Had my Asterisk instance stop responding to incoming/outgoing calls today. Had to kill -9 the asterisk process and restart it to get it back. Not really looking for help on that as the instance is version 1.6 and is due to be replaced with an upgraded version shortly. However, this does make me wonder, do you restart periodically to try to avoid issues or do you just let things run until there is a problem? This box had 119 days of up time on the Asterisk process. I have a client that I installed an Elastix instance on and the last time I checked it, it was up to almost 500 days of up time without an asterisk restart. Kevin Larsen - Systems Analyst - Pioneer Balloon - Ph: 316-688-8208 -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.digium.com/pipermail/asterisk-users/attachments/20130111/c889bc44/attachment.htm>
The general rule seems to be, don't restart it unless there's a problem or you hear of memory leaks. I had a version of 1.4 that I restarted every night because I read about memory leaks, but I hear of 1.2 installs that have been running continuously for 10 years. From: asterisk-users-bounces at lists.digium.com [mailto:asterisk-users-bounces at lists.digium.com] On Behalf Of Kevin Larsen Sent: Friday, January 11, 2013 3:07 PM To: asterisk-users at lists.digium.com Subject: [asterisk-users] How often to restart Asterisk... Had my Asterisk instance stop responding to incoming/outgoing calls today. Had to kill -9 the asterisk process and restart it to get it back. Not really looking for help on that as the instance is version 1.6 and is due to be replaced with an upgraded version shortly. However, this does make me wonder, do you restart periodically to try to avoid issues or do you just let things run until there is a problem? This box had 119 days of up time on the Asterisk process. I have a client that I installed an Elastix instance on and the last time I checked it, it was up to almost 500 days of up time without an asterisk restart. Kevin Larsen - Systems Analyst - Pioneer Balloon - Ph: 316-688-8208 -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.digium.com/pipermail/asterisk-users/attachments/20130111/c81245e0/attachment.htm>
On Fri, Jan 11, 2013 at 2:06 PM, Kevin Larsen < kevin.larsen at pioneerballoon.com> wrote:> However, this does make me wonder, do you restart periodically to try to > avoid issues or do you just let things run until there is a problem? This > box had 119 days of up time on the Asterisk process. I have a client that I > installed an Elastix instance on and the last time I checked it, it was up > to almost 500 days of up time without an asterisk restart. >I've had boxes run for years, and others have problems in a month or two. I have a general practice of having a reboot cron job on critical servers at 3am on Sunday. Our customer SLA allows for a maintenance period during this time. -- Carlos Alvarez TelEvolve 602-889-3003 -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.digium.com/pipermail/asterisk-users/attachments/20130111/8f755f3c/attachment.htm>
>Veering off topic, but still curious :) > >Since an AGI only exists for part of the life of a single call, how >does >it accumulate enough resources to be a problem?Call goes away...AGI thread stays back in some zombi state waiting for something (most of the times response from another web service) That leaves you with an interpreter (probably PHP) fully loaded plus whatever buffers/magic it had acquired stuck in memory. If you are *really lucky* it will be stuck in a blocking function bringing the core its running to 100% utilization and you might notice it. if not you are in for a "random(15) days" crash ....