running two networkmanagementsystems at same time?!

[makemeasandwich]

Hello,
i am trying out another open source networkmanagementsystem besides zabbix on my ubuntu just for fun and I want to run both at the same time.

My zabbix uses the postgresql.

The other network managementsystem is also using postgresql.

However both system save monitored data in the database and both seem to work just fine?

How does that work? The data and the database dont get overritten?

Both processes can be seen that they are running.

Thanks for the help!

[Linwood]

Most likely both use separate databases.

This can work fine, though if the two applications do more global maintenance (e.g. stopping the database server at some point) then you can get a collision.

You also will not be able to do functions that collide on ports, e.g. if you use SNMP traps, both will try to listen on the same port.

You should consider separating these, on either VM's or separate physical servers. While probably 95% of everything should work, the smaller overlaps that may collide will be confusing.

[makemeasandwich]

Most likely both use separate databases.

This can work fine, though if the two applications do more global maintenance (e.g. stopping the database server at some point) then you can get a collision.

You also will not be able to do functions that collide on ports, e.g. if you use SNMP traps, both will try to listen on the same port.

You should consider separating these, on either VM's or separate physical servers. While probably 95% of everything should work, the smaller overlaps that may collide will be confusing.

Thanks for the quick answer!

Both use one database for example postgrse but separate databases ? How does that work tho?

I have only one version of it installed when I check in cluster. However I have 2 users running one is called zabbix and it looks like he is using zabbix related stuff while the other one is called postgres. The other user postgres is related to the other system

[Linwood]

I am not sure I understand your question completely but let me try.

MySQL and Postgres are both database engines, not databases. In general database engines can have many databases present at one time, both databases related to each other (for one application) and databases completely unrelated to each other.

This is all mostly transparent to the client software, each knows only its own database. Think database engine = disk drive, and database = file (that is not a very good analogy in many ways but maybe it helps).

Applications which are "well behaved" with respect to databases tend to ignore each other's databases and access and everything works fine. The problem comes in usually when an application, frankly, tries to be helpful -- where an application takes over some aspect of the database engine (not just database), for example starts and stops it on some schedule for backup (which is not necessary but some applications still do it). There are also some global settings, which vary by database engine, which either cannot be set individually by database, or may not be set individually by the application. When this occurs you can have issues running two un-coordinated applications against the same database engine (with different databases).

The problem you have with two unrelated applications is there is generally no way to know in advance if they will be badly or well behaved in this regard. The same is true of other global aspects unrelated to the database - temp area usage and file names, network listening ports, etc. It is one reason that it is often more stable and simpler to run such applications in their own system (or vm).

Even when applications are well behaved there may be designed in conflicts. In network management systems, the port for SNMP traps is almost always the same. This means on any one system with one IP, you cannot have both network management systems listening for SNMP traps at the same time on the same port.

To your question on users -- that depends a lot on configuration, and what user account is being used for the database engines, for web access, for actual user logins, etc. Beyond what I can comment on, as it is your setup choices that influence that. But be aware that all the forms of access and use (both internal and from the real human user) generally show up under some user ID, most of which are under your configuration control.