[cisco-voip] MRA getting served up publisher as TFTP server after upgrade from v9 to v11

Tue Jul 17 22:28:25 EDT 2018

Hmmmm, I was pretty sure the _cisco-uds._tcp.acme.com records were simply hosts you want your jabber clients to begin talking with, not necessarily the TFTP server itself. They are currently set to our two less busiest subscribers.

https://www.cisco.com/c/en/us/td/docs/voice_ip_comm/jabber/11_0/CJAB_BK_C04C09E7_00_cisco-jabber-110-planning-guide/CJAB_BK_C04C09E7_00_cisco-jabber-110-planning-guide_chapter_0101.html

_cisco-uds

  *   Provides the location of Cisco Unified Communications Manager version 9.0 and later.
  *   The client can retrieve service profiles from Cisco Unified Communications Manager to determine the authenticator.

But I think you’re spot on, it’s less about MRA than it is about UDS in general.

The UDS query returns two hosts for TFTP, the publisher and the primary tftp server.

Very weird.

It’s likely that this has been going wrong all along but we just noticed the speed issue after the upgrade.

---
Lelio Fulgenzi, B.A. | Senior Analyst
Computing and Communications Services | University of Guelph
Room 037 Animal Science & Nutrition Bldg | 50 Stone Rd E | Guelph, ON | N1G 2W1
519-824-4120 Ext. 56354 | lelio at uoguelph.ca<mailto:lelio at uoguelph.ca>

www.uoguelph.ca/ccs<http://www.uoguelph.ca/ccs> | @UofGCCS on Instagram, Twitter and Facebook

[University of Guelph Cornerstone with Improve Life tagline]

From: Ryan Huff <ryanhuff at outlook.com>
Sent: Tuesday, July 17, 2018 10:07 PM
To: Lelio Fulgenzi <lelio at uoguelph.ca>
Cc: voyp list, cisco-voip <cisco-voip at puck.nether.net>
Subject: Re: [cisco-voip] MRA getting served up publisher as TFTP server after upgrade from v9 to v11

I think this has less to do with MRA as much as it does with UDS in general.

Do you have a _cisco-uds SRV record that points at the publisher (where tftp is not running)? I believe you only want UDS SRV records pointing at nodes running TFTP (apologies for the screen shot as I am mobile).

[image1.png]

https://www.cisco.com/c/en/us/td/docs/voice_ip_comm/jabber/11_0/CJAB_BK_D657A25F_00_deployment-installation-guide-jabber-110/CJAB_BK_D657A25F_00_deployment-installation-guide-jabber-110_chapter_01010.html#CJAB_TK_S5764985_00

Thanks,
Ryan

Sent from my iPhone

On Jul 17, 2018, at 21:46, Lelio Fulgenzi <lelio at uoguelph.ca<mailto:lelio at uoguelph.ca>> wrote:

It was an off-line upgrade. We basically built our entire environment off-line (install v9 then restore from backup) then upgrade to v11. then turn off v9 online servers and turn on new v11 servers while connected to online network. It’s a tried and true process. ;) works great except for the change freeze and historical reports with uccx.

No changes were made to uc services as far as I can tell.

Oh - and TFTP service is deactivated on publisher.

I could try turning on the old pub in the offline network to see what uds returns. It might be that it was returning both but we’re only seeing the slowdown now?

Definitely weird.
-sent from mobile device-

Lelio Fulgenzi, B.A. | Senior Analyst
Computing and Communications Services | University of Guelph
Room 037 Animal Science & Nutrition Bldg | 50 Stone Rd E | Guelph, ON | N1G 2W1<x-apple-data-detectors://1/0>
519-824-4120 Ext. 56354<tel:519-824-4120;56354> | lelio at uoguelph.ca<mailto:lelio at uoguelph.ca>

www.uoguelph.ca/ccs<http://www.uoguelph.ca/ccs> | @UofGCCS on Instagram, Twitter and Facebook

[University of Guelph Cornerstone with Improve Life tagline]

On Jul 17, 2018, at 9:30 PM, Ryan Huff <ryanhuff at outlook.com<mailto:ryanhuff at outlook.com>> wrote:
Was it an “in place” upgrade? If not check if anything changed with the CM Groups during the upgrade. Sometimes people just upgrade the publisher, and build in new subscribers (if this happened, I’d also make sure any custom MoH files are on all nodes and not just the publisher). When you do that, it will remove everything from the CM  groups and the nodes have to be manually added back and in the correct order.

Did anything change with the UC Services?

Thanks,

Ryan
Sent from my iPhone

On Jul 17, 2018, at 20:47, Lelio Fulgenzi <lelio at uoguelph.ca<mailto:lelio at uoguelph.ca>> wrote:

shot in the dark here - we're working with a partner who has upgraded our cluster from v9 to v11.5.

we noticed that MRA (first time) logins are taking significantly longer. some captures show that MRA clients are being served up the ip address of the publisher for tftp services as well as the primary (only) tftp server.

we use _cisco-uds SRV records for on-prem so the deployment document confirms no settings of tftp server as i recollected. the MRA deployment guide talks about refreshing the nodes - but that's been done a few times.

we'll be opening a case in the morrow but just wondering if anyone has run into this?

---
Lelio Fulgenzi, B.A.
Senior Analyst, Network Infrastructure
Computing and Communications Services (CCS)
University of Guelph

519-824-4120 Ext 56354
lelio at uoguelph.ca<mailto:lelio at uoguelph.ca>
www.uoguelph.ca/ccs<http://www.uoguelph.ca/ccs>
Room 037, Animal Science and Nutrition Building
Guelph, Ontario, N1G 2W1
_______________________________________________
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
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://puck.nether.net/pipermail/cisco-voip/attachments/20180718/4acc4ee1/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 1297 bytes
Desc: image001.png
URL: <https://puck.nether.net/pipermail/cisco-voip/attachments/20180718/4acc4ee1/attachment.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image002.png
Type: image/png
Size: 248578 bytes
Desc: image002.png
URL: <https://puck.nether.net/pipermail/cisco-voip/attachments/20180718/4acc4ee1/attachment-0001.png>