> @@ -1365,7 +1377,7 @@ static struct mi_root* ds_mi_list(struct mi_root*
> cmd_tree, void* param)
> 9, part_it->name.s, part_it->name.len);
>
> if (part_node == NULL
> - || ds_print_mi_list(part_node, part_it) < 0) {
> +
> @@ -2110,6 +2123,27 @@ int ds_print_mi_list(struct mi_node* rpl,
> ds_partition_t *partition)
> if(node1 == NULL)
> goto error;
> }
> +
> + if (flags & 1) {
Please define the flag "1"
`-DDBG_QM_MALLOC` was restructured (see
https://github.com/OpenSIPS/opensips/pull/736), so it does not have any impact
anymore on master.
In short, each of the 3 current stable allocators now has its debugging
version, available through `-DDBG_MALLOC`. Readability of the entire fragment
dump
Hi Bogdan, please drop me a note if possible when you merge that patch
into public branch[es], I'll adjust my test case to verify that
accordingly.
On the second issue, it basically about handling of INVITE
transactions in the tm module. The issue is that given unique, new
INVITE transaction
Hello everyone!
Jan 12-13, the YUM repository (yum.opensips.org) will be down due to
maintenance.
--
Nick
___
Devel mailing list
Devel@lists.opensips.org
http://lists.opensips.org/cgi-bin/mailman/listinfo/devel
Hi Maxim,
A Happy New Year to you too ! Sorry for the slow reply, many things to
catch up with.
So, I understand the original problem (retransmissions and cancelling
for slowly answered invites) is solved by the patch I sent you. Hoping
the are no side effects on this change, I can start
Hello!
I use opensips 211:
version: opensips 211 (x86_64/linux)
flags: STATS: On, DISABLE_NAGLE, USE_MCAST, SHM_MMAP, PKG_MALLOC, F_MALLOC,
FAST_LOCK-ADAPTIVE_WAIT
ADAPTIVE_WAIT_LOOPS=1024, MAX_RECV_BUFFER_SIZE 262144, MAX_LISTEN 16,
MAX_URI_SIZE 1024, BUF_SIZE 65535
poll method support: poll,
Branch: refs/heads/master
Home: https://github.com/OpenSIPS/opensips
Commit: b32a14e6536e4a651eb9fcc0d26dcf034a8b8cda
https://github.com/OpenSIPS/opensips/commit/b32a14e6536e4a651eb9fcc0d26dcf034a8b8cda
Author: ionutrazvanionita
Date: 2016-01-11
Hey Raz,
I repeated my previous test: open few tabs, close them, repeat. but I dumped
the statistics, and did a list_tcp_conns between each test. All the tcp conns
seem to tear down properly, but pkg memory use does grow over time. As you can
see by the connection number in my gist, I had to
Earlier, one process handled nat_traversal timeouts, so static
iteration counter correctly distributed keepalives across timeout
interval.
Now timeouts are sent to various processes, each of which had its own
counter, resulting in erratic keepalive delivery.
Fix moves the iteration counter into
Hi, Eric!
I see a lot of memory used by the structures that store the connections. Could
you try to list all connections just before running out of memory and let me
know how many active connections you have? You can use `opensipsctl fifo
list_tcp_conns`.
Also, how many children are you
I am only using 2 children.
I will try to get it to run out of memory again and see how many active
connections there are.
---
Reply to this email directly or view it on GitHub:
Branch: refs/heads/master
Home: https://github.com/OpenSIPS/opensips
Commit: f471785b3e75925d78b6bda63c1ee1d704fc93c7
https://github.com/OpenSIPS/opensips/commit/f471785b3e75925d78b6bda63c1ee1d704fc93c7
Author: Cerghit Ionel
Date: 2016-01-08 (Fri, 08
Merged #748.
---
Reply to this email directly or view it on GitHub:
https://github.com/OpenSIPS/opensips/pull/748#event-511290683___
Devel mailing list
Devel@lists.opensips.org
http://lists.opensips.org/cgi-bin/mailman/listinfo/devel
Thank you for the comments Bogdan! I made the changes.
---
Reply to this email directly or view it on GitHub:
https://github.com/OpenSIPS/opensips/pull/748#issuecomment-170584084___
Devel mailing list
Devel@lists.opensips.org
RTPProxy has the Q command ( both in 12 and 20 releases ) which allows for
fetching statistics on a per call basis
Implement a mechanism ( function or pvar ) for fetching them - eg fetching the
call stats at BYE time, for storing them in CDRs or for attaching them to a BYE
header ( similar to
I recompiled with -DBG_MALLOC as suggested by liviu.
I started the proxy, opened a few tabs to register webrtc clients over wss,
closed the tabs and repeated the process till the proxy ran out of memory. I
then sent sigusr1 signals to all the pids and got the following log output.
17 matches
Mail list logo