push server failing to connect

[melbruki]

I have setup nextcloud aio in kubernetes, everything was working perfectly for 2 months then suddenly my files stopped coming through, I am able to login through the web and all my files are there. The push server is resolving to this ::1 instead of the actual ip address of the pod, I have a single stack ipv4 network. I am not using a reverse proxy , I have a public ip address directly on the apache port of 443.
below is a self test from push. I did a test connection using the test push client, it gave me the following error

[2024-07-22 12:15:09.995455 -07:00] INFO [test_client] test_client/src/main.rs:70: Received: Authentication timeout
Error:   x WebSocket protocol error: Connection reset without closing handshake
  `-> Connection reset without closing handshake

however i was sure that the credentials were correct, so i did a push self test in nextcloud pod , below is what i received

Warning: Failed to set memory limit to 0 bytes (Current memory usage is 2097152 bytes) in Unknown on line 0
The current PHP memory limit is below the recommended value of 512MB.
✓ redis is configured
✓ push server is receiving redis messages
✓ push server can load mount info from database
✓ push server can connect to the Nextcloud server
🗴 push server is not a trusted proxy by Nextcloud or another proxy in the chain.
  Nextcloud resolved the following client address for the test request: "::1" instead of the expected "1.2.3.4"
  The following trusted proxies are currently configured: "127.0.0.1", "::1", "10.244.4.0/24", "10.99.242.1/32"
  The following x-forwarded-for header was received by Nextcloud: "10.244.4.6"
    from the following remote: ::1

✓ All proxies in the chain appear to be trusted by Nextcloud
  One of the proxies is the chain (probably 10.244.4.6) seems to have stripped the x-forwarded-for header
  Please configure the reverse proxy at 10.244.4.6 to not strip the x-forwarded-for header

  If you're having issues getting the trusted proxy setup working, you can try bypassing any existing reverse proxy
  in your setup by setting the `NEXTCLOUD_URL` environment variable to point directly to the internal Nextcloud webserver url
  (You will still need the ip address of the push server added as trusted proxy)
nextcloud-aio-nextcloud-84c94c8d9c-q258z:/var/www/html# 

NAME                                            READY   STATUS    RESTARTS   AGE   IP             NODE            NOMINATED NODE   READINESS GATES
nextcloud-aio-apache-86cb4774f7-d6628           1/1     Running   0          22d   10.244.4.31    talos-fa1-gf4   <none>           <none>
nextcloud-aio-collabora-6657d56bd6-lz48m        1/1     Running   0          22d   10.244.4.159   talos-fa1-gf4   <none>           <none>
nextcloud-aio-database-6d6dd56565-5mlwm         1/1     Running   0          22d   10.244.3.13    talos-2m3-rai   <none>           <none>
nextcloud-aio-fulltextsearch-7b66db7b94-5ht72   1/1     Running   0          22d   10.244.4.216   talos-fa1-gf4   <none>           <none>
nextcloud-aio-nextcloud-84c94c8d9c-q258z        1/1     Running   0          22d   10.244.4.14    talos-fa1-gf4   <none>           <none>
nextcloud-aio-notify-push-98bfbbdd6-4578v       1/1     Running   0          26m   10.244.4.6     talos-fa1-gf4   <none>           <none>
nextcloud-aio-redis-65dd49dc78-tbh59            1/1     Running   0          22d   10.244.3.242   talos-2m3-rai   <none>           <none>
nextcloud-aio-talk-b57f8b967-vjk9f              1/1     Running   0          22d   10.244.4.100   talos-fa1-gf4   <none>           <none>
root@ip-172-26-1-22:/home/ubuntu# 

[szaimen]

Hi, which AIO version is this? Can you upgrade to at least v9.1.0 and check if that fixes things?

[melbruki]

I am on version 28.0.5 of nextcloud and I believe that corresponds with aio version 8.2 if I'm not mistaken I don't want to upgrade to version 29 just yet

[szaimen]

All right. Moving to discussions then since I am pretty sure that it is fixed with v9.1.0 and above

[melbruki]

Ok so I'm on aio version 8.2 , you are saying that it's fixed in version 9. However is there a way to fix this without upgrading to Nextcloud 29

[szaimen]

I dont think so

[miconx]

i have the same problem wit AIO 9.6.0 and nextcloud version 30.0.0
is the problem back again?

[vauvenal5]

Hi!
I am having the same or a similar issue on the latest AIO v11.6.0 with Nextcloud Hub 10 (31.0.8)

Notify push container log:

[2025-08-23 15:32:52.012108 +00:00] ERROR [notify_push] src/main.rs:84: Self test failed: Error while communicating with nextcloud instance: error sending request for url (https://my.cloud.com/index.php/apps/notify_push/test/version)

Push self-test results:

nextcloud-aio-nextcloud:/var/www/html$ ./occ notify_push:self-test
✓ redis is configured
✓ push server is receiving redis messages
✓ push server can load mount info from database
✓ push server can connect to the Nextcloud server
🗴 push server is not a trusted proxy by Nextcloud or another proxy in the chain.
  Nextcloud resolved the following client address for the test request: "my.public.ip.addr" instead of the expected "1.2.3.4" test value.
  The following trusted proxies are currently configured: "127.0.0.1", "::1", "172.19.0.0/16"
  The following x-forwarded-for header was received by Nextcloud: "my.public.ip.addr"
    from the following remote: 127.0.0.1

✓ All proxies in the chain appear to be trusted by Nextcloud
  One of the proxies is the chain (probably my.public.ip.addr) seems to have stripped the x-forwarded-for header
  Please configure the reverse proxy at my.public.ip.addr to not strip the x-forwarded-for header

  If you're having issues getting the trusted proxy setup working, you can try bypassing any existing reverse proxy
  in your setup by setting the `NEXTCLOUD_URL` environment variable to point directly to the internal Nextcloud webserver url
  (You will still need the ip address of the push server added as trusted proxy)

My setup is pretty straight forward. Just running the AIO master container. No reverse proxy in place. My infrastructure has DDNS setup in place and my URL my.cloud.com is being resolved to the public IP of my router.

Only thing is that my server has multiple network ports and I have bound the AIO setup only to one interface by setting APACHE_IP_BINDING and adapting the respective port forwards. However, stopping the other service and removing the APACHE_IP_BINDING did not change anything.

I also tried resetting the docker network. Any hints will be much appreciated.

[vauvenal5]

One update here. By sheer coincidence I decided to try occ notification:test-push user and discovered that some of my devices do receive the push notifications. After deleting and re-installing the app everything seems to work.

So now I even understand less of what the self-test is trying to tell me.