High Availability Backup & Restore Module

I suspected that is odd… I was guessing they registered as part of the backup job. Since they are registered on the main server I assumed the warm spare was keeping the connection up. The phones do not call out the warm spare. What can I do to test or confirm this.

Thanks tonyclewis. Very good to know and will be helpful! However, this still doesn’t fix the problem entirely. I want to be able to bring the backup online with a phone call - so that I can do it from anywhere. I had set it up with the backup loaded ready to go, I can dial a number, enter a code, and a “reload” will be issued to bring it online. Unfortunately, there was a power glitch that rebooted it an brought the trunks online. ACK! Calls were going to two PBXs.

If I restored it with the trunks offline (using this new feature), I would have to either do another restore with the trunks active, or manually enable trunks. I don’t know how to a restore or enable trunks via asterisk dialplan commands.

The only solution I can come up with is to write a php script to load a backup, and invoke that from a dialplan.

I can’t find the “disable trunks on restore” option. I have Backup & Restore 2.10.0.41. Can someone describe where it the setting is please?

Hello everybody,

I have just one question, which is the same as robfantini from 7/10/2010 at the article here:

http://www.freepbx.org/news/2010-05-30/high-availability-backup-and-restore

How can I tell the backup module to exclude the NAT IP address and the IP address info in the OSS endpoint manager?

I use hosted PBX systems we create in a datacenter. I have 45+ (and growing fast) customers who all can’t be down for more than 60 seconds. We use routers that support SRV records and local registrations to the said routers. Basically, here is how our failover works.

Router (at customer premise) - Edgewater Networks Edgemarc 4500 or 4550 on 11.6.x firmware

  • the SIP server under SIP ALG (on the router) is set to go to my-pbx.mycompany.com
  • the SIP phones are configured to register to the Edgemarc on VLAN 100 at its local IP on said VLAN at 192.168.96.254.
  • we have two SRV records created in our public (or private) DNS configure like below:

_sip._udp.my-pbx.mycompany.com 0 0 5060 my-pbx.mycompany.com
_sip._udp.my-pbx.mycompany.com 1 0 5060 my2-pbx.mycompany.com

note that the second line shows 1 0 5060 - this means the second line is lower priority, and the second pbx is my2-pbx.mycompany.com

  • the survivability section of the Edgemarc automatically looks up this SRV record, and lists the two pbxes, but with my-pbx.mycompany.com as the primary. Survivability in the Edgemarc is set to auto-refresh every XX seconds - which we have set for 60 seconds. When the first PBX is down (or the datacenter is down), then all phones register to the second PBX listed as active. As soon as the primary comes back online, the phones switch back to the primary.

The big caveat here, however, is we need both PBXs to remain configured with Asterisk’s IP configuration for that specific PBX IP we configured. Both are in different datacenters and plugged directly into the internet. Obviously, we cannot just change an IP address and expect it to work.

Thank you for any and all input you may have,

Jon

I found that the table sipsettings in mysql is where the bind and public IP settings are, and also that under endpointman_global_vars, the public IP of the sip server is present there.

It would be nice to only tell the backup system to exclude the rows containing the ip addresses as opposed to the entire table, but we can manually configure these two pages. It isn’t a big deal. Here is what I added to my exclude field in the backup settings for the ‘Config Server’ backup item:

exclude:

backup
backup_cache
backup_details
backup_items
backup_server_details
backup_servers
backup_template_details
backup_templates
sipsettings
endpointman_global_vars

My apologies for not having responded to the OP. I had experienced some similar issues… but now that we are using FreePBX 2.11, those issues seem to have completely disappeared.

Regarding your GUI-related issues, are you using Google Chrome? This is the only browser I would recommend to use.