Cloudron Forum

I tested this again more carefully.

The TURN service does indeed get stopped after app restart because of stopUnusedServices, but PeerTube itself continues to work normally and the socket/WebRTC behavior seems fine.

So my earlier assumption was probably wrong: the TURN service stopping does not appear to be the cause of the PeerTube issue I saw before. That may have been a separate glitch.

So it seems TURN is currently unused in my setup, and Cloudron stopping it is expected with the new unused-services cleanup behavior.

Thanks for clarifying.

Thanks, this likely explains it.

What confused me is that this only started happening after a Cloudron upgrade. Before that, restarting PeerTube did not stop the TURN service, so P2P continued to work normally.

I did not manually configure TURN inside PeerTube itself. I assumed the existing Cloudron TURN service would continue to be available to PeerTube as before.

So from my side, there is no custom TURN config to share. The main issue is that after the Cloudron upgrade, restarting PeerTube now triggers stopUnusedServices and TURN gets stopped, which breaks PeerTube P2P/WebRTC for users behind strict NAT.

So it seems the package currently does not declare TURN as an addon dependency, and this only became visible after the newer Cloudron service cleanup behavior.

Hi,

I noticed an issue when restarting my PeerTube app.

After triggering an app restart, Cloudron runs stopUnusedServices and stops several services:

services: stopUnusedServices: stopping mysql (no apps use it)
services: stopUnusedServices: stopping mongodb (no apps use it)
services: stopUnusedServices: stopping turn (no apps use it)

This happens consistently right after restarting PeerTube.

While MySQL and MongoDB might not be used, TURN is required for PeerTube WebRTC P2P functionality.

When TURN is stopped, P2P connections fail (especially for users behind strict NAT), and PeerTube falls back to HTTP streaming.

Expected behavior:
TURN should remain running if PeerTube requires it.

Actual behavior:
TURN is stopped automatically after app restart and must be restarted manually.

It seems Cloudron does not detect TURN as a dependency of PeerTube.

Is this expected behavior, or is there a misconfiguration or bug in addon detection?

Thanks.

Same problem in my peertube instance.

its working now. but with a little trick for my server. so i make postgree not start by making wrong custom.conf postgresql. the problem is when postgres start the server immediately crashes and can't do ssh or ping.
then i can enter dashboard and choose postgres recovery mode.

thanks for your help.

here some error log 
2025-03-24T13:21:29Z server started
2025-03-24T13:21:29Z ==> Waiting for Postgres
2025-03-24T13:21:29Z pg_ctl: server is running (PID: 15)
2025-03-24T13:21:29Z /usr/lib/postgresql/16/bin/postgres "-D" "/var/lib/postgresql/16/main"
2025-03-24T13:21:29Z ==> Postgres is running
2025-03-24T13:21:29Z ==> Resetting root password
2025-03-24T13:21:29Z ALTER ROLE
2025-03-24T13:21:29Z ==> Upgrading pgvectors
2025-03-24T13:21:29Z ERROR:  database "cloudronpgvectorupdate" already exists
2025-03-24T13:21:29Z 2025-03-24 13:21:29.801 UTC [30] root@postgres ERROR:  database "cloudronpgvectorupdate" already exists
2025-03-24T13:21:29Z 2025-03-24 13:21:29.801 UTC [30] root@postgres STATEMENT:  CREATE DATABASE cloudronpgvectorupdate
2025-03-24T13:22:30Z ==> Detected existing installation
2025-03-24T13:22:30Z ==> Copying updated config
2025-03-24T13:22:30Z ==> Updating existing postgresql
2025-03-24T13:22:30Z waiting for server to start....2025-03-24 13:22:30.157 GMT [15] LOG:  skipping missing configuration file "/var/lib/postgresql/16/main/custom.conf"
2025-03-24T13:22:30Z 2025-03-24 13:22:30.157 GMT [15] LOG:  skipping missing configuration file "/var/lib/postgresql/16/main/custom.conf"
2025-03-24T13:22:30Z 2025-03-24 13:22:30.186 UTC [15] LOG:  starting PostgreSQL 16.6 (Ubuntu 16.6-0ubuntu0.24.04.1) on x86_64-pc-linux-gnu, compiled by gcc (Ubuntu 13.2.0-23ubuntu4) 13.2.0, 64-bit
2025-03-24T13:22:30Z 2025-03-24 13:22:30.186 UTC [15] LOG:  listening on IPv4 address "0.0.0.0", port 5432
2025-03-24T13:22:30Z 2025-03-24 13:22:30.186 UTC [15] LOG:  listening on IPv6 address "::", port 5432
2025-03-24T13:22:30Z 2025-03-24 13:22:30.189 UTC [15] LOG:  listening on Unix socket "/var/run/postgresql/.s.PGSQL.5432"
2025-03-24T13:22:30Z 2025-03-24 13:22:30.195 UTC [18] LOG:  database system was interrupted; last known up at 2025-03-24 13:21:29 UTC
2025-03-24T13:22:30Z 2025-03-24 13:22:30.315 UTC [18] LOG:  database system was not properly shut down; automatic recovery in progress
2025-03-24T13:22:30Z 2025-03-24 13:22:30.318 UTC [18] LOG:  redo starts at D/D9A09748
2025-03-24T13:22:30Z 2025-03-24 13:22:30.318 UTC [18] LOG:  invalid record length at D/D9A0AE20: expected at least 24, got 0
2025-03-24T13:22:30Z 2025-03-24 13:22:30.318 UTC [18] LOG:  redo done at D/D9A0ADC8 system usage: CPU: user: 0.00 s, system: 0.00 s, elapsed: 0.00 s
2025-03-24T13:22:30Z 2025-03-24 13:22:30.326 UTC [16] LOG:  checkpoint starting: end-of-recovery immediate wait
2025-03-24T13:22:30Z 2025-03-24 13:22:30.337 UTC [16] LOG:  checkpoint complete: wrote 3 buffers (0.0%); 0 WAL file(s) added, 0 removed, 0 recycled; write=0.003 s, sync=0.003 s, total=0.013 s; sync files=4, longest=0.001 s, average=0.001 s; distance=5 kB, estimate=5 kB; lsn=D/D9A0AE20, redo lsn=D/D9A0AE20
2025-03-24T13:22:30Z 2025-03-24 13:22:30.342 UTC [15] LOG:  database system is ready to accept connections
2025-03-24T13:22:30Z done
2025-03-24T13:22:30Z server started
2025-03-24T13:22:30Z ==> Waiting for Postgres
2025-03-24T13:22:30Z pg_ctl: server is running (PID: 15)
2025-03-24T13:22:30Z /usr/lib/postgresql/16/bin/postgres "-D" "/var/lib/postgresql/16/main"
2025-03-24T13:22:30Z ==> Postgres is running
2025-03-24T13:22:30Z ==> Resetting root password
2025-03-24T13:22:30Z ALTER ROLE
2025-03-24T13:22:30Z ==> Upgrading pgvectors
2025-03-24T13:22:30Z 2025-03-24 13:22:30.509 UTC [30] root@postgres ERROR:  database "cloudronpgvectorupdate" already exists
2025-03-24T13:22:30Z 2025-03-24 13:22:30.509 UTC [30] root@postgres STATEMENT:  CREATE DATABASE cloudronpgvectorupdate
2025-03-24T13:22:30Z ERROR:  database "cloudronpgvectorupdate" already exists

i am using ubuntu 22 on Dual Xeon E5-2670 96GB memory.

i give 36GB
how to view postgress error log. maybe it can be done through system rescue. but where is the config that I have to edit to stop postgress on booting..
i cant do anything in my server. just crashed. after postgressql start.

its about 32GB. raise to 8GB.
i have new problem now. server crashed now. already remove custom.conf .
If I look at htop, it seems that when starting postgressql, the server cannot be accessed immediately.

@joseph said in cant start peertube after updateing latest cloudron:

@awnzky interesting, so you had to create an empty custom.conf ?

i just copy paste from default conf. and rise temp_file_limit

now my peertube is running. problem solved.

after create custom custom.conf in /home/yellowtent/platformdata/postgresql/16/main# error message is gone.

my peertube still not responding.
error log
2025/03/24 02:04:50 [error] 316#316: *357 connect() failed (111: Connection refused) while connecting to upstream, client: 172.18.0.1, server: default_server, request: "GET / HTTP/1.1", upstream: "http://127.0.0.1:9000/", host:

from htop showing process postges create index.

error log

Mar 24 07:37:21 2025-03-24 00:37:21.628 UTC [1762] userf8ee153b7f89431b8a74be747b18e88c@dbf8ee153b7f89431b8a74be747b18e88c ERROR: temporary file size exceeds temp_file_limit (4194304kB)
Mar 24 07:37:21 2025-03-24 00:37:21.628 UTC [1762] userf8ee153b7f89431b8a74be747b18e88c@dbf8ee153b7f89431b8a74be747b18e88c STATEMENT: CREATE UNIQUE INDEX "local_video_viewer_url" ON "localVideoViewer" ("url")
Mar 24 07:37:21 2025-03-24 00:37:21.633 UTC [1763] FATAL: terminating background worker "parallel worker" due to administrator command
Mar 24 07:37:21 2025-03-24 00:37:21.633 UTC [1763] STATEMENT: CREATE UNIQUE INDEX "local_video_viewer_url" ON "localVideoViewer" ("url")
Mar 24 07:37:23 2025-03-24 00:37:23.380 UTC [47] LOG: background worker "parallel worker" (PID 1763) exited with exit code 1

I tried the FLUSHALL command from cloudron terminal, and it seems to be back to normal. I will monitor whether it will go up again.

yes. is busy peertube.

here is the log and while postgresql sql memory usage is very small

redis log

Mar 20 23:59:51 13:M 20 Mar 2025 16:59:51.327 * RDB memory usage when created 15005.03 Mb
Mar 21 00:00:23 Redis service endpoint listening on http://:::3000
Mar 21 00:06:22 13:M 20 Mar 2025 17:06:22.756 * Done loading RDB, keys loaded: 33349358, keys expired: 998.
Mar 21 00:06:22 13:M 20 Mar 2025 17:06:22.756 * DB loaded from disk: 391.667 seconds

@girish yes, thank you...

@girish at the moment it looks like the postgres restore process is complete. and in the app log there are no running tasks.