asterisk users - Apr 2010 - long return times from System() calls with 1.6.2.6?

I've just upgraded to 1.6.2.6 on one of my test systems. I started out
happy, with some improvements in transfers to Local() channels from a
SIP channel, and much nicer verbose fax handling.

However, something is really weird when I need to do System() calls.
It was really, really weird. This was also affecting AGI, when I
needed to read system variables from asterisk into an AGI Perl script.

I have a second test system, with asterisk-1.6.2.6, and there are not
these problems with that system.

So I suspect something whacky that really probably has nothing to do
with asterisk.

It almost feels like delay in reading loopback, or running out of
available files on the system, or something like that. I'm rebooted,
and the problem did not go away.

I'm doing really, really innocent things, like:

exten => s,1,Verbose(EXTENSION is: ${EXTEN})
exten => s,n,Set(MESSAGE_PATH=/path/to/message/)
exten => s,n,System(test -e ${MESSAGE_PATH}${EXTEN})
exten => s,n,Verbose(System call result was ${SYSTEMSTATUS})
exten => s,n,ExecIf($[${SYSTEMSTATUS} =
SUCCESS]?Playback(${OVERFLOW_GENERIC}))
exten => s,n,ExecIf($[${SYSTEMSTATUS} = SUCCESS]?HangUp())
exten => s,n,Goto(Next_context,s,1)

That runs lightning-fast on every system, but not on this one. There
is a huge pause, like two seconds, waiting for the System() call to
return. Dead air is not cool when setting up messaging on a phone
system.

Ideas?

On Thu, Apr 8, 2010 at 4:30 PM, David Backeberg <dbackeberg at gmail.com>
wrote:
> However, something is really weird when I need to do System() calls.
> It almost feels like delay in reading loopback, or running out of
> available files on the system, or something like that. I'm rebooted,
> and the problem did not go away.
>
> I'm doing really, really innocent things, like:
>
> exten => s,n,System(test -e ${MESSAGE_PATH}${EXTEN})
So I did some more testing. Same dialplan, reverted to
asterisk-1.6.0.13, and the contexts that do these test -e calls runs
lightning fast. It's like maybe there's something going on where it
needs to run sudo or something?

Took iptables down, no change.

I run asterisk as non-root.

I then tried running asterisk as root. Same problem, so that doesn't
seem to be it.

went looking around asterisk.conf, didn't see any pertinent settings,
but then again, it's been a long time since I've built 'make
examples', so maybe there's a new setting in there that controls calls
out to the system?

I am investigating High Availability solutions for my front end servers.

I got into a discussion regarding "replicated local databases" versus
"
a single fiber connected shared database" on an EMC. 

Is anyone running a shared database on a SAN? Care to comment on your
findings...

Thanks,

Robert

In case anybody was following this thread,
wanted to let people know that the fix made it into SVN,
and is packaged into
1.6.2.8-rc1

Huge thanks to Kevin and Tilghman

On Wed, Apr 21, 2010 at 3:40 PM, David Backeberg <dbackeberg at gmail.com>
wrote:
> issue opened.
>
> https://issues.asterisk.org/view.php?id=17223
>