If you look at the Allstar blue status pages (i.e. http://stats.allstarlink.org/nodeinfo.cgi?node=2530) and note the Actual Uptime you'll see it’s not more than a couple of weeks in most instances. Note: the *CLI> core show uptime shows the same statistic. Also note: this is not the same as the Linux system uptime which has never been a problem. That uptime will be as long as the box has not lost power or been reset. 

The fact that Asterisk reboots at an all to frequent rate has always bugged the hell out of me. Just the other day when our group was supporting the Baker to Vegas Challenge Cup Race one of the repeaters rebooted during the passing of important traffic. Granted, most of the time you won’t even know this is occurring because Asterisk only takes 20 seconds or less to restart or the restart happens when no one is using the repeater. But it’s really messed up when a hub reboots and drops all the connections. I equate that to Asterisk crashing and dropping all the phone calls causing everyone to have to redial. 

For many years I’ve been bugging Jim about this and we’ve been trying all sorts of things to isolate the problem. Long story short, we built a new test environment a few months ago and have found a problem. What was happening is that if there was more than one simultaneous permanent connection reconnect there would be a core dump. Jim added code to hold off reconnects such that only one can occur at a time.  

I don’t know if this is going to fix all the uptime woes but I am optimistic. I’ll be interested to hear what results others have. 

The fix has been posted in svn and you may apply it via the normal update process. 
--
Tim
:wq