Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
Dear Qlik Experts,
Environment Details:
-----------------------------------------------------------------------------------------------------------------------------------------------------
Central node on Windows 2019 server with QS Nov 2023 Patch 9
URL - www.centralnode.com and IP is 1.2.3.4
RIM node (FailOver) on Windows 2019 server with QS Nov 2023 Patch 9
URL is - www.rimnode.com and IP is 5.6.7.8
Repository Db(not within Qlik Sense Central) / Postgredb 14.11 version is in Linux server
-----------------------------------------------------------------------------------------------------------------------------------------------------
Manual Failover:
-----------------------------------------
1. Currently, I am doing a manually failover by forcing Qlik services stopping in central node to take the control by FailOver node (RIM node) which is working fine.
2. Before Fail over qmc/hub URL was - www.centralnode.com/hub
3. After Fail Over qmc/hub is working with - (www.rimnode.com) which means fail over is working, but I don't want to communicate or force users to use (www.rimnode.com/hub) or feel the difference of environment usage hence to use the www.centralnode.com/hub after fail over so I have contacted the network team to switch the IPs of Central to RIM node (Fail Over) so users will be using same URL (www.centralnode.com/hub). The moment Central Node IP is mapped to FailOver node, QMC/hub url displayed with http error with "Qlik logo + 400 Bad Request The http request header is incorrect." with url lock showing as "not secure" (it was secure before failover).
Also before starting Fail Over I did below:
-------------------------------------------------------------
1. Copied the Central node Qlik Client and Service Cluster Certficate to RIM node
2. White listed Central and RIM nodes in virtual proxies of both Central and RIM.
3. Browser/URL SSL Thumbprint of Central and RIM is different as we have different certificates.
Can someone help/assist me if any other certificate/thumbprint/something else to be done to make it work? Did I miss something ? Please suggest
4. Once the above theory of IP switching works, I will map 2 IP addresses to Central node so fail over will become automatic the moment central goes down.
Thanks,
Anil
I don't think you should be switching IPs of machines. It will cause conflicts within the DNS records which might cause issues for the network team and have to be addressed later.
If you open cmd and type nslookup followed by the server IP or machine name you will see the name or IP of the server. If you give the a server's IP to another, 2 machine names might refer to 1 IP and will cause issues.
I suggest you relay your requirements to the network team and let them handle the issue.
Why not put them behind a load balancer with a URL such as https://qlikbi.mycompany.com?
The load balancer will redirect to either working server without changing the URL. It can be either a physical based (hardware) or software based load balancer.
You can then make a virtual proxy or edit existing virtual proxies to include the new URL.
Hi @M_B,
Many thanks for your reply.
May I ask what exactly I have to ask the load balancer team to do? Can you explain with below URLs ?
Current setup :
------------------------
URL - www.centralnode.com (FQDN) and IP is 1.2.3.4 
URL is - www.rimnode.com (FQDN) and IP is 5.6.7.8
What Changes I need to request for the above setup?
Thanks
Ail
Designate a site such as qlikbi.mycompany.com with an IP 1.3.5.7 and add the 2 IPs (1.2.3.4 and 5.6.7.8) under the new URL. They should know what to do if you ask them to provide 1 URL to access both machines and redirect in case either of them fail.
I find the URLs to be strange. Usually a machine name is appended to the main domain URL like https://qlikcentral.mycompany.com and https://qlikrimnode.mycompany.com. You can then give them a unified URL like https://qlikbi.mycompany.com with a load balancer.
Honestly, it is not something I have first hand experience with setting up but that is how the service is set up for our users.
Hi @M_B ,
You are right.
As I should not disclose the actual name in the open forums so I gave random names here for the topic to understand. But sorry for the confusion for random naming.
Actually our URLs look like below I simulated of ours.
FQDNs
qliksense.systems.us.hdp (Central) - and the URL Users currenlty using is qliksense.systems.us.hdp/hub
qliksense1.systems.us.hdp (RIM) (Fail Over)
qliksense11.systems.us.hdp (RIM)
----
At the moment your suggest setup is not done yet.............
Currently I am testing on UAT thinking that switching of IP works but it is not working. So from my initial query does the switching of IP from Central to RIM wont help for the Fail Over to work ?
Any thoughts what am I missing ?
I will do your suggested setup by contacting network team but before that I am trying to achieve above switching of IPs.
Please suggest
Thanks for your reply
Anil
.
I don't think you should be switching IPs of machines. It will cause conflicts within the DNS records which might cause issues for the network team and have to be addressed later.
If you open cmd and type nslookup followed by the server IP or machine name you will see the name or IP of the server. If you give the a server's IP to another, 2 machine names might refer to 1 IP and will cause issues.
I suggest you relay your requirements to the network team and let them handle the issue.
Hi @M_B ,
Automatic Fail over is working after using Network Load balancing
Thank you for your suggestion
Good to hear that everything is working.
Glad I could help.