Task Summary not populating or disappears

Currently on Octopus server 2020.1.18, and recently started to see behaviour where the Task Summary disappears.

image

Currently if I refresh the page (Ctrl+F5 in Chromium Edge), the items may load in. Even if they do, they seem to disappear on their own after a few more seconds.

image

This page automatically switched to the task log which shows:
image

Mostly seen on deployments that require manual intervention, although they are the projects we tend to look at most so this may not be unique to this set of projects.

Is this a known bug?

Hi @tkenyon,

Welcome to the Octopus Boards!

This definitely seems like interesting behavior.

Does this behavior only happen in Chromium Edge? Would you be able to test this in Firefox?

Are you by chance running an HA Setup? Do you have a load balancer? This sort of behavior is commonly caused by one of those two things.

Would you be able to get me a HAR file of the session to take a look?
Here are instructions on how to export the HAR file with Edge: https://support.zendesk.com/hc/en-us/articles/204410413-Generating-a-HAR-file-for-troubleshooting#Edge

If you want to DM me the HAR file for privacy, please feel free to do so.

Thanks,
Jeremy

I’ve tried reopening the same tasks from yesterday and they are now loading fine.

To replicate the issue, I reran a deploy that has a step that fails. Once the step failed and requested manual intervention, I started seeing this behaviour again.
Later, I ran a long running deploy and saw similar issues on this too. Edge and FireFox both see the issue.
I’ve also had a report that when i changed somebodys project permissions and made more available to them, the UI would show the new projects, then hide them, then show them.

I’ve also replicated the issue on FireFox.

A good while after something has completed (anywhere from 10 minutes - a couple hours), the pages look to be stable.

Setup wise - We are using 2 node HA with an f5 load balancer.

I will send the HAR file via DM shortly.

Hi @tkenyon,

There are two things you could check. Is your load balancer doing any caching? It’s possible that you are seeing the cached version of the task log page.

In your HA Server task logs, are you seeing any errors about access denied, locks, or “can’t find” like found in this documentation?

Please let me know if one of those resolves it for you or if we need to dig deeper.

Thanks,
Jeremy

I’ve had a look through the task log while a task was showing the bad behaviour. No errors seen in there.
I double checked the 2 nodes we are using and confirmed they both have the same settings when running
Octopus.Server.exe show-configuration

Just to confirm, you’re suggesting we turn off caching on our load balancer to see if this fixes it?
Apologies for the delay. I’m having to work with other teams.

Hi @tkenyon,

Yes, that is correct. If caching is enabled you may get the results you’re seeing. Please let me know if you can test that and if you can, if it works.

Let me know if you have any questions or concerns.

Thanks,
Jeremy

Hi @jeremy.miller

So, as I couldn’t change the F5 settings at the time, I decided to do the one thing people should always try on a windows machine… Turned it off and on again.

Since rebooting both nodes, the website has been looking a lot more stable.

If the issue reoccurs, we will take a look at the load balancer if possible, but for now I think we’re ok.

Thanks for your help
Tim

1 Like

Hey @tkenyon,

Thanks so much for updating me. I think we all doubt the power of the reboot sometimes. I’m glad to hear it’s performing better.

Have a great rest of your week!

Thanks,
Jeremy