Sipxcom High Availability Configuration Primer on Wiki

[pmkr...@gmail.com]

I've put together the first pass of a wiki document found here http://wiki.sipxcom.org/display/sipXcom/Sipxcom+High+Availability+Primer on configuring Sipxcom for high availability. There are two phases to the work, and I'm about 3/4 through phase 1, so this document will be updated as more results become available.

There are 2-3 people on this forum who have been incredibly supportive and influential in my Sipxcom work - I won't names (you know who you are) in order to maintain decorum and a sense of humility. I'll only offer five words - "a very sincere thank you".

Feedback and insights appreciated - hope this helps others.

All the best
Peter

[Michael Picher]

Hi Peter,

Shaping up nicely. When complete this should probably live here: http://wiki.sipxcom.org/display/sipXcom/High+Availability

I think with respect to DNS that all of the DNS duties can be performed properly with DNS Manager and having servers in proper bind views. We would normally put DNS on all of the servers which also helps keep the server speed up.

Unfortunately DNS Manager is one of those things that needs some better (or just some) documentation as well...

Thanks,

  Mike

Michael Picher, VP of Product Innovation
eZuce, Inc.

300 Brickstone Square

Suite 104

Andover, MA. 01810

O.978-296-1005 X2015 
M.207-956-0262
@mpicher <http://twitter.com/mpicher> 

linkedin
www.ezuce.com

Notice: This transmittal and/or attachments may be privileged or confidential. It is intended solely for the addressee(s) named above. Any dissemination, or copying is strictly prohibited. If you received this transmittal in error, please notify us immediately by reply and immediately delete this message and all its attachments. Thank you. FMS

--
You received this message because you are subscribed to the Google Groups "sipxcom-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sipxcom-users+unsubscribe@googlegroups.com.
To post to this group, send email to sipxco...@googlegroups.com.
Visit this group at https://groups.google.com/group/sipxcom-users.
To view this discussion on the web visit https://groups.google.com/d/msgid/sipxcom-users/51dfab54-1b6e-4849-beda-9a776b6bfdf9%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Michael Picher]

I've put some good base info for DNS Management into the wiki...

http://wiki.sipxcom.org/display/sipXcom/DNS+Management

Thanks,

  Mike

Michael Picher, VP of Product Innovation
eZuce, Inc.

300 Brickstone Square

Suite 104

Andover, MA. 01810

O.978-296-1005 X2015 
M.207-956-0262
@mpicher <http://twitter.com/mpicher> 

linkedin
www.ezuce.com

Notice: This transmittal and/or attachments may be privileged or confidential. It is intended solely for the addressee(s) named above. Any dissemination, or copying is strictly prohibited. If you received this transmittal in error, please notify us immediately by reply and immediately delete this message and all its attachments. Thank you. FMS

[pmkr...@gmail.com]

Thanks Mike for the details here - I worked from this description implyingr that unmanaged DNS applies to everything and not just the phones - Company or ITSP DNS servers to resolve ALL names instead of local DNS servers. This statement may need to be refreshed to state that unmanaged DMS applies only to phones.

In working with the default DNS record view and defining failover plans, I was unable to get failover plans applied to the default.name server file generated by Sipxconfig in the /var/named directory. I'm doing more work this evening, and will double-check this. The only way I could get failover plans to override the default plan is by defining regions.

There is some upside to the DNS struggles - it made the Sipxcom DNS principals documents come to life as one really needed to think through which DNS records are used by which Sipx services. If I had to train someone new on HA, my first assignment to them would be to build the system using all external DNS servers first. Once successful, then use the onboard tools to understand what they do and how they apply to the HA cluster.

Thanks again.

All the best
Peter

To unsubscribe from this group and stop receiving emails from it, send an email to sipxcom-user...@googlegroups.com.

[Michael Picher]

It's definitely good for folks to get a better understanding of DNS and views.

With respect to the default view vs. others, I would expect in a larger installation that servers would be in default - like views and phones then are in views where you control the failover as you'd like.

Mike

To unsubscribe from this group and stop receiving emails from it, send an email to sipxcom-users+unsubscribe@googlegroups.com.

To post to this group, send email to sipxco...@googlegroups.com.
Visit this group at https://groups.google.com/group/sipxcom-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/sipxcom-users/5a23c5ff-6594-40e8-bf2b-7875396eaac7%40googlegroups.com.

[pmkr...@gmail.com]

I can create DNS failover plans but cannot seem to find a way to change the weighting from the default 100 percent - I can't type a value in the field or adjust downwards using the arrow icon. The behavior from Firefox, Chrome, and Edge browsers is identical.

Have tried this from 16.08 and 16.12 systems. Is there a browser setting I need to change or something else I am missing?

Many thanks in advance.

All the best
Peter

[pmkr...@gmail.com]

Thought I'd share this very preliminary result from phase 2 testing in the wiki. There are three servers - one primary and secondary on one LAN segment, and another secondary server in a separate LAN segment. All traffic is going through the router.  The traffic graph is taken from the perspective of LAN interface to the second subnetwork. In this test, there are only two phones registered to the secondary server in the second subnet.

The primary voice server was down when I started this test and brought it back up. After the primary server comes back online, there is ~8M burst of data as state information is exchanged from the secondary server in the second subnet back to the primary. The second peak represents a push of server profiles from the primary to the secondary server. These are 100MB+ connections between the subnets.

All the best
Peter

[Michael Picher]

If you add another server to the same section, i think you can hover over the 100 or on the field below and you'll see some arrows appear.

That is built with Dart and it acts a little wonky.

Mike

To view this discussion on the web visit https://groups.google.com/d/msgid/sipxcom-users/3c37be2a-2b07-4fdd-ac10-0d80e759f6c1%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

Thanks,

   Mike

-----------------------------------------------------------------------------------------

There are 10 types of people in this world...  Those who understand binary and those who don't.

[pmkr...@gmail.com]

I am trying to build a system that looks as follows:
Site 1 - pbx primary, pbx2 secondary, no region assigned - global databases - subnet1
Site 2 - pbx3 secondary, local region, regional database - subnet 2

All databases appear to have good connectivity, but for pbx3 the sipregistrar does not come up - here is what appears in the sipregistrar log file,

"2017-03-24T14:15:06.662075Z":7:SIP:NOTICE:pbx3.lvtest.com::7f1a24251700:sipxregistry:"SipUserAgent::handleCancelQueue - STARTED"
"2017-03-24T14:15:06.667745Z":8:SIP:CRIT:pbx3.lvtest.com:SipRegistrar:7f1a2705b700:sipxregistry:"[150-ISN] SipRedirectorISN::readConfig BASE_DOMAIN parameter missing or empty"
"2017-03-24T14:15:06.668364Z":9:SIP:CRIT:pbx3.lvtest.com:SipRegistrar:7f1a2705b700:sipxregistry:"[160-ENUM] SipRedirectorENUM::readConfig BASE_DOMAIN parameter missing or empty"
"2017-03-24T14:15:06.675052Z":10:SIP:ERR:pbx3.lvtest.com:SipRegistrar:7f1a2705b700:sipxregistry:"SipRegistrar::operationalPhase Exception: Failed to call findOne, no good nodes in sipxecs"
"2017-03-24T14:15:06.675151Z":11:SIP:EMERG:pbx3.lvtest.com:SipRegistrar:7f1a2705b700:sipxregistry:"SipRegistrar::run Unable to initialize server"
"2017-03-24T14:15:07.650381Z":12:SIP:NOTICE:pbx3.lvtest.com::7f1a31ade800:sipxregistry:"main: cleaning up."
"2017-03-24T14:15:07.650677Z":13:KERNEL:NOTICE:pbx3.lvtest.com::7f1a31ade800:sipxregistry:"Exiting sipxregistry"

Any thoughts on how to correct this - I've tried a variety of permutations including fresh installations yesterday.

Many thanks in advance.

All the best
Peter

[pmkr...@gmail.com]

Some more observations - I just went through the DNS management wiki info posted earlier this week and believe I have a better understanding of what is happening with the Sipxcom side regarding DNS management. First, the arrows on the failover plans only appear to work when you add two additional plans - otherwise adjustments do not work. I adjusted values and then deleted unused fail-over policies. In the Ezuce screen snapshots on overriding the default DNS plan, when a custom plan is created and then you back and open the default DNS record view, the menu presents a slider window of custom DNS record plans that wish to be included. This menu never appears in the 16.12 Sipxcom release.  Finally, I clicked on the unmanaged DNS service option and provisioned the IP address of the standalone DNS server. I ran a tcpdump DNS trace on the unmanaged DNS server expecting to see queries from Sipxcom when lines were registering - there were none. I also did an audit of the /etc/resolv.conf and /var/named/default.. files and saw no references to the IP address of the unmanaged DNS server.

It almost appears as if there are components missing between the Ezuce and Sipxcom streams ...

All the best
Peter

[pmkr...@gmail.com]

The High Availability primer has been moved http://wiki.sipxcom.org/display/sipXcom/Sipxcom+High+Availability+Primer and updated with initial phase 2 results. Unfortunately, I am still struggling with the DNS management tools on Sipxcom so the primer was written assuming that Sipxcom uses standalone DNS servers for the moment. The document will be updated when a better handle on the Sipxcom DNS management tools is obtained. I have figured out how to adjust the weights on the failover plans (you need to add an entry and then can adjust weights). I have also just figured out how to bring up the attached menu which may be the recipe for over-writing the default records. Will work with this today.

Any insights on where to look in Mongo, commands, etc for display replication dataset from primary to secondary is appreciated.

All the best
Peter

[pmkr...@gmail.com]

I have updated the high availability primer again and captured how to use some of the managed DNS tools on Sipxcom.

All the best
Peter

[Joe Micciche]

On 03/28/2017 09:09 AM, pmkr...@gmail.com wrote:
> Any insights on where to look in Mongo, commands, etc for display
> replication dataset from primary to secondary is appreciated.
>
> All the best
> Peter

Go into mongo shell:

mongo
then
rs.status()
rs.config()
then
exit

joe