Hi All, I'm interested in how people are "clustering" Asterisk, if that's possible, or how you might be achieving a redundant solution. I've a single Asterisk server driving the company. Its well backed-up, and I've a cloned machine that (in theory) with a DNS change could take over operations. However I'd like to achieve something more automated if possible. I know that some of my VoIP trunk providers "cluster" IAX connections, but I'm not sure how they would do that. Any ideas? Adrian Marsh ?
Adrian Marsh wrote:> I'm interested in how people are "clustering" Asterisk, if that's > possible, or how you might be achieving a redundant solution. > I've a single Asterisk server driving the company. Its well > backed-up, and I've a cloned machine that (in theory) with a DNS > change could take over operations. > > However I'd like to achieve something more automated if possible.I haven't looked into it in any detail, but how about the standard Linux HA solution with a heartbeat monitor, a shared file-system and IP take-over? /Per Jessen, Z?rich -- http://www.spamchek.com/ - your spam is our business.
Adrian Marsh wrote:> I'm interested in how people are "clustering" Asterisk, if that's possible, or how you might be achieving a redundant solution. > I've a single Asterisk server driving the company. Its well backed-up, and I've a cloned machine that (in theory) with a DNS change could take over operations. > > However I'd like to achieve something more automated if possible.Maybe my post at http://lists.digium.com/pipermail/asterisk-users/2007-August/195339.html could provide you with some answers. I don't want to quote my text as not to spam the list (although it's all GPL). There's a nice countdown at http://www.amooma.de/gemeinschaft/ but we're all quite busy at the moment (that page still needs to be translated). Regards, Philipp Kempgen -- amooma GmbH - Bachstr. 126 - 56566 Neuwied - http://www.amooma.de Let's use IT to solve problems and not to create new ones. Asterisk? -> http://www.das-asterisk-buch.de My pick of the month: rfc 2822 3.6.5 Gesch?ftsf?hrer: Stefan Wintermeyer Handelsregister: Neuwied B 14998
It's nice to see Asterisk redundancy being discussed. A year and half ago, when I posed the question of Asterisk redundancy, I was looked at like I was from outer space. ----- Original Message ---- From: Jared Smith <jsmith at digium.com> To: Asterisk Users Mailing List - Non-Commercial Discussion <asterisk-users at lists.digium.com> Sent: Tuesday, September 25, 2007 7:27:37 AM Subject: Re: [asterisk-users] Asterisk Redundancy On Tue, 2007-09-25 at 15:59 +0200, Per Jessen wrote:> I haven't looked into it in any detail, but how about the standard Linux > HA solution with a heartbeat monitor, a shared file-system and IP > take-over?It's been my experience that this usually works fairly well for stateless protocols like HTTP, but doesn't do so well on stateful protocols like SIP and IAX, and in general is a much more difficult problem to solve. Most people tend to use some combination of SIP proxies (such as SER and OpenSER), DUNDi, shared storage, redundant databases with replication, T1/E1 failover boxes, and horizontal scaling to make Asterisk more highly-available. Of course, I haven't really gone into much detail here, but hopefully it helps answer your question. (It's also my personal experience that people who know how to build such solutions are making enough money off of selling their solution that they aren't real eager to give away all their secrets.) In reality though, you say the word "cluster" and it means five different things to five different people. To really be able to answer the original poster's question, we'd really have to know a lot more about his architecture and his potential points of failure. -- Jared Smith Community Relations Manager Digium, Inc. _______________________________________________ Sign up now for AstriCon 2007! September 25-28th. http://www.astricon.net/ --Bandwidth and Colocation Provided by http://www.api-digital.com-- asterisk-users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users ____________________________________________________________________________________ Check out the hottest 2008 models today at Yahoo! Autos. http://autos.yahoo.com/new_cars.html -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20070925/7da20c93/attachment.htm
Nagios.... that's not redundancy. ----- Original Message ---- From: Dave Walker <DaveWalker at ubuntu.com> To: Asterisk Users Mailing List - Non-Commercial Discussion <asterisk-users at lists.digium.com> Sent: Tuesday, September 25, 2007 9:09:46 AM Subject: Re: [asterisk-users] Asterisk Redundancy On Tue, 2007-09-25 at 18:01 +0200, Philipp Kempgen wrote:> Adrian Marsh wrote: > > > so maybe it's a case of looking at > > Linux-HA. > > If I remember this correctly a normal ping is all Linux HA can > do. It does not check whether Asterisk or other services are > alive and respond to queries.Have you looked at: http://www.voip-info.org/wiki-Asterisk+monitoring My personal favourite would be nagios (not that I have used the SIP plugin, but do use nagios for other services) Kind Regards, Dave Walker ____________________________________________________________________________________ Tonight's top picks. What will you watch tonight? Preview the hottest shows on Yahoo! TV. http://tv.yahoo.com/ -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20070925/fe181816/attachment.htm
>----- Original Message ---- >From: Atis Lezdins <atis at iq-labs.net> >To: Asterisk Users Mailing List - Non-Commercial Discussion <asterisk-users at lists.digium.com> >Sent: Tuesday, September 25, 2007 2:11:10 PM >Subject: Re: [asterisk-users] Asterisk Redundancy > >On 9/25/07, Philipp Kempgen <philipp.kempgen at amooma.de> wrote: >> Adrian Marsh wrote: >> >> > I'm interested in how people are "clustering" Asterisk, if that's possible, or how you might be achieving a redundant solution. >> > I've a single Asterisk server driving the company. Its well backed-up, and I've a cloned machine that (in theory) with a DNS change could take over >operations. >> > >> > However I'd like to achieve something more automated if possible.. >> >> Maybe my post at >> http://lists.digium.com/pipermail/asterisk-users/2007-August/195339.html >> could provide you with some answers. > > >Hi, >This seems nice way of sharing settings, however it wouldn't take over >calls in progress. For us, currently the greatest problem is that >whenever Asterisk crashes, calls are lost, and that means - lost >money. Are there any ideas?You might want to take Asterisk out of the media path then. If it crashes, calls will stay up, although your CDR's will be screwed. If screwed CDR's still means lost money... your still screwed! Doug. ____________________________________________________________________________________ Pinpoint customers who are looking for what you sell. http://searchmarketing.yahoo.com/ -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20070925/a7d5aeb9/attachment.htm
>----- Original Message ---- >From: Scott Moseman <scmoseman at gmail.com> >To: Asterisk Users Mailing List - Non-Commercial Discussion <asterisk-users at lists.digium.com> >Sent: Wednesday, September 26, 2007 6:07:06 AM >Subject: Re: [asterisk-users] Asterisk Redundancy > >On 9/26/07, SIP <sip at arcdiv.com> wrote: >> >> No. It's not. But there still exists the possibility even in a >> relatively stable situation that the software could crash or that >> hardware could fail. It's best, when planning a highly-available >> solution, to plan for the unforeseen and not assume you can >> avoid all mishaps. Let's assume, for the sake of argument, that >> the software will NEVER fail. Hardware still might, and that would >> still mean a lost call unless there's a way to switch running calls >> over to a new server seamlessly. >> > >Also be sure that you have a very redundant network configuration. >Too often I see people spend a great deal of time and money to get >redundant servers when their switches, firewalls, routers, etc are not >even capable of handling a failed network element.You can achieve this at the application level. ____________________________________________________________________________________ Moody friends. Drama queens. Your life? Nope! - their life, your story. Play Sims Stories at Yahoo! Games. http://sims.yahoo.com/ -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20070927/9f93252c/attachment.htm
>----- Original Message ---- >From: SIP <sip at arcdiv.com> >To: Asterisk Users Mailing List - Non-Commercial Discussion <asterisk-users at lists.digium..com> >Sent: Wednesday, September 26, 2007 4:31:08 AM >Subject: Re: [asterisk-users] Asterisk Redundancy > >Per Jessen wrote: >> Atis Lezdins wrote: >> >> >>> This seems nice way of sharing settings, however it wouldn't take over >>> calls in progress. For us, currently the greatest problem is that >>> whenever Asterisk crashes, calls are lost, and that means - lost >>> money. Are there any ideas? >>> >> >> Perhaps investigate/diagnose the craches? Software instability is not >> solved with a high-availability solution. IMHO. >> >> >> /Per Jessen, Z?rich >> >> >No. It's not. But there still exists the possibility even in a >relatively stable situation that the software could crash or that >hardware could fail. It's best, when planning a highly-available >solution, to plan for the unforeseen and not assume you can avoid all >mishaps. Let's assume, for the sake of argument, that the software will >NEVER fail. Hardware still might, and that would still mean a lost call >unless there's a way to switch running calls over to a new server >seamlessly. > >Are there such ways? IP calls are especially troublesome in that regard.Don't set your goals too high. I've worked for a few companies with Asterisk now and just having an architecture that can recover within a few seconds and process new calls almost seamlessly is a workable goal. Having an architecture that can seamlessly fail over and keep calls up is kinda like the whole grail of redundancy with Asterisk. Hint... you might be able to do it with SIP reinvites... Doug. ____________________________________________________________________________________ Yahoo! oneSearch: Finally, mobile search that gives answers, not web links. http://mobile.yahoo.com/mobileweb/onesearch?refer=1ONXIC -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20070927/383d6c3c/attachment.htm
Douglas Garstang wrote:>>Also be sure that you have a very redundant network configuration. >>Too often I see people spend a great deal of time and money to get >>redundant servers when their switches, firewalls, routers, etc are not >>even capable of handling a failed network element. > > You can achieve this at the application level.How do you do that when your single network connection is gone? When considering redundancy it is essential that you have no single point of failure. Depending on how far you want to go, this means right from your dual-box asterisk setup to dual diesel-generators and two multi-homed datacenters. /Per Jessen, Z?rich -- http://www.spamchek.com/ - your spam is our business.
Adrian Marsh wrote: interested in how people are "clustering" Asterisk, if that's possible, or how you might be achieving a redundant solution.> I've a single Asterisk server driving the company. Its well backed-up, and I've a cloned machine that (in theory) with a DNS change could take over operations. > > However I'd like to achieve something more automated if possible.The Carrier Class project, which is based on Asterisk and OpenSER, looks promising: http://www.carrierclass.net/