Skip to main content

Hi All,

We are working on adding new FME Engines on a Separate Linux Machine in a fault-tolerance environment where two FME Cores have been configured. 

Following the FME Flow Administrator’s Guide

https://docs.safe.com/fme/html/FME-Flow/AdminGuide/Adding_FME_Engines_on_Separate_Machine.htm

The install instructions have this step on the Linux section,

The installer prompts you for the name of the <coreHost>. In a fault-tolerant installation, the primary <coreHost> should always be specified. Do not specify the failover host.
Here we are not very clear. We have two FME Flow Cores.

first, how do we know which one is supposed to be the primary or the failover host between two Cores?

Secondly, if say that the first Core to start can be considered the primary, what happens to the engines on a Separate Engine if the primary Core is down?

Can anyone give us some helps on this?

Thanks,

-John

Hi @missp29,

The install instructions have this step on the Linux section,

The installer prompts you for the name of the <coreHost>. In a fault-tolerant installation, the primary <coreHost> should always be specified. Do not specify the failover host.
Here we are not very clear. We have two FME Flow Cores.

In a failover scenario, it doesn't matter which core is designated as primary or failover; the engine simply needs to be registered with a core. The current wording is a bit unclear, I've contacted our publication team to see if we can clarify this.

 

Secondly, if say that the first Core to start can be considered the primary, what happens to the engines on a Separate Engine if the primary Core is down?

The configuration settings for the failover host are stored in the database. If a core goes offline, the distributed engines will first try to reconnect to that host before attempting to connect to the secondary host. Any engines local to that host will also go down unless they are manually reassigned or redirected to an active host using the REST API.

 

See A Guide to Choosing your FME Server Deployment Architecture for more information and diagrams of this kind of scenario.

 

Hope that helps!

 

 


Hi dylan.at.safe,

Thank you very much for your helps here.

As you explained below,

The configuration settings for the failover host are stored in the database. If a core goes offline, the distributed engines will first try to reconnect to that host before attempting to connect to the secondary host. …

 

I am not very clear here. The distributed server only has the primary Core info.  if the primary Core goes down, how could the Distributed Server know where the second Core is? As you mentioned that the configuration settings for the failover host are stored in the database, does this mean the Distributed Servers will get the second host directly from the database?

Thanks,

-John


Hi @missp29,

 

The distributed server only has the primary Core info.  if the primary Core goes down, how could the Distributed Server know where the second Core is? As you mentioned that the configuration settings for the failover host are stored in the database, does this mean the Distributed Servers will get the second host directly from the database?

You are correct, the distributed server will get the second host from the database.

 

 


Hi dylan.at.safe,

I see. Again I really appreciate your help!

Thanks,

-John


Reply