Unsolved Can't set up a Directus instance at all

[shrey]

Just installed a fresh Directus app (2nd one on this server) and then tried logging in with the provided First-Time info.

But, getting an error:
"Invalid Credentials"

Haven't made any customisations, anywhere, yet.

[girish]

@shrey Is this with or without Cloudron user management ? Do you see any error in the logs?

[girish]

Just to double check, be sure to have 'Default' in the login dropdown. I checked with and without Cloudron user management and things seem to work.

[girish]

Another thing: can you go to Services -> Postgresql -> Increase memory limit. And then re-install directus? Maybe the migration is not completing because Postgresql is restarting or something.

[shrey]

@girish

When i visit the Directus admin page for the first time, i do notice this error in the console:

POST https://my-domain/auth/refresh 400`

{"errors":[{"message":"\"refresh_token\" is required in either the JSON payload or Cookie","extensions":{"code":"INVALID_PAYLOAD"}}]}

[shrey]

@girish i don't have this option, Default/LDAP.

[shrey]

@girish
increased it from 2GB to 4GB

Still getting the same error when trying to login: Invalid Credentials

[shrey]

@girish this is the only error that i see in the server logs, right after installation:

Feb 21 16:39:28 ==> Enabling postgis extension
Feb 21 16:39:29 CREATE EXTENSION
Feb 21 16:39:29 ==> New install, initializing
Feb 21 16:39:32 Killed
Feb 21 16:39:32 npm notice
Feb 21 16:39:32 npm notice New major version of npm available! 8.19.2 -> 9.5.0
Feb 21 16:39:32 npm notice Changelog: https://github.com/npm/cli/releases/tag/v9.5.0
Feb 21 16:39:32 npm notice Run npm install -g npm@9.5.0 to update!
Feb 21 16:39:32 npm notice
Feb 21 16:39:32
Feb 21 16:39:33 ==> Enabling postgis extension
Feb 21 16:39:33 NOTICE: extension "postgis" already exists, skipping
Feb 21 16:39:33 CREATE EXTENSION
Feb 21 16:39:33 ==> New install, initializing
Feb 21 16:39:36 [11:09:35.848] INFO: Initializing bootstrap...
Feb 21 16:39:36 [11:09:35.898] INFO: Database already initialized, skipping install
Feb 21 16:39:36 [11:09:35.898] INFO: Running migrations...
Feb 21 16:39:36 [11:09:35.906] INFO: Done
Feb 21 16:39:37 ==> DB Migration
Healtheck error: Error: connect ECONNREFUSED 172.18.18.24:80552023-02-21T11:09:40.000Z [11:09:40.218] INFO: Running migrations...
Feb 21 16:39:40 [11:09:40.241] INFO: Database up to date
Feb 21 16:39:41 ==> Starting Directus
Feb 21 16:39:45 [11:09:44.403] INFO: Server started at http://0.0.0.0:8055
Feb 21 16:39:51 [11:09:50] GET /server/ping 200 20ms
Feb 21 16:39:13 Starting supervisor

[girish]

@shrey said in Can't set up a Directus instance at all:

Feb 21 16:39:29 ==> New install, initializing
Feb 21 16:39:32 Killed

Right, it's running out of memory and thus the migration is not completing. I think you will see some oom message in dmesg output. Can you check how much memory the server has in free -m output?

[girish]

I see directus has basically no specific memory limit set. Maybe we can bump this...

[shrey]

@girish where do i run this command free -m ?

[girish]

@shrey sorry, you have to ssh into the server and check.

[shrey]

@girish

[girish]

I will link the other thread here - https://forum.cloudron.io/topic/8697/unable-to-connect-to-a-digitalocean-managed-database/9 . Cloudron is not designed to use an external database and you will hit all sorts of random issues. Sorry about that, maybe we should make this clear somewhere.

[nebulon]

The latest package has a higher default memory limit and is more reliable now during initial setup. So please try to install a fresh instance.