VoIP Mailing List Archives

[srinivas.ksvreddy at g...]

Hi All,

please let me know implementation of failover in freeswitch.

Thanks
Srinivas

[dwhite at olp.net]

On 29/09/09 16:37 +0530, srinivasula reddy wrote:

[msc at freeswitch.org]

This topic has been beaten to death recently. Search the archives for things like "redundancy" and "failover" and you'll see lots of discussions. The bottom line is that your needs will dictate how much time, effort, and money you are willing to sink into this. If you want professional assistance then email consulting@freeswitch.org (consulting@freeswitch.org). If you want to do your own research then I'd say start with the ClueCon videos, specifically Day 3, Presentation #5. Here's the torrent:
http://files.freeswitch.org/cluecon_2009/presentations/cluecon_2009.torrent

-MC

On Tue, Sep 29, 2009 at 8:16 AM, Dan White <dwhite@olp.net (dwhite@olp.net)> wrote:

[dwhite at olp.net]

I've downloaded 0Kb in 56 minutes. Anyone mind setting up a seeder? I'm on
a pretty fast connection.

While waiting, I found the following discussions:

I found the following threads while searching for redundancy and failover:
http://lists.freeswitch.org/pipermail/freeswitch-users/2007-September/001499.html
(round robin)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-April/013069.html
(load balancing, scalability)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-August/017349.html
(a discussion of setting up multiple FS boxes in multiple datacenters)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-December/009723.html
(comparing FS to a commercial product)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-February/011770.html
(WAN redundancy)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-February/010924.html
(Comparing against commercial SBCs)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-September/006710.html
(Hardware redundancy)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-January/001934.html
(Generic heartbeat failover question)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-August/017972.html
(Failing over an originating call based on originate_timeout)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-August/005740.html
(HA via Ultra Monkey)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-April/013097.html
(gateway redundancy)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-August/005688.html
(redundancy via openser)
http://lists.freeswitch.org/pipermail/freeswitch-users/2007-August/001363.html
(failover based on xml-curl)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-July/016904.html
(multiple gateways)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-February/002146.html
(dead gateway detection)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-June/003979.html
(load balancing and failover via mod_xml_curl)
http://lists.freeswitch.org/pipermail/freeswitch-users/2007-September/001445.html
(multiple gateways)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-June/015363.html
(hot failover)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-May/014071.html
(failover extension)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-November/008532.html
(an interesting discussion on failing over registrations)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-October/thread.html#7734
(clustering via DNS SRV)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-November/007845.html
(hot failover and 6 9s!)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-September/005874.html
(high availability clustering)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-October/007376.html
(outbound call failover)

Of these, the two discussing hot failover are closest to what I was
referring to.

Last night, we upgraded our Acme Packet Session Border Controllers. We
upgraded to a new major version, including rebooting each one in turn. No
registrations (and presumably no calls) were dropped during the process.
They are tied together via ethernet.

An SBC's function is admittedly much more narrowly focused. Has there been
any work in FS in this area? I presume there is a mechanism for dumping a
registration database from a FS box. Is there a way to load that database
into another one?

Thanks.

On 29/09/09 10:34 -0700, Michael Collins wrote:

[mgg at giagnocavo.net]

Those discussions are probably it. It goes something like:

"I want hot failover. Can't FS just load state?"
"No, we think it'd take about $100K of work, and considerable time"
"Oh. XXX does it. I'd think we just need a database."
"No, it'll take a lot of effort. Just use a SIP proxy and redirect traffic around. Run FS on OpenVZ to avoid scheduled hardware maintenance. There's a money back guarantee on FS anyways."

So, until someone gets funding and dev time scheduled, it probably won't go anywhere. Now if someone shows up with the people to actually implement it and wants some pointers on where to start after having gotten into the FS core, then the responses might be slightly different.

-Michael

-----Original Message-----
From: freeswitch-users-bounces@lists.freeswitch.org [mailto:freeswitch-users-bounces@lists.freeswitch.org] On Behalf Of Dan White
Sent: Tuesday, September 29, 2009 1:05 PM
To: freeswitch-users@lists.freeswitch.org
Subject: Re: [Freeswitch-users] Freeswitch Failover

I've downloaded 0Kb in 56 minutes. Anyone mind setting up a seeder? I'm on
a pretty fast connection.

While waiting, I found the following discussions:

I found the following threads while searching for redundancy and failover:
http://lists.freeswitch.org/pipermail/freeswitch-users/2007-September/001499.html
(round robin)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-April/013069.html
(load balancing, scalability)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-August/017349.html
(a discussion of setting up multiple FS boxes in multiple datacenters)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-December/009723.html
(comparing FS to a commercial product)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-February/011770.html
(WAN redundancy)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-February/010924.html
(Comparing against commercial SBCs)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-September/006710.html
(Hardware redundancy)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-January/001934.html
(Generic heartbeat failover question)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-August/017972.html
(Failing over an originating call based on originate_timeout)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-August/005740.html
(HA via Ultra Monkey)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-April/013097.html
(gateway redundancy)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-August/005688.html
(redundancy via openser)
http://lists.freeswitch.org/pipermail/freeswitch-users/2007-August/001363.html
(failover based on xml-curl)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-July/016904.html
(multiple gateways)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-February/002146.html
(dead gateway detection)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-June/003979.html
(load balancing and failover via mod_xml_curl)
http://lists.freeswitch.org/pipermail/freeswitch-users/2007-September/001445.html
(multiple gateways)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-June/015363.html
(hot failover)
http://lists.freeswitch.org/pipermail/freeswitch-users/2009-May/014071.html
(failover extension)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-November/008532.html
(an interesting discussion on failing over registrations)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-October/thread.html#7734
(clustering via DNS SRV)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-November/007845.html
(hot failover and 6 9s!)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-September/005874.html
(high availability clustering)
http://lists.freeswitch.org/pipermail/freeswitch-users/2008-October/007376.html
(outbound call failover)

Of these, the two discussing hot failover are closest to what I was
referring to.

Last night, we upgraded our Acme Packet Session Border Controllers. We
upgraded to a new major version, including rebooting each one in turn. No
registrations (and presumably no calls) were dropped during the process.
They are tied together via ethernet.

An SBC's function is admittedly much more narrowly focused. Has there been
any work in FS in this area? I presume there is a mechanism for dumping a
registration database from a FS box. Is there a way to load that database
into another one?

Thanks.

On 29/09/09 10:34 -0700, Michael Collins wrote:

[nicolas at medularis.com]

Just finished downloading the whole torrent and seeding now.


On Tue, Sep 29, 2009 at 3:04 PM, Dan White <dwhite@olp.net (dwhite@olp.net)> wrote: