Qlik Replicate Server Load Balancing and Failover

[iti-attunity-sup]

Hello Experts,

I would like to inquire about load balancing and failover on Qlik Replicate servers.

The customer would like to use two Replicate servers.

Under normal circumstances, they want to use these servers for load balancing.

If one server goes down, they want to failover to the other server and continue running the tasks that were previously handled by the failed server.

The idea is as follows:

Both servers are running without any issues → load balancing

Server#1: TASK#A
Server#2: TASK#B

When Server#1 fails, the tasks should failover to Server#2:

Server#2: TASK#A, TASK#B

I believe it's possible to pre-define both tasks on each server and start the tasks on the other server.

However, in this case, I think a full reload would be required to start the tasks. Is that correct?

I believe we cannot know the final state of the task when it was running on the failed server.

Or if there are any other methods to achieve this, could you please let us know?

Any advice would be appreciated.

Regards,

Kyoko Tajima

[john_wang]

Hello Kyoko Tajima, @iti-attunity-sup ,

I do not think the current versions of Qlik Replicate support load balance straightly. The HA is supported, please check the section: Setting up Qlik Replicate in a cluster environment. 

Hope this helps.

John.

[iti-attunity-sup]

@john_wang 

Thank you for your prompt update.

Though I described it as load balancing, it doesn't mean running a single task on multiple servers. Instead, it divides the replicated tables into multiple tasks and runs them on separate servers.

In that case, would it be possible to run the tasks on another server in the event of a failure, as I mentioned in my previous update?

Regards,

Kyoko Tajima

[john_wang]

Hello Kyoko Tajima, @iti-attunity-sup 

Thanks for the update.

In a high-availability (HA) setup, distributing tasks across multiple nodes is impossible, even for predefined tasks. This is because only one node remains active and handles all tasks, while the other node is in standby mode and cannot take on any jobs.

Regards,

John.

[iti-attunity-sup]

Hello @john_wang 

Thank you for your prompt reply.

I apologize for asking repeatedly.

Also, I apologize if my previous question was unclear.

I understand that it is not possible to execute tasks concurrently on multiple nodes in an HA (cluster) configuration.

The customer is running two replicate servers, which are NOT  HA (cluster) configuration.

In the event of a failure on one of the servers, they want the task to be taken over and executed by the other server.

In this case, I believe it is technically possible if the tasks are pre-defined and started with a full load.

However, I would like to confirm if there are any other potential concerns.

Regards,

Kyoko Tajima

[DesmondWOO]

Hi @Tajima san, @iti-attunity-sup,

Based on your description, your idea is similar to creating a new task. I think it should be fine, but the customer should ensure that the server capacity is sufficient to handle the additional load. They should also ensure that the network and firewall settings are properly configured to allow connections for the new source and target endpoints.

Regards,
Desmond

[iti-attunity-sup]

Hello @DesmondWOO 

Thank you for your update.

I understand the points to consider. (capacity and  configurations) 

Regards,

Kyoko Tajima