Description

 

This article describes the steps to recover the Virtual Machines (VMs) in Contrail Service Orchestration (CSO) in case of any failure.

 

Symptoms

 

VRR1 VM is not coming up after a server reboot. I am not able to telnet to my VRR.

root@host:/var/log/libvirt/qemu# virsh console vrr
error: failed to get domain 'vrr'
error: Domain not found: no domain with matching name 'vrr'

 

Solution

 

In CSO all-in-one deployment, there are six VMs that are provisioned during deployment as shown below: 

From the CSO installation server, perform virsh list .

root@host:~# virsh list 
 Id    Name                           State
----------------------------------------------------
 1     k8-microservices1              running
 2     contrail_analytics1            running
 3     k8-infra1                      running
 4     monitoring1                    running
 5     startupserver1                 running
 7     vrr1                           running

 

 

Perform the following:

  1. Log in to the startup server VM from the installation server. 
  2. Go to the Contrail_Service_Orchestration_5.1.2 folder and run the recovery.sh script.

root@startupserver1:~/Contrail_Service_Orchestration_5.1.2# ./recovery.sh 

Note: recovery.sh is a utility provided by CSO BU and is added automatically during installation.  

root@startupserver1:~/Contrail_Service_Orchestration_5.1.2# ./recovery.sh 

This tool assists you with recovering your CSO setup.

The following components can be recovered:

  • cassandra

  • mariadb

  • vrr

  • saltstack

  • icinga

  • rabbitmq

  1. Specify one of the components to recover (in number) : 3 (in our example)

INFO     Started recovering vrr component at 2021-01-22 01:08:41.412041 ...
INFO     VRR recovery is initiated...
ERROR    Vrr - 192.168.10.29 is unhealthy
INFO     Recovery takes time, please be patient
INFO     VRR recovery started. Please wait...
INFO     Vrr console recovered for vrr1
INFO     VRR config sync 
INFO     Completed recovering vrr component at 2021-01-22 01:21:27.054657 .
INFO     Time taken to recover 0:12:45.642616 

From the above logs, it is confirmed that the VM has been recovered. 

root@host:~# virsh list
 Id    Name                           State
----------------------------------------------------
 1     k8-microservices1              running
 2     contrail_analytics1            running
 3     k8-infra1                      running
 4     monitoring1                    running
 5     startupserver1                 running
 7     vrr1                           running

You will now be able to console to VRR1 successfully.

root@host:~# virsh console 7
Connected to domain vrr1
Escape character is ^]

root>
root> show version 
set version 15.1R6.7