2014


I have one opensips server deployed on Amazon cloud.

I am using AWS health monitoring service which pings opensips intermittently. (TCP port 5060)

The following error showed up in the log after a few days:
"ERROR:core:tcpconn_new: No more SHM for send chunks pointers"

Tried to know more about this leak (http://www.opensips.org/Documentation/TroubleShooting-OutOfMem)

19:32:24 params(0x7f1e16e97000, 0x7f1e18d53bc0), called from tcp_main.c: tcpconn_new(508)
19:32:24 freeing frag. 0x7f1e18d53b90 alloc'ed from tcp_main.c: tcpconn_new(453)
19:32:24 ERROR:core:handle_new_connect: tcpconn_new failed, closing socket
19:32:25 INFO:core:sig_usr: signal 15 received
19:32:25 Memory status (pkg):
19:32:25 qm_status (0x7f1e1dde7010):
19:32:25  heap size= 2097152
19:32:25  used= 426784, used+overhead=913736, free=1183416
19:32:25  max used (+overhead)= 913736
19:32:25 dumping all alloc'ed. fragments:
19:32:25       0. N  address=0x7f1e1de1c528 frag=0x7f1e1de1c4f8 size=144 used=1
19:32:25             alloc'd from tls/tls_domain.c: tls_new_domain(180)
19:32:25         start check=f0f0f0f0f0f0f0f0, end check= c0c0c0c0c0c0c0c0, abcdefedabcdefed
19:32:25       1. N  address=0x7f1e1de1c618 frag=0x7f1e1de1c5e8 size=144 used=1

But can't understand what could be causing it.
Can anyone advise how to proceed further?

I am using "opensips 1.11.3-tls (x86_64/linux)" git revision: 7e5bbcf

Best Regards,
- Gary

More information about the Users mailing list