Hi Everyone,
I recently had an issue with users being unable to view apps in the hub until I updated virtual proxy settings.Users were able to see the streams they had access to, but all the streams were 'empty' even though the apps existed in the QMC.
Our current Qlik configuration has 2 engine rim nodes, at the time of the incident the affect virtual proxy was only load balancing to a single Engine rim node.
Initially, I did not see the problem as I was accessing the hub through a different virtual proxy which I use for testing purposes, but end users were affected when they went to their default virtual proxy. After realizing the issue was happening at the virtual proxy level I made 2 changes.
1) I added the second engine to affected virtual proxy, since at the time of the incident it was only load balancing to a single engine node
2) I updated the host whitelist to mimic the whitelist of my test virtual proxy. After the whitelists were updated the proxy nodes were restarted by QMC and the issue resolved.
Speaking with support it seems like the adding of the second engine node is likely what fixed it, but why is this the case? Both engine nodes are able to access the fileshare for the apps listed in the service cluster configs in QMC, and my understanding is load balancing to a single node would only effect performance as more users would be sent to a single engine rather than splitting the load between them.
Any idea what could've caused this behavior? Or where to look to track this down?
