Workaround:

docker cp aaf_aucore_1:/opt/AuCore/static/nginx.conf.j2 /tmp/​

vi /tmp/nginx.conf.j2 ​

Change proxy_buffer_size value from 8k to 16k, so it must be:​
proxy_buffer_size 16k;​

Add a couple of new lines below:​
proxy_busy_buffers_size 24k;​
proxy_buffers 64 4k;​

docker cp /tmp/nginx.conf.j2 aaf_aucore_1:/opt/AuCore/static/​

systemctl restart aauth​

The proposed values may not be enough if a user is a member of about 100 groups. We are currently working to estimate the optimal values.

docker cp aaf_aucore_1:/opt/AuCore/static/nginx.conf.j2 /tmp/​  ==> this returns a file without name in the tmp directory.

docker cp aaf_aucore_1:/opt/AuCore/static/nginx.conf.j2 /tmp/​nginx.conf.j2 ==> this returns a file with the correct name in the tmp directory.

But when I do a vi on that file it still edits a 'new' empty file. Though the filesize is 11145 bytes.

Did I miss something?

I succeeded to implement the workaround by specifying the filename in the destination and using another folder. But it does not make it better.

Please check if the command will show the three parameters with the new values?

cat /var/lib/docker/volumes/aaf_webd-config/_data/nginx.conf | grep proxy_bu

proxy_buffer_size 16k;

proxy_busy_buffers_size 24k;

proxy_buffers 64 4k;

This is the output:

Last login: Mon Apr 12 14:54:32 2021 from 10.0.0.111
fs-naaf:~ # cat /var/lib/docker/volumes/aaf_webd-config/_data/nginx.conf|grep proxy_bu
            proxy_buffer_size 16k;
            proxy_busy_buffers_size 24k;
            proxy_buffers 64 4k;

Seems correct?

Yes.

Do you still see the error "WebAuth feature is not running" or this was changed?

How big your membership in the repository groups?

If you will enable the Developer Tools in your browser, how much length will be the URL in the operation before getting the error? 

Here you are.

The zip seems to be corrupted. It's only 230 bytes.

zuuut..

Based on the errors I see, I recommend creating a new OAuth 2.0 event and apply the new Event name, Client ID, and Client Secret on the Filr side.

If the issue still not solved after that and you see the same error, did you experience it in the previous version you had and which version you had before? 6.3.3?

I just ran into this same issue with Access manager (4.5.3.3).  I tried creating a new event but that didn't help.  I increased the nginx header size as suggested and then ran into the header size issue in access manager and had to increase that as well.  Once the headers were increased on both systems it worked.

Is this suggested nginx change persistent in the container?  

I had this issue in my production environment even though dev and test worked fine.  

Is this a know issue in 6.3.4? 

Thanks,
Jeremiah

Hello George and   others,

The thing with the groups is one thing. Hopefully another defect, known as defect 327040 will be solved with this patch.

On Thursday I had a conversation with development, they state it should be fixed as well.

Thanks for the update.  Does the access manager integration with advanced auth pass groups?