Salt Master_Active Active

Study Active-Active Configuration For Salt-Master

Source: EOS-9550

Procedure

1. Intall salt-master on all nodes

   yum install salt-master
   

2. Copy the keys from primary node located at /etc/salt/pki to non-primary nodes

   [root@srvnode-2 ~]# scp -r /etc/salt/pki srvnode-2:/etc/salt/pki
   master.pub                                        100%  451     1.3MB/s   00:00
   master.pem                                        100% 1675     5.8MB/s   00:00
   

3. Restart salt-master on all nodes

   systemctl restart salt-master
   

4. Install salt-minion on both nodes

   yum install salt-minion
   

5. Configure master and related values in configuration file /etc/salt/minion

   master:

   • srvnode-1
   • srvnode-2

   master_type: str
   random_master: True
   master_shuffle: False

6. Restart salt-minion

   systemctl restart salt-master
   

7. Checked for key acceptance on both nodes Node-1:

   [root@srvnode-2 ~]# salt-key -L
   Accepted Keys:
   Denied Keys:
   Unaccepted Keys:
   srvnode-1
   srvnode-2
   Rejected Keys:
   
   [root@srvnode-2 ~]# salt-key -A -y
   The following keys are going to be accepted:
   Unaccepted Keys:
   srvnode-1
   srvnode-2
   Key for minion srvnode-1 accepted.
   Key for minion srvnode-2 accepted.
   
   [root@srvnode-2 ~]# salt "*" test.ping
   srvnode-1:
       True
   srvnode-2:
       True
   
   [root@srvnode-2 demo]# salt "*" grains.get master
   srvnode-1:
       srvnode-2
   srvnode-2:
       srvnode-2
   

   Node-2:

   [root@srvnode-1 ~]# salt-key -L
   Accepted Keys:
   Denied Keys:
   Unaccepted Keys:
   srvnode-1
   srvnode-2
   Rejected Keys:
   [root@srvnode-1 ~]# salt-key -A -y
   The following keys are going to be accepted:
   Unaccepted Keys:
   srvnode-1
   srvnode-2
   Key for minion srvnode-1 accepted.
   Key for minion srvnode-2 accepted.
   [root@srvnode-1 ~]# salt-key -L
   Accepted Keys:
   srvnode-1
   srvnode-2
   Denied Keys:
   Unaccepted Keys:
   Rejected Keys:
   [root@srvnode-1 ~]# salt "*" test.ping
   srvnode-1:
       True
   srvnode-2:
       True
   [root@srvnode-1 ~]# salt "*" grains.get master
   srvnode-1:
       srvnode-1
   srvnode-2:
       srvnode-1
   

Observations

1. On node-1, at times auto-detection of the keys doesn't list all keys after waiting for couple of mins and restarting services. Firing salt-call state.highstate command helps detection of the minions on node-1. ( needs Exploration.) This didn't happen on secondary node though. The keys were auto-detected by its master.

2. Created a basic pillar file and a state file and verified it is accessible across both nodes. It was observed that, in a failover scenario the pillar files and srv(both srv & srv_user) are required in sync across all nodes.