Closed pwFoo closed 5 years ago
I just tested again both MySQL 0.15.10 -> 0.15.11-rc3 and RethinkDB 0.15.10 -> 0.15.11-rc3 and both worked as expected. Both correctly dropped and recreated the databases and then correctly launched the server.
If you are still seeing this issue please describe steps to reproduce.
Same problem update rc2 to rc4. Container stops after startet.
$ docker logs tinode
2018/12/30 17:40:27 Database exists, DB version is correct. All done. mysql
2018/12/30 17:40:49 Database exists, DB version is correct. All done. mysql
Looks like a bug with rc3 / rc4...
Please give me exact steps to reproduce.
It stops after ~ 6 seconds
Please list the entire setup: database name and version, steps you used to start the containers.
I tried to reproduce from 0.15.10 to 0.15.11-rc3 and it worked fine.
Thanks. You probably don't want to publish your password here. I removed it but it was still available in history. So I deleted the entire message. Here is your message without the password etc:
docker run --name tinode-mysql --env MYSQL_ALLOW_EMPTY_PASSWORD=yes -d mysql:5.7
docker run -d --name tinode -e VIRTUAL_HOST=<DOMAIN -e VIRTUAL_PORT=18080 -e CADDY_PROXY_OPTS=websocket --link tinode-mysql:mysql -e SMTP_SERVER=<SERVER> -e SMTP_PORT=25 -e SMTP_SENDER=<ADDY> -e SMTP_PASSWORD=<PW> --env FCM_CRED_FILE=/fcm/<FILE> -v tinode-fcm:/fcm --env FCM_SENDER_ID=<ID> -v tinode-logs:/var/log/tinode -v tinode-botdata:/botdata tinode/tinode-mysql:0.15.11-rc2
Reproduce:
docker rm -f tinode && docker run -d --name tinode -e VIRTUAL_HOST=<HOST> -e VIRTUAL_PORT=18080 -e CADDY_PROXY_OPTS=websocket --link tinode-mysql:mysql -e SMTP_SERVER=<SMTP_SERVER> -e SMTP_PORT=25 -e SMTP_SENDER=<SENDER> -e SMTP_PASSWORD=<REMOVED> --env FCM_CRED_FILE=/fcm/tinode-1586b-firebase-adminsdk-vima7-373057298f.json -v tinode-fcm:/fcm --env FCM_SENDER_ID=<REMOVED> -v tinode-logs:/var/log/tinode -v tinode-botdata:/botdata tinode/tinode-mysql:0.15.11-rc4
No info in logs. It fails from 0.15.10 or 0.15.11.rc2 to rc3 / rc4
I still cannot reproduce:
genes-mbp:chat gene$ docker run --name tinode-mysql --env MYSQL_ALLOW_EMPTY_PASSWORD=yes -d mysql:5.7
223f340b8c78d62d8a4410e01b45c52811f1567e7964caa8b4ed2023c8f9064c
genes-mbp:chat gene$ docker run -d --name tinode --link tinode-mysql:mysql -v tinode-logs:/var/log/tinode -v tinode-botdata:/botdata tinode/tinode-mysql:0.15.10
cc59eb019b09daa7361e5eb490bc4d437dca1c876cf6027a04bf354c8c24f174
genes-mbp:chat gene$ docker rm -f tinode
tinode
genes-mbp:chat gene$ docker run -d --name tinode -e VIRTUAL_PORT=18080 -e CADDY_PROXY_OPTS=websocket --link tinode-mysql:mysql -v tinode-logs:/var/log/tinode -v tinode-botdata:/botdata tinode/tinode-mysql:0.15.11-rc4
bb321a5e68ca0fe4cbf93d77c32052edb2c6b6f18e04a802150e9936908aefe5
genes-mbp:chat gene$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
bb321a5e68ca tinode/tinode-mysql:0.15.11-rc4 "/bin/sh -c ./entryp…" 3 minutes ago Up 3 minutes 16061/tcp, 18080/tcp tinode
223f340b8c78 mysql:5.7 "docker-entrypoint.s…" 4 minutes ago Up 4 minutes 3306/tcp, 33060/tcp tinode-mysql
You are using the following variables:
-e VIRTUAL_HOST=<DOMAIN> -e VIRTUAL_PORT=18080 -e CADDY_PROXY_OPTS=websocket
Which means you are running a modified docker container. Is it possible that something is off in your docker?
That is just a reverse proxy in front of the container to configure the automatic reverse proxy. So image isn't modified. You removed the FCM part. Could you test it with FCM?
I missed to remove the password... have to change it now. Thanks!
It's SMTP related! Any changes with SMTP / email? Works with <= rc2, stops with rc3/rc4
Yes, there is a change related to SMTP: domains whitelist: https://github.com/tinode/chat/blob/devel/docker/tinode/config.template#L85
Now that you mentioned it I think I know why it's failing.
Fix released as https://github.com/tinode/chat/releases/tag/v0.15.11-rc5
Container ist running now. SMTP related Problem is fixes.
Because of upgrade / downgrade and DB schema incompatibility I need to register a new user.
Register fails with
auth secret is not unique policy
Docker Image related?
That error means you are trying to register a user with a duplicate login. It looks like the database has not been reset. Can you check if the database has been updated?
deleted and recreated container with RESET_DB=1
2019/01/03 13:03:23 Database reset requested
2019/01/03 13:03:25 Generating users...
2019/01/03 13:03:26 Generating group topics...
2019/01/03 13:03:27 Generating P2P subscriptions...
2019/01/03 13:03:27 Generating group subscriptions...
2019/01/03 13:03:27 Inserting messages...
2019/01/03 13:03:30 Inserting forms as tino fsw1rVhWfXM
2019/01/03 13:03:30 All done.
Same error if I try to register a new user
Error: policy violation (422)
auth DB uname
mysql> select uname from auth;
+-------------+
| uname |
+-------------+
| basic:alice |
| basic:bob |
| basic:carol |
| basic:dave |
| basic:eve |
| basic:frank |
| basic:tino |
| basic:xena |
+-------------+
What's the login/password of the user you are trying to register?
Two restrictions were added to basic auth: min_login_length
and min_password_length
:
https://github.com/tinode/chat/blob/devel/docker/tinode/config.template#L45
Maybe you are trying to use a login/password which is too short.
Ok... I use a 3 characters short username :( Tried other passwords, but not changed the username...
I suppose I can change it to 3.
Yeah, would be nice :) 👍
Changed to 3 in devel. I'll release it in the 0.15.11 in a couple of days.
Maybe known bug / problem because it's a pre-release.
Updated my docker container from 0.15.10 (stable release) to latest 0.15.11-rc3, but container stops after some seconds.
Restarted container and it stops again after some seconds.
So I downgraded to the earlier pre-release 0.15.11-rc2 and it looks better. After update to the new pre-releases the DB is emtpy as logged during first start with the pre-release container.
That is no problem because it's just a test server