[asterisk-dev] [Code Review] dont poke registered sip peers immediately after sip reload to avoid packet storm

[schmidts]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/1652/
-----------------------------------------------------------

Review request for Asterisk Developers.


Summary
-------

with a big amount of sip peers registered doing a sip reload causes a packet flood of sip OPTIONS messages cause they are immediately sent out when the peer is loaded back from the astdb.
This fix prevents this packet storm and just schedule a poke in a random value between 1 ms and peers qualifyfreq or the global qualifyfreq. 
Another very small fix is to schedule a poke on peers with qualify disabled. This doesnt cost much but its just unnecessary to do this.

maybe it would be even better to dont call the poke function from source_db on a reload cause poking all peers is done after this again.


This addresses bug ASTERISK-19154.
    https://issues.asterisk.org/jira/browse/ASTERISK-19154


Diffs
-----

  branches/1.8/channels/chan_sip.c 349449 

Diff: https://reviewboard.asterisk.org/r/1652/diff


Testing
-------

tested with 4500 peers with qualify=no and no poke is scheduled after a sip reload.
tested with 2500 peers with qualify=yes and pokes are scheduled in the right time (1 to 60000 ms)


Thanks,

schmidts

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/asterisk-dev/attachments/20120105/1f1f3b7a/attachment-0001.htm>