asterisk users - Feb 2008 - IAX2 trunks unreliable becoming UNREACHABLE after a time

I have a network of offices using Asterisk that are connected via IAX2
trunks. The trunks work great for a day or two then for no reason at all one
end of the trunk will become UNREACHABLE while the other end is still
connected. The only way to fix the problem is to shutdown Asterisk completly
then start it backup again. The end that dies is not always the same, some
times it is server A and some times it is server B. Never have I seen that
both ends die, just one. The side that is still connected can make calls to
the end that died but not the other way. If you call from the server with
the dead IAX2 trunk you here "All circuts are busy now." All networks
have
static IP addresses and their firewalls are setup to allow UDP 4569 to come
in to the Asterisk systems.

I have been doing a lot of research into this problem. I found this bug
tracker http://bugs.digium.com/view.php?id=5912 that talks about it being an
old problem with  version 1.2.1 using rand() and it not being thread safe.
This I can understand. The thread proposed using rand_r() or ast_random() in
place of rand(), that sounds like a good idea. So when I look at my newer
1.2.18 version I find that it is still using rand() and the bug tracker
continues to be opened and closed and reopened again and again.

Do I dare ask if anyone has a reliable IAX2 trunk? If so how? Should I avoid
using IAX2 all together? I know SIP trunking is an option but it becomes a
real management problem with trying to deal with all the many ports that
need to be open through the firewalls, IAX2 seems like a better way to go if
only it was reliable.

-- 
Open Source: To innovate then create
Proprietary: To imitate then litigate
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.digium.com/pipermail/asterisk-users/attachments/20080210/94e8f77e/attachment.htm

Royce Souther wrote:
>
> Do I dare ask if anyone has a reliable IAX2 trunk? If so how? Should I 
> avoid using IAX2 all together? I know SIP trunking is an option but it 
> becomes a real management problem with trying to deal with all the 
> many ports that need to be open through the firewalls, IAX2 seems like 
> a better way to go if only it was reliable.

Until recently, we were running 1.2.24 between 2 remote facilities and 1 
home office (Mine) with iax2 trunking.  I've never had a trunk fail on 
me.  We're running it over openvpn.  I've since gone to 1.4.15 and have 
had no problems with it dropping either.  We aren't running huge amounts 
of traffic over it.  Maybe 35 to 40 calls per day from each facility.

Doug

-- 
Ben Franklin quote:

"Those who would give up Essential Liberty to purchase a little Temporary
Safety, deserve neither Liberty nor Safety."

I'm having the same exact problem......2 sites connected via the internet, 2
remote sites are unreachable, but the home site finds and can make calls just
fine to the 2 remotes.
 
-Darren
 

________________________________

From: asterisk-users-bounces at lists.digium.com on behalf of Royce Souther
Sent: Sun 2/10/2008 2:33 PM
To: asterisk-users at lists.digium.com
Subject: [asterisk-users] IAX2 trunks unreliable becoming UNREACHABLE aftera
time


I have a network of offices using Asterisk that are connected via IAX2 trunks.
The trunks work great for a day or two then for no reason at all one end of the
trunk will become UNREACHABLE while the other end is still connected. The oving
nly way to fix the problem is to shutdown Asterisk completly then start it
backup again. The end that dies is not always the same, some times it is server
A and some times it is server B. Never have I seen that both ends die, just one.
The side that is still connected can make calls to the end that died but not the
other way. If you call from the server with the dead IAX2 trunk you here
"All circuts are busy now." All networks have static IP addresses and
their firewalls are setup to allow UDP 4569 to come in to the Asterisk systems.

I have been doing a lot of research into this problem. I found this bug tracker
http://bugs.digium.com/view.php?id=5912 that talks about it being an old problem
with  version 1.2.1 using rand() and it not being thread safe. This I can
understand. The thread proposed using rand_r() or ast_random() in place of
rand(), that sounds like a good idea. So when I look at my newer 1.2.18 version
I find that it is still using rand() and the bug tracker continues to be opened
and closed and reopened again and again.

Do I dare ask if anyone has a reliable IAX2 trunk? If so how? Should I avoid
using IAX2 all together? I know SIP trunking is an option but it becomes a real
management problem with trying to deal with all the many ports that need to be
open through the firewalls, IAX2 seems like a better way to go if only it was
reliable.

-- 
Open Source: To innovate then create
Proprietary: To imitate then litigate 

This message was sent from D2 Technology, INC.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/ms-tnef
Size: 4797 bytes
Desc: not available
Url :
http://lists.digium.com/pipermail/asterisk-users/attachments/20080210/be99ab4e/attachment.bin

I may have found a solution to why this problem is happening to me. All my
IAX trunks are up and working and have been for over a day now. If there are
still up and running with no problems in a week I will post again and let
everyone know.

At this point in time it seems the problem was caused by a poorly
constructed initrd file.

My servers all run RAID-1 and my /var/ is mounted on it's own RAID-1. I hand
crafted a new initrd to ensure that RAID-1 started properly and that /var/
was mounted before init runs, it was not before. As of now I am very hopeful
that the problem is gone, all indicators are that this was the solution I
needed.


On Sun, Feb 10, 2008 at 12:33 PM, Royce Souther <osgnuru at gmail.com>
wrote:
> I have a network of offices using Asterisk that are connected via IAX2
> trunks. The trunks work great for a day or two then for no reason at all
one
> end of the trunk will become UNREACHABLE while the other end is still
> connected. The only way to fix the problem is to shutdown Asterisk
completly
> then start it backup again. The end that dies is not always the same, some
> times it is server A and some times it is server B. Never have I seen that
> both ends die, just one. The side that is still connected can make calls to
> the end that died but not the other way. If you call from the server with
> the dead IAX2 trunk you here "All circuts are busy now." All
networks have
> static IP addresses and their firewalls are setup to allow UDP 4569 to come
> in to the Asterisk systems.
>
> I have been doing a lot of research into this problem. I found this bug
> tracker http://bugs.digium.com/view.php?id=5912 that talks about it being
> an old problem with  version 1.2.1 using rand() and it not being thread
> safe. This I can understand. The thread proposed using rand_r() or
> ast_random() in place of rand(), that sounds like a good idea. So when I
> look at my newer 1.2.18 version I find that it is still using rand() and
> the bug tracker continues to be opened and closed and reopened again and
> again.
>
> Do I dare ask if anyone has a reliable IAX2 trunk? If so how? Should I
> avoid using IAX2 all together? I know SIP trunking is an option but it
> becomes a real management problem with trying to deal with all the many
> ports that need to be open through the firewalls, IAX2 seems like a better
> way to go if only it was reliable.
>
> --
> Open Source: To innovate then create
> Proprietary: To imitate then litigate



-- 
Open Source: To innovate then create
Proprietary: To imitate then litigate
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.digium.com/pipermail/asterisk-users/attachments/20080220/a278ea9a/attachment.htm

I am personally Waiting u :) - 

Thanks in advance.

Regards
Bilal
-----------------------
I may have found a solution to why this problem is
happening to me. All
 my
IAX trunks are up and working and have been for over a
day now. If
 there are
still up and running with no problems in a week I will
post again and
 let
everyone know.

At this point in time it seems the problem was caused
by a poorly
constructed initrd file.

My servers all run RAID-1 and my /var/ is mounted on
it's own RAID-1. I
 hand
crafted a new initrd to ensure that RAID-1 started
properly and that
 /var/
was mounted before init runs, it was not before. As of
now I am very
 hopeful
that the problem is gone, all indicators are that this
was the solution
 I
needed.


On Sun, Feb 10, 2008 at 12:33 PM, Royce Souther
 <osgnuru at gmail.com>
 wrote:
> I have a network of offices using Asterisk that are
connected via
 IAX2
> trunks. The trunks work great for a day or two then
for no reason
 at
 all one
> end of the trunk will become UNREACHABLE while the
other end is
 still
> connected. The only way to fix the problem is to
shutdown Asterisk
 completly
> then start it backup again. The end that dies is not
always the
 same,
 some
> times it is server A and some times it is server B.
Never have I
 seen
 that
> both ends die, just one. The side that is still
connected can make
 calls to
> the end that died but not the other way. If you call
from the
 server
 with
> the dead IAX2 trunk you here "All circuts are busy
now." All
 networks
 have
> static IP addresses and their firewalls are setup to
allow UDP
 4569
 to come
> in to the Asterisk systems.
>
> I have been doing a lot of research into this
problem. I found
 this
 bug
> tracker http://bugs.digium.com/view.php?id=5912 that
talks about
 it
 being
> an old problem with  version 1.2.1 using rand() and
it not being
 thread
> safe. This I can understand. The thread proposed
using rand_r() or
> ast_random() in place of rand(), that sounds like a
good idea. So
 when I
> look at my newer 1.2.18 version I find that it is
still using
 rand()
 and
> the bug tracker continues to be opened and closed
and reopened
 again
 and
> again.
>
> Do I dare ask if anyone has a reliable IAX2 trunk?
If so how?
 Should
 I
> avoid using IAX2 all together? I know SIP trunking
is an option
 but
 it
> becomes a real management problem with trying to
deal with all the
 many
> ports that need to be open through the firewalls,
IAX2 seems like
 a
 better
> way to go if only it was reliable.
>
> --
> Open Source: To innovate then create
> Proprietary: To imitate then litigate




     
____________________________________________________________________________________
Be a better friend, newshound, and 
know-it-all with Yahoo! Mobile.  Try it now. 
http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ