Fresh V8.0.0 install login not working

SuiteCRM Forum - English Language
1

Hi people,
i just setup my new container behind a nginx reverse proxy with v8.0.0 .
The setup process went smooth and everything is fine.
But when i try to login to suitecrm it logs me out immediately.

In the console i get an error when i click login (with the right credentials)

GET https://mydomain.com/api/index.php?action=Login&module=Users 404 (Not Found)

I guess it is because of the proxy and the somehow not working mod_rewrite function
Any ideas? Did somebody solve this already?

2

Hi @AlexHammerschmied,

The following thread may help:

Note: Please note that nginx is not officially supported.

3

Thank you Clemente,
this is interesting, because i am running suitecrm on an apache server. It is just behind a reverse proxy which runs with nginx.
But the directives are actually handled by the apache server.
At least that is what i think :slight_smile:

Did someone else run into “problems” like this already?

4

Hi @AlexHammerschmied,

No problem : )

Haven’t heard of any issue with using nginx as reverse proxy yet.

I’m not able to help much more, sorry

5

in the browser console it gives me these errors:

Cookie “LEGACYSESSID” has been rejected because it is already expired.

Cookie “PHPSESSID” has been rejected because it is already expired.

Cookie “PHPSESSID” has been rejected because it is already expired.

6

Hi there!

I’m having the same issue, but in my case, I’m trying to deploy SuiteCRM in Kubernetes using a GKE cluster.

I’m using a LoadBalancer service type, in my case GKE’s Network LoadBalancer, and SuiteCRM 8 fails with the same issue described in the initial comment here:

GET http://mydomain.com/api/index.php?action=Login&module=Users [404 Not Found 295ms]

When attempting to log in through the load balancer, the UI will display the error: ‘Error occurred while fetching metadata’.

Another significant detail is that I have been able to log in using /legacy, and login also works normally if I access directly using kubectl port-forward instead of the load balancer.

Please see below screenshots from the failing session (through the gke load balancer), compared with the successful session (direct connection):

Session failed:

photo_2022-02-11_16-35-42
photo_2022-02-11_16-35-421280×616 98.5 KB

In the requests, I see the following:

  • Several Graphql 200, but the latest, requesting modStrings I see cookie LEGACYSSID deleted and empty response.
  • Graphql 301 requested userPreferences.
  • Error 404 GET http://mydomain.com/api/index.php?action=Login&module=Users

I have noticed the following differences between the successful and failed sessions:

  • Cookie “PHPSESSID” has been rejected because it is already expired. is displayed twice, while in the successful session it is only displayed once.

  • Cookie LEGACYSSID deleted in graphql response.

  • Failed session used:
    GET http://mydomain.com/api/index.php?action=Login&module=Users

    while successful call used
    GET http://localhost:8080/legacy/index.php?module=Users&action=SetTimezone

If need further details please let me know. I attempted to upload more screenshots but the website didn’t let me.

7

I think this happens because of a load-balancer/reverse-proxy.
I am not sure though. I gave it up to get up and running quickly and created a simple vps with apache like “in the old times”. No containers for now :frowning: