Qlik Sense hub and QMC not reachable after installing a Windows Server update

[Mika_2000]

Hello everyone reading this!

Over the past few days, we've noticed something related to Qlik Sense environments. Four of our clients have reported an incident where the Qlik Sense Hub and QMC are not reachable, this is something uncommon and happens rarely. Once logged into their servers we tried the usual stuff.

1. Restarting the Qlik Sense services.
2. Rebuilding the Qlik Sense service certificates.
3. Rebooted the server.
4. Checked the firewall

Etc etc, for these four clients these fixes did not work. We went to take a look at the Event Viewer and noticed that PostgreSQL kept logging the following error.

FATAL: no pg_hba.conf entry for host <Ipv6 address of the server> iser “qliksenserepository”, database “QSMQ”, SSL off.

Investigating this event on the internet we saw that we needed to change this pg_hba.conf file to accept the server’s ipv6 address which seems logical. However, this ipv6 address is already registered in this .conf file.

Looking further on our client’s server we noticed that this issue occurred after upgrading their Windows server.

It seems that this update has triggered this issue to all four clients affected by this issue. They all installed this update, rebooted the server and tried to start Qlik Sense but their browser just says “This webpage is not avaible” also when trying to reach https://localhost:443/qmc or /hub.

We have told our clients to restore a backup from 08-11-2022 and 09-11-2022 but this did not have any affect, once we rebuild the certificates, their environments did work again.

We think this issue is triggered by the Windows update because all clients affected by this issue run on a different Qlik Sense release. Has anyone else come across this issue? How did you fix it? Is this issue known by the Qlik Team or is it just coincidence that all four clients ran across the same issue after updating their environment? We would like to hear what you think about this issue :D. Once we find out what really happened we will of course update this ticket.

If you have any questions feel free to ask me!

[Mika_2000]

Alright!

In about 30 minutes I have a meeting with a customer with this problem.

I will try to add this entry! Thanks for your reply 🙂

[Rakesh_HB · Accepted solution]

Hello All,

If you see the below KB Article from Microsoft, there are a few changes Microsoft had made for DNS and HNS with this Windows Patch. Please read it carefully.

• November 8, 2022—KB5019966 (OS Build 17763.3650)

Thank you!

Rakesh

[Mika_2000]

Thanks @jonas_brorstad this did fix the issue with our client I will mark this as a solution 😄

@Rakesh_HB, thanks for your reply.

I will take a look at this page, right now we are going to update a couple of our internal servers to KB50109166 and see what happens, if these servers also get the same issue I will let yo know. Fingers crossed...

[Mika_2000]

So we have updated some internal servers but none ran into this issue, it is worth noting that the servers we updated are not part of a domain and do not have a domain controller. The customers affected by this issue do have a domain and domain controller. In the link you provided there are some known issues with domain controllers.

After installing updates released on November 8, 2022 or later on Windows Servers with the Domain Controller role, you might have issues with Kerberos authentication. This issue might affect any Kerberos authentication in your environment. Some scenarios that might be affected:

Maybe this could be something?

[Rakesh_HB]

Hello @Mika_2000 ,

Domain/Domain controller uses IPv6 address.

[maxime_belanger_UQ · Accepted solution]

We had the same issue, on Qliksense February 2022, Windows Server 2019.

Our achitecture has 4 Qliksense nodes, with the primary node having the problem with the postgresql service not starting after the November 8, 2022—KB5019966 (OS Build 17763.3650) WIndows Patch that was installed on November 11th in our case.

Adding the ipv6 ::0/0 address in the pg_hba.conf as stated in the solution fixed the problem after restarting the server. For other people having a hard time finding the file, it is in a hidden folder by default; for us it was here:  C:\ProgramData\Qlik\Sense\Repository\PostgreSQL\12.5\pg_hba.conf  )