Scheduler problem

[YoussefBelloum]

hi everyone,

I hope you're doing well.

I'm administrating à multi node Qliksense site.

1 central

1 failover

2 consumers

On the scheduling, the cental is MANAGER and the rest of the node are WORKERS.

Everyday users complains that they can't run reload tasks (the restart of the scheduler service on all nodes solves the issue) and today I found a weird message on the logs:

The log is saying that : Scheduler is not an active master

I've made some research on that and I found nothing on the concept of "active master".

What does it mean "Active master" ? and how does a MASTER is supposed to run the task knowing that the MANAGER is the role that can launch the tasks...

Anyone faced this issue ?

Thanks 

Youssef

[Miguel_Angel_Baeyens]

I have seen that error message when there are one or more failover candidates and one of them assume the role of Central when the original central disconnects or fails or its services were shut down for some reason, like maintenance or upgrades.

Check this article to verify which node is central and make sure that one is set to manager or manager and worker:

https://community.qlik.com/t5/Official-Support-Articles/Manually-Failover-the-Central-Node/ta-p/1714...

[YoussefBelloum]

Hi @Miguel_Angel_Baeyens ,

Thanks for your reply.

Actually the central node is already set as "manager and worker" and the failover node is set to "worker".

Maybe I need to set the failover node as "Manager and worker" also ?

[Miguel_Angel_Baeyens]

Sorry for the late reply, but yes, the failover candidate must have all services, settings and setup the same as the original central node. Note that setting up the failover candidate as "manager and worker" will help sending tasks to other schedulers and also being able to perform reloads itself. If the central/failover node is small (which is not infrequent) you want it to be Manager, but not Manager and Worker, so the heavy lifting of the reloads is taken care of by better sized Scheduler nodes.