migrate backend from zabbix standalone to zabbix distribuited (master and proxy)

[yayo]

We use and love zabbix... Monitored devices by zabbix are grown and now we need to switch from standalone (zabbix01) to distribuited architecture (1 master e 2 proxy for now).

We have just created the new architecture (called zabbixnew) and we need to switch off the zabbix01 server, and import data from old backend to the new one.

There is any problem to do this?

My second question:

After migration and switch off are completed (and all works fine .. i hope! ) if i switch the agent configurations to allow new server to pull the data, zabbixnew will resume to pull data from the last time before migration? I know that we will have a hole between the stop of the zabbix01 and the start of the zabbixnew (and for us, is ok)

Is that the right way to move to new architecture?

Thank you!

[Pada]

There is a big difference between having a distributed setup (using nodes) and having a standalone server with proxies.

You can start using zabbix proxy servers with a standalone zabbix server without having to create a new server or import settings.
All you'll need to do is:

• add the proxy via the frontend
• change the zabbix server address on the active zabbix agents to point to the proxy's address instead and then restart those agents
• set the proxy (via zabbix frontend) for the hosts that you want to monitor via the proxy

You typically go for proxies when you have hosts sitting in a different network/behind a firewall. Like we use proxies in our environment, where we have a single proxy at every data centre, so that it will continue to collect data even when the network link is down to our zabbix server. The other advantage is that it takes away some of the load off our zabbix server.

What is the reason you want to go for a proxy or distributed architecture?

[yayo]

no nodes in our new configuration so this scenario is not considered...

we have created new environment (1 master 2 proxies) with better hardware configuration for more device to manage.

We need some kind of steps to migrate (or copy or switch off) from old standalone to new master ...

Have this any problem? And, when new master startup and pull data from devices no data will be lost? I think that data lost will be between stop of old standalone server and start new master.