Did I break Confluence Cloud? (aka Site restore not possible, aka long running task never done)

[Janek Schumann]

Dear Community,

Did anyone ever do successful Site Restore in Confluence Cloud?

It seems I broke my site beyond repair and don't know what to do no.
After 3x2 hours of waiting for an error message to pop up, the 4th try was the charm. However, now it is telling me that

A long-running task has been initiated by another user. To avoid conflicts, you will not be able to update apps until this task has completed

Since this is Cloud, I cannot do anything about compared to the good old Server days...

---

First things first, what I did to end up here:

1. I wanted to test out a Site Restore. I don't trust the backup system of Cloud platforms in general. I wanted to make sure a backup actually recovers everything, *especially* Custom Entity Content aka custom app data / properties. (which Atlassian back in 2021 did not backup - they do now - but I just ran into the issue yesterday that a Site import *did not* recover app data. I see it in the backup data. I can query it via API in the productive system. But it is empty in the test system after import. Therefore I decided to go nuclear and do a Site Import instead, maybe that's the way to go, I thought)

2. I tried this and the restore process acquired a lock on the the system (cool), deleted all content (cool, I guess...), deleted all users (okay, fine), ......... to get stuck. This whole thing needed 65 seconds, and then it took another 2 hours 20 minutes or so to say "error". For the entire time it was stuck at "Existing data removal completed".

3. I tried this 2 more times with the same result. Waiting 2 hours and 20 minutes each time. After all, the best approach is retry, isn't it?

4. Since I don't give up that easily I gave it one more shot. It seems this time it worked.

HOWEVER, I am now stuck at the point where the apps need to get installed!

And this is blocked by above error message.

What shall I do? Just wait? For a day? For a week?

based on the timestamp "(Initiated by undefined on 3/10/2023, 11:41:07 AM)" it seems it was triggered by the first site import, not the 4th. It is now almost 36 hours later and the task is still running and blocking app install and configuration.

Any tips are much appreciated,

J.

[Janek Schumann · Accepted solution]

7 days and counting.

No response from Atlassian.

Deleted the product now. Moving on to the next bug I found in Cloud: app data isn't imported back from a space export. Yay. All data lost. Thank you Atlassian. Wonderful...

[Janek Schumann]

@atlassian: any chance this can get resolved?

we are at 104 hours of runtime for that task.

FYI: feature request: have a timeout for long-running tasks. ;)

[Janek Schumann]

They weren't kidding with "long-running task": 50 hours and counting. I will check once a day and see if there is a timeout at some point.