Tzafrir Cohen wrote:
What I say is that you have the worse of both worlds:
- downtime of at ~1/2 a minute (avarage, if a cron runs every minute).
In the case a restart is all it take.
- A bigger downtime in case a restart is not what it takes. Because your
logs will be flooded.
- And a most unpredicatable behaviour.
So why would Asterisk crash? And if so: would a simple restart really
do?
Tzafrir;
Well I would seriously disagree. Assuming that asterisk just hit a
bump in the road and needs a restart then an average downtime of 30
seconds is minimal. I'm assuming that scripts like safe_asterisk
include restart delays anyway, just to let things settle before trying
again. So there is really is no difference here. As for bigger
downtime, well if asterisk won't restart your startup method isn't gonna
change anything. Most monitoring scripts will keep trying just as cron
will the difference is that the script can be set to stop after a
certain number of time (maybe a dozen or so depends on the config). So
your logs will be just as full anyway. And the behavior is in no way
unpredictable. If anything it is predictable with clock-work efficiency
(please excuse the pun).
Lets face it that in a production environment if asterisk dies the
manager that is on will be calling from their cell phone to get someone
to fixit ASAP. I did say originally that this was a simple and less
then perfect way to handle things, but it does work well. If you don't
like it that is just fine. I have found that it is a reliable way to
insure that asterisk will restart on a stable system that does not
support "inittab".
Mark C.
_______________________________________________
--Bandwidth and Colocation provided by Easynews.com --
asterisk-users mailing list
To UNSUBSCRIBE or update options visit:
http://lists.digium.com/mailman/listinfo/asterisk-users
