cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
58
Views
1
Helpful
6
Replies
Highlighted
Cisco Employee

Template Application Issue: NCS-HA using BGP Quagga

 

Hello NSO Team

 

 

We are facing a weird issue, where the template-to-be-applied is not in sync with the HA status at NCS. We have two devices in the same AS, but not connected directly to each other. We call them NSO-1 (Master) and NSO-2 (Slave). Diagram attached for reference.

 

 

When we shutdown the NCS process at master, the slave node will become the new ‘Master’

 

 

At new ‘Master’ (Previously slave)

 

admin@ncs% request ha commands status

 

status LAB-NFV-NSO-2[master]

 

[ok][2016-08-22 19:48:40]

 

 

However, when we telnet to the quagga device to see the applied template, we see this-

 

[root@LAB-NFV-NSO-2 state]# telnet 10.255.22.6 2605

 

Trying 10.255.22.6...

 

Connected to 10.255.22.6.

 

Escape character is '^]'.

 

 

Hello, this is Quagga (version 0.99.22.4).

 

Copyright 1996-2005 Kunihiro Ishiguro, et al.

 

 

 

User Access Verification

 

 

Password:

 

SLAVE>

 

SLAVE> en

 

SLAVE#

 

 

Instead of seeing the template ‘failover-master’ applied to the new ‘master’ (previously slave).

 

 

When we check the ncs-java-vm.log to see the process beneath, we see this-

 

 

<DEBUG> 22-Aug-2016::19:48:36.791 BgpAnycast Timer-0: - Querying BGP Prefix result:

 

BGP routing table entry for 10.255.22.8/32

 

Paths: (1 available, best #1, table Default-IP-Routing-Table)

 

Advertised to non peer-group peers:

 

10.255.22.5

 

Local

 

0.0.0.0 from 0.0.0.0 (10.255.11.21)

 

Origin IGP, metric 0, localpref 100, weight 32768, valid, sourced, local, best

 

Last update: Fri Aug 19 00:21:10 2016

 

 

SLAVE#

 

<DEBUG> 22-Aug-2016::19:48:36.791 Cluster Timer-0: - UpdateLocalNode - beSlave

 

<ERROR> 22-Aug-2016::19:48:36.792 Ha Timer-0: - failed to call HA

 

Exception

 

<DEBUG> 22-Aug-2016::19:48:40.685 Cluster Timer-0: - UpdateLocalNode – beMaster

 

 

The BGP is now showing only one path for the prefix (as expected). But the following two messages are suspicious. The last message is to declare itself as a master, which is failed to reflect at the quagga device in applying the template.

 

 

Also, if the HA mode is deactivated in the master, will the slave recognize itself as ‘master’ or ‘none’? In our testing, the ‘slave’ node always goes to ‘none’ if the HA is deactivated at the acting ‘master’.

 

 

Above all, it would be very helpful if you can let us know how to find out the reason behind ‘configuration data base locked’ error, and to unlock it.

 

 

Please let me know if you need any more information from our setup to help us solve this situation.

 

 

Other outputs that may help

 

 

admin@ncs% show ha

 

token cisco123;

 

local-user admin;

 

bgp {

 

anycast-prefix   10.255.22.8/32;

 

anycast-path-min 1;

 

}

 

member LAB-NFV-NSO-1 {

 

address         10.255.22.2;

 

default-ha-role master;

 

cluster-manager true;

 

quagga-device   srdm-quagga;

 

}

 

member LAB-NFV-NSO-2 {

 

address         10.255.22.6;

 

default-ha-role slave;

 

failover-master true;

 

relay-name      LAB-NFV-NSO-1;

 

quagga-device   srdm-quagga2;

 

}

 

[ok][2016-08-22 20:04:42]

 

 

Thank you very much.

 

6 REPLIES 6
Highlighted
Cisco Employee

Re: Template Application Issue: NCS-HA using BGP Quagga

 

Hi Krishna,

 

 

Not sure what version are you using. We found similar behaviour with VMS 1.x (NSO 3.x) and filled couple of defects (like CSCuu90400), but AFAIK nothing got resolved….

 

 

BR,

 

Bostjan

 

Highlighted
Cisco Employee

Re: Template Application Issue: NCS-HA using BGP Quagga

 

Hi Bostjan,

 

 

We are on NCS 4.1 and tailf-hcc package version 4.1 as well.

 

 

Thanks,

 

Vineet

 

Highlighted
Cisco Employee

Re: Template Application Issue: NCS-HA using BGP Quagga

 

Hello Bastjan

 

Can you give us the list of CDETS ticket, to check on the what are the defect you experienced so far ?

 

Regards

 

 

Highlighted
Cisco Employee

Re: Template Application Issue: NCS-HA using BGP Quagga

 

Hi Krishna and Linshu,

 

 

Just out of curiosity, could you share template setting (show run devices template)?

 

 

Best regards,

Hiro

 

Highlighted
Cisco Employee

Re: Template Application Issue: NCS-HA using BGP Quagga

 

Hello Hiro –san

 

Here it is

 

admin@ncs# show running-config devices template

devices template hcc-failover-master

config

  quagga-bgp:hostname FAILOVER-MASTER

  quagga-bgp:route-map SET-AS-PATH 10

   set as-path prepend as-container as-number 65001

  !

!

!

devices template hcc-master

config

  quagga-bgp:hostname MASTER

  quagga-bgp:route-map SET-AS-PATH 10

   set as-path prepend as-container as-number 65001,65002

  !

!

!

devices template hcc-none

config

  quagga-bgp:hostname NONE

  quagga-bgp:route-map SET-AS-PATH 10

   set as-path prepend as-container as-number 65001,65002,65003,65004

  !

!

!

devices template hcc-slave

config

  quagga-bgp:hostname SLAVE

  quagga-bgp:route-map SET-AS-PATH 10

   set as-path prepend as-container as-number 65001,65002,65003

  !

!

!

Highlighted
Cisco Employee

Re: Template Application Issue: NCS-HA using BGP Quagga

 

Hello Linshu,

 

 

Thank you for sharing config.

 

The template looks good to me.

 

 

Maybe that’s a bug then.

 

I will let other experts comment.

 

 

Best regards,

 

Hiro

 

Content for Community-Ad
Cisco Community April 2020 Spotlight Award Winners