datopian / datahub-qa

:package: Bugs, issues and suggestions for datahub.io
https://datahub.io/
32 stars 6 forks source link

Users cannot log in to datahub.io (server error: 524) #260

Closed sglavoie closed 4 years ago

sglavoie commented 4 years ago

When I try to log in to datahub.io, I want to be able to access the website quickly and perform actions in my account so that I can work with datasets from the browser and the command line.

Acceptance

Tasks

Analysis

Error observed

Screenshot_2020-04-30_13-23-41

User is logged in (was previously logged out)

Screenshot_2020-04-30_13-24-19

sglavoie commented 4 years ago

FIXED. This is a duplicate of #259. As reported in that other issue:

rufuspollock commented 4 years ago

@sglavoie how is this a duplicate of #259? they seem quite different ...

sglavoie commented 4 years ago

Maybe there was some confusion on my part. It looks to me like @monikappv and I are reporting the exact same issue resulting in a 524 server error. It is very likely that not being able to log in as a user is due to an error on the server (HTTP status code >= 500) that prevents access to other parts of the website, including the dashboard and the login route.

I will of course re-open this issue if you see it fit. :+1:

rufuspollock commented 4 years ago

@sglavoie i don't think you know these are the same source - 524 is a really generic error. I would reopen and investigate. I think you'll need to ask on Datopian devops b/c it will prob relate to internal apis in some way.

Also an aside it should not be FIXED but DUPLICATE (if indeed it were a duplicate).

sglavoie commented 4 years ago

Yes, you are absolutely right. I will be careful not to make such assumptions from now on and re-read the docs about closing issues as I certainly forgot about that other label. Thank you.

sglavoie commented 4 years ago

FIXED: All actions can be performed as normal (log in, loading pages, uploading datasets, etc.) and there is no more timeout issue when the user is logging in.


Solution by @zelima:

  1. Go to GCP -> CPC Network -> External IP addresses
  2. If the issue is the same as above you should see 4 Unused external IPs (In use by - None) and 4 used ephemeral IPs (2 for prod and 2 for testing).
  3. Click change on the right of the External IP and chose appropriate service (prod for prod and test for testing, from top to bottom)
  4. Save and wait for a bit
rufuspollock commented 4 years ago

@sglavoie can you check if this is happening again?

sglavoie commented 4 years ago

@rufuspollock I believe we are dealing with the same issue again. I let @zelima know about it since he was able to provide a fix last time on GCP (solution linked above).

zelima commented 4 years ago

@sglavoie Should be fixed.

Aside: Please ping me in the Gchat if I do not respond in a reasonable time

sglavoie commented 4 years ago

@zelima All working again! :tada:

Please ping me in the Gchat if I do not respond in a reasonable time

I will. Thank you. :slightly_smiling_face: