how to sync Nxcloud nodes

[Guest]

Hi

I have a Nxcloud cluster with 1 master and 3 slaves.

I realize that I make some config changes, and only the master node takes the configuration.
For example: If I add a new user, only the master node responses the DNS requests... the slaves nodes doesnt responses. the slaves nodes doesnt have the last changes of configuration.

I guess I have many issues relates to the syncronization of the nodes of my nxcloud cluster. right?
so.
Is there a command to force the sync between nodes?
How can I check the sync in any node?

REgards

[support200]

There are 2 conditions for your slave nodes to find out the changes in your master node and reload its config.

1. Time synch

2. DB connection

Are they in the same time zone and you get the same result when you run 'date' command? Maybe we should check this in database in future.

Do you see your slave node active in 'System > Cluster'?

If you are not sure with these, we can check it for you. Contact us at 'support @ nxfilter.org'.

[support200]

Sorry, actually Time Synch doesn't matter here. Do you see this kind of log in /nxfilter/log/nxfilter.log on your slave nodes?

INFO [2022-04-20 01:38:54] - ConfigLoader, Still running..

Since your slave nodes supposed to detect the changes in DB by ConfigLoader, it should be reloaded when you change something while it's running.

[support200]

I think you already contacted us for this problem once. We have many users running clusters of NxFilter and NxCloud. Some of they run 8 slave nodes in their cluster. You are the first one who brought this kind of problem. Very weird.

We made ConfigLoader restarting by itself when it's stopped last time for you. So I guess it's running. Last time, we confirmed your config loading without any problem.

Contact us. We will look into the problem and see if there's any real problem there.

[support200]

I looked into your server report on our side. Your servers worked fine until 2022-06-06 and the next day your slave nodes have a problem with DB connection I guess.

Your slave node access its master node DB through TCP/19003 and it writes a log for DB access. You can see it active or not on 'System > Cluster'.

[support200]

You may have the problem while you see it active on 'System > Cluster'. It means that TCP/19004 for node communication is fine but you have the problem with TCP/19003.

At the moment, you slave node sends an alert email for master connection when both ports closed. We may need to change it sending 2 kinds of alert emails for each case.

[Guest]

Hi. solved after restarting the OS.