[Bug]: Wrong command for backup restore in documentation

[cencio91]

Error Message and Logs

Hi,

i'm trying disaster recovery procedure, and i notice that the official guide at
documentation

say to restore backup by using the command:

docker exec -it coolify-db pg_restore --verbose --clean --no-acl --no-owner -U coolify -d coolify < /path/to/your_backup_file

But it gives the following error (on ubuntu):

the input device is not a TTY

This way should works instead:

cat /path/to/your_backup_file | docker exec -i coolify-db pg_restore --verbose --clean --no-acl --no-owner -U coolify -d coolify

If you confirm this, i think that documentation should be updated.

Also, as mentioned in #5314 there is a bug after restoring the instance

If a manual backup of the server keys was necessary, it should be written in the documentation

P.S: I've tryed both using the APP_PREVIOUS_KEYS env variable, than replacing entire .env, as suggested here coollabsio/coolify#3761

Thank you for your work,
have a nice day

Coolify Version

I'm using 4.418

Are you using Coolify Cloud?

No (self-hosted)

Operating System and Version (self-hosted)

ubuntu 24.04 lts

Additional Information

No response

[ShadowArcanist]

Hey @cencio91 thanks for letting us know about this issue!

Could you please explain what exactly you mean by "If a manual backup of the server keys was necessary, it should be written in the documentation" ?
The doc already explains to backup the APP_KEY on step 2:

[marwie]

Hello, I'm currently having the same issue - after restoring the backup the server is unreachable.

I've also had the the input device is not a TTY error which was fixed with the command provided in the issue description.

Just in case it's relevant here's some logs during restoring:

pg_restore: while PROCESSING TOC:
pg_restore: from TOC entry 3997; 2606 16508 CONSTRAINT servers servers_pkey coolify
pg_restore: error: could not execute query: ERROR:  cannot drop constraint servers_pkey on table public.servers because other objects depend on it
DETAIL:  constraint ssl_certificates_server_id_foreign on table public.ssl_certificates depends on index public.servers_pkey
HINT:  Use DROP ... CASCADE to drop the dependent objects too.
Command was: ALTER TABLE ONLY public.servers DROP CONSTRAINT servers_pkey;
pg_restore: dropping CONSTRAINT server_settings server_settings_pkey
...
pg_restore: error: could not execute query: ERROR:  cannot drop table public.servers because other objects depend on it
DETAIL:  constraint ssl_certificates_server_id_foreign on table public.ssl_certificates depends on table public.servers
HINT:  Use DROP ... CASCADE to drop the dependent objects too.
Command was: DROP TABLE public.servers;
...
pg_restore: error: could not execute query: ERROR:  relation "servers" already exists
Command was: CREATE TABLE public.servers ( ...
...
pg_restore: from TOC entry 4386; 0 16499 TABLE DATA servers coolify
pg_restore: error: COPY failed for table "servers": ERROR:  duplicate key value violates unique constraint "servers_pkey"
DETAIL:  Key (id)=(0) already exists.
CONTEXT:  COPY servers, line 1

---

I've tried to use another SSH key which works at first but when I try to edit my applications and save I get other errors like the following:

SQLSTATE[42703]: Undefined column: 7 ERROR: column "is_http_basic_auth_enabled" of relation "applications" does not exist LINE 1: ...t "fqdn" = $1, "custom_nginx_configuration" = $2, "is_http_b... ^ (Connection: pgsql, SQL: update "applications" set "fqdn" = http://p0wkgk4.65.109.143.120.sslip.io, "custom_nginx_configuration" = , "is_http_basic_auth_enabled" = ?, "http_basic_auth_username" = ?, "http_basic_auth_password" = ?, "updated_at" = 2025-06-01 14:50:32 where "id" = 1)

I'm not sure if any of this is related to restoring a backup of coolify 391 to coolify 418 I've now updated both instances to 418 and still get the same error as above

[cencio91]

Hi @ShadowArcanist ,

APP_KEY backup is well documented, as you show in your screenshot.
With "server keys" I didn't mean the APP_KEY, but the server private/public key, as ex0ns sad in this comment coollabsio/coolify#3761 (comment)

The issue is that, after completing the restore procedure as shown in the guide, coolify can't restore connection with server, and i

I can see all configurations and projects, but the connection to server is not working.

NOTE: I'm restoring coolify in a different server. The issue is probably related to this, but I think it's a common scenario during a disaster recovery or migration.

Thank you very much,
Francesco

[ShadowArcanist]

Thank you all for the feedback and discussion here! Based on your reports, I’ve updated the Backup & Restore guide to address:

1. The “input device is not a TTY” error (using the cat … | docker exec -i … approach instead of -it).
2. Explicitly backing up and restoring Coolify’s SSH key files under /data/coolify/ssh/keys so that servers remain reachable.
3. Added instructions to pin the installation script to the same Coolify version (e.g. bash -s 4.0.0-beta.400) to avoid schema mismatches and “column does not exist” errors.

These changes will go live in the public documentation in next few days. Thanks again for helping make the procedure clearer and more reliable!

[cencio91]

Hi @ShadowArcanist and thank you for your work.
Just to clarify: I've not tested if backing up the /data/coolify/ssh/keys is enough to restore coolify server connection.
I can try in next days if needed. If you have tested it yourself instead, the issue can be closed

Best regards,
Francesco

[cencio91]

Hi @ShadowArcanist , I've tested the new backup/restore procedure, but i still get the same error (permission denied)

P.S:
According to new documentation, During restoration, step 2 says to stop all coolify containers..
but a step 2.1 is needed to restart the database container

docker stop $(docker ps -q --filter "name=coolify")
docker start coolify-db

otherwise, (the DB restore command (step 3) cannot be executed

[ShadowArcanist]

I will take a closer look in a day or two