Organization/status api request return 500


#1

Issue Summary

Hi, Just installed the latest version with docker-compose and everything works brilliant, instead of just one proxied request! In npm console got:

[HPM] Error occurred while trying to proxy request /api/organization/status from localhost:8080 to http://localhost:5000/ (HPE_HEADER_OVERFLOW) (https://nodejs.org/api/errors.html#errors_common_system_errors)

Try this request with curl and same, if just replace 8080 to 5000 and send curl from the terminal, everything ok, returned normal json.

Technical details:

  • Redash Version: 6
  • Browser/OS: Chrome 71 / MacOs Mohave
  • How did you install Redash: docker-compose

Understand that it can be something stupidly simple, but can’t find what I’m doing wrong.

Thanks


#2

Hi! :slight_smile:

First, it seems you are using a development environment with the version from github.com/getredash/redash, is that correct?

This issue also happened to me in my dev server lately… I’ve investigated some of it and it seems the node proxy for the backend has a limit of 80 KB for response headers. I tried to find a way augment it, however without any success.

By doing some inspect on the header, I noticed that the field with more data was X-ChromeLogger-Data. And yep, commenting the lines related to it on redash/__init__.py made this error stop on dev environment.

@arikfr do you know anything that could’ve caused this overhead recently?


#3

Hi all
did you solve the problem, i have the same ?
can you help me please


#4

Hi @makremhammali,

The way I found to workaround this in my development environment was to use npm run watch instead of npm run start. The server will be available in http://localhost:5000 (not http://localhost:8080).

If this is not your case, please give some more details (Redash version, How did you install Redash, which guide you’re following, etc.)


#5

thank you @gabrieldutra


#6

I came across this issue too. Thank you @gabrieldutra for the workaround. I do think this may be a bug though. Especially since the issue isn’t reproducible in older versions (e.g. earlier than 90a0a7d39).

Perhaps we need to re-open https://github.com/getredash/redash/issues/3538?


#7

I’m pretty sure in #3538 they didn’t start the Flask server.

But I agree we should understand the underlying issue and fix it. @emtwo can you reliably reproduce this? Any idea what the STR are?


#8
  1. docker-compose run server manage database create_tables to setup tables
  2. docker-compose up
  3. In another tab, npm run start
  4. In browser navigate to http://localhost:8080

Actual Result: The page tries to reload multiple times but does not succeed. The url has redirected to http://localhost:8080/login?next=/ and there is an error 429 too many requests In the tab from step 3, the error is [HPM] Error occurred while trying to proxy request /setup from localhost:8080 to http://localhost:5000/ (ECONNRESET) (https://nodejs.org/api/errors.html#errors_common_system_errors) on the flask server tab we see https://pastebin.com/a8pHKeEe

Expected outcome: The login page should be loaded