[cisco-voip] UCCX HA and CUIC Historical Data Source

Wed Apr 12 23:17:05 EDT 2017

Hi Anthony,

Yup, unfortunately we don’t do a very good job of documenting this. Ideally, it should configure itself and the reason the secondary tab is not populated is that secondary tab is meant for standalone CUIC servers itself where each server needs to have a primary/backup, but in UCCX we don’t utilize that concept.

I have asked the team to create an external facing technotes covering this behavior and how the troubleshooting should be done. I will share the document here once it is done for review.

Regards,
Abhiram Kramadhati
Technical Solutions Manager, CCBU
CCIE Collaboration # 40065

From: Matthew Loraditch <MLoraditch at heliontechnologies.com>
Date: Wednesday, 12 April 2017 at 1:05 AM
To: Anthony Holloway <avholloway+cisco-voip at gmail.com>, "Abhiram Kramadhati (akramadh)" <akramadh at cisco.com>, Nick Britt <nickolasjbritt at gmail.com>
Cc: Cisco VoIP Group <cisco-voip at puck.nether.net>
Subject: RE: [cisco-voip] UCCX HA and CUIC Historical Data Source

11.5, I installed SU1 after install, but didn’t look at CUIC until after.

Matthew G. Loraditch – CCNP-Voice, CCNA-R&S, CCDA
Network Engineer
Direct Voice: 443.541.1518

Facebook<https://www.facebook.com/heliontech?ref=hl> | Twitter<https://twitter.com/HelionTech> | LinkedIn<https://www.linkedin.com/company/helion-technologies?trk=top_nav_home> | G+<https://plus.google.com/+Heliontechnologies/posts>

From: Anthony Holloway [mailto:avholloway+cisco-voip at gmail.com]
Sent: Tuesday, April 11, 2017 3:34 PM
To: Matthew Loraditch <MLoraditch at heliontechnologies.com>; Abhiram Kramadhati (akramadh) <akramadh at cisco.com>; Nick Britt <nickolasjbritt at gmail.com>
Cc: Cisco VoIP Group <cisco-voip at puck.nether.net>
Subject: Re: [cisco-voip] UCCX HA and CUIC Historical Data Source

What version Matt?

On Tue, Apr 11, 2017 at 1:34 PM Matthew Loraditch <MLoraditch at heliontechnologies.com<mailto:MLoraditch at heliontechnologies.com>> wrote:
Just on#1 and 2 because I happened to just be building 2 CCX servers today, it did configure itself.

Matthew G. Loraditch – CCNP-Voice, CCNA-R&S, CCDA
Network Engineer
Direct Voice: 443.541.1518<tel:(443)%20541-1518>
Facebook<https://www.facebook.com/heliontech?ref=hl> | Twitter<https://twitter.com/HelionTech> | LinkedIn<https://www.linkedin.com/company/helion-technologies?trk=top_nav_home> | G+<https://plus.google.com/+Heliontechnologies/posts>

From: cisco-voip [mailto:cisco-voip-bounces at puck.nether.net<mailto:cisco-voip-bounces at puck.nether.net>] On Behalf Of Anthony Holloway
Sent: Tuesday, April 11, 2017 2:29 PM
To: Abhiram Kramadhati (akramadh) <akramadh at cisco.com<mailto:akramadh at cisco.com>>; Nick Britt <nickolasjbritt at gmail.com<mailto:nickolasjbritt at gmail.com>>

Cc: Cisco VoIP Group <cisco-voip at puck.nether.net<mailto:cisco-voip at puck.nether.net>>
Subject: Re: [cisco-voip] UCCX HA and CUIC Historical Data Source

I do appreciate the follow up, but I have a bone to pick here...actually, a few.  First, I was not in a situation where datasources were pointing at each other, I did have them pointing to a single node.  And that's where failover failed me.  I had both nodes pointing at the default Node 1, which was my Master, and when failover happened, reporting was down.

1) Where in the documentation does it say to reconfigure the datasource to point to the non-master node when deploying HA?

2) Why doesn't it configure itself?

3) Why doesn't it update itself like LiveData does?

4) Why can't we use the secondary source tab to populate both nodes?

5) What if I don't want all my reporting traffic traversing the WAN in HAoW deployments?  I.e., My Mast is located where most of my Agent and Supervisors are, and I want reporting to stay local to them.  The HA node is there just in case of a datacenter failure.
On Tue, Apr 11, 2017 at 3:03 AM Abhiram Kramadhati (akramadh) <akramadh at cisco.com<mailto:akramadh at cisco.com>> wrote:
Hi guys,

So, here is the expected behavior and we are taking 11.5 as reference:

Historical datasource:

The historical datasource on either node should always point to the non-master Node. So, assume N1 is the master. CUIC on N1 should be pointing to N2 datasource host and CUIC on N2 should be pointing to N2 datasource host. Also, unlike older versions we don’t redirect users to the secondary UCCX IP to login to run reports. You can login to any node CUIC and run the report, but the datsource host should be the secondary server.
However, we have seen cases where the nodes point to each other: N1 pointing to N2 (expected), N2 pointing to N1 (problem). This is opened as a defect and we are investigating why this can happen and I am guessing that’s what has happened in Anthony’s case in the original email thread.

CSCvd22513 is the defectID for your reference. I will share more information on the outcome of the investigation.

LiveData:

This is always pointing to the local node by default and need not be changed. The LD datasource uses REST API to figure out the engine master and will obtain the feed from the current active server. The LD failover too should happen automatically. So if the engine failover happens and you are still logged into N1 (slave now), LD should still be available and usable.

Regards,
Abhiram Kramadhati
Technical Solutions Manager, CCBU
CCIE Collaboration # 40065

From: "Abhiram Kramadhati (akramadh)" <akramadh at cisco.com<mailto:akramadh at cisco.com>>
Date: Friday, 7 April 2017 at 2:52 PM
To: Anthony Holloway <avholloway+cisco-voip at gmail.com<mailto:avholloway%2Bcisco-voip at gmail.com>>, Nick Britt <nickolasjbritt at gmail.com<mailto:nickolasjbritt at gmail.com>>

Cc: Cisco VoIP Group <cisco-voip at puck.nether.net<mailto:cisco-voip at puck.nether.net>>
Subject: Re: [cisco-voip] UCCX HA and CUIC Historical Data Source

I will get back on this by Monday with details.

Regards,
Abhiram Kramadhati
Technical Solutions Manager, CCBU
CCIE Collaboration # 40065

From: cisco-voip <cisco-voip-bounces at puck.nether.net<mailto:cisco-voip-bounces at puck.nether.net>> on behalf of Anthony Holloway <avholloway+cisco-voip at gmail.com<mailto:avholloway%2Bcisco-voip at gmail.com>>
Date: Friday, 7 April 2017 at 6:01 AM
To: Nick Britt <nickolasjbritt at gmail.com<mailto:nickolasjbritt at gmail.com>>
Cc: Cisco VoIP Group <cisco-voip at puck.nether.net<mailto:cisco-voip at puck.nether.net>>
Subject: Re: [cisco-voip] UCCX HA and CUIC Historical Data Source

Nope. Abhiram?

On Thu, Apr 6, 2017 at 2:31 PM Nick Britt <nickolasjbritt at gmail.com<mailto:nickolasjbritt at gmail.com>> wrote:
Hi Anthony,

Sorry to grave dig but just wondered if you ever got an answer for this?

I am about to go down this rabbit hole myself as a customer wants a better explanation (documentation) as to how this is supposed  to be configured and how this should behave.

Cheers

Nick

On Wed, Jun 15, 2016 at 10:22 PM, Anthony Holloway <avholloway+cisco-voip at gmail.com<mailto:avholloway+cisco-voip at gmail.com>> wrote:
There seems to be zero documentation for UCCX that mentions changing or adding Data Source configuration in CUIC when running HA; whether HAoL or HAoW.

However, I have heard from Cisco employees, forums posts, and colleagues, these two points:
1.      The Data Source should point to the secondary node, and you have to manually change it, as the default is pointing to the primary node.
2.      The Data Source's Secondary tab is defaulted to disabled, and not populated.  It shouldn't be used, and CUIC takes care of updating the Data Source during a failure.
First off, where in the documentation is that explained?  I cannot find a good explanation, sans ambiguity, to save my life.  Are people just spreading rumors and old wives tales?

Also, I do know that back in the HRC days, the client would handle the connection to the secondary server automatically.  So, I can see where this tale comes from.

Now, with HAoW, I tested failover with the server shutdown.  Not in slave, but actually powered off.  What I observed was, the Data Source was not automatically updated, and I could run any reports, despite being logged in to the secondary CUIC server.  The Data Source connection test failed, obviously, and reports failed, obviously.

I did consider take a leap of faith and confiure the CUIC Data Source's Secondary tab, but the user account to connect to the DB instance was not in my control, and I don't know the password.  I'm sure I could get it, but it was a show stopper nonetheless.

So, has anyone here actually tested with a failed node or island mode with their HA setup, or is it all just speculation, like this post:

https://supportforums.cisco.com/discussion/12473981/ha-uccx-cuic-data-only-one-server

_______________________________________________
cisco-voip mailing list
cisco-voip at puck.nether.net<mailto:cisco-voip at puck.nether.net>
https://puck.nether.net/mailman/listinfo/cisco-voip

--
- Nick
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://puck.nether.net/pipermail/cisco-voip/attachments/20170413/8860f6f4/attachment.html>