Forums

Hi!

I’m having trouble using ASANA API on a PythonAnywhere website, this works perfectly locally, and on their python 3.7 shell.

but, when Ran on the server I can use my API key, and confirm authentication - after that it halts.

import asana


@action('index', method=['GET', 'POST'])
def index():
    user = auth.get_user()
    message = T("Hello {first_name}".format(**user))
    key = "MYKEY"
    asana_api = asana.Client.access_token(key)
    print("asana_api, ",asana_api)
    print(asana_api.session.authorized)
    me = asana_api.users.me()                                 #halts on this line, or any use of the asana_api
    return dict(message=message)

my console logs are

2021-02-08 22:24:37 asana_api, <asana.client.Client object at 0x7f536c5f0750>
2021-02-08 22:24:37 True

and the webpage stalls until it eventually times out.

This will also time out,

for workspace in asana_api.workspaces.find_all():
        print ("workspace, ",workspace)
        workspace_gid = workspace.get('gid')

pretty much any ASANA API call except for the ones shown above will time out…

Thanks so much! Willing to provide any other necessary details

after 5 minutes or so, the logs are:

2021-02-08 22:24:37 asana_api,  <asana.client.Client object at 0x7f536c5f0750>
2021-02-08 22:24:37 True
2021-02-08 22:29:39 Mon Feb  8 22:29:38 2021 - *** HARAKIRI ON WORKER 2 (pid: 8, try: 1) ***
2021-02-08 22:29:39 Mon Feb  8 22:29:38 2021 - HARAKIRI !!! worker 2 status !!!
2021-02-08 22:29:39 Mon Feb  8 22:29:38 2021 - HARAKIRI [core 0] 10.0.0.124 - GET /simpleASANA/index since 1612823077
2021-02-08 22:29:39 Mon Feb  8 22:29:38 2021 - HARAKIRI !!! end of worker 2 status !!!
2021-02-08 22:29:39 DAMN ! worker 2 (pid: 8) died, killed by signal 9 :( trying respawn ...
2021-02-08 22:29:39 Respawned uWSGI worker 2 (new pid: 13)
2021-02-08 22:29:39 spawned 2 offload threads for uWSGI worker 2

Looks like .me() method is not returning anything and just running and request times-out after 5 minutes. What is it supposed to do?

From Glad Jones ---- This is an issue with web4py mokey-patching requests in a way that is not compatible with how we run code in web apps. We had this from Massimo DiPierro (the author of web4py):

"I tested this and if I comment gevent.monkey.patch_all() from core.py, then pythonanywhere requests works fine for me. gevent is also a problem on google app engine and we are not using it unless we have multiple workers. For now I have eliminated this dependency and picked tornado as default."

So if you patch your installation of web4py so that it does not call gevent.monkey_patch_all, it should work. Alternatively, it sounds from the above quote like Massimo may have made the change in web4py already, so maybe just updating your installation of web4py will fix it.

This worked!!!! Thank you PythonAnywhere team