Upgrading nightly version inside docker

[Fogapod]

I wanted to upgrade my nightly docker container because there were no releases for 2 days, a few commits happened till then.
I ran into a few issues:

• status command

# edgedb server status
Error: No instance "default" found. Run:
  edgedb server init default

I have a running database, not sure what is wrong here.

• upgrade --nightly

# edgedb server upgrade --nightly
[2020-07-27T15:02:38Z WARN  edgedb::server::upgrade] No instances found. Nothing to upgrade.

• server init default

# edgedb server init default
┌──────────────────────┬────────────────────────────────────────┐
│ Instance Name        │ default                                │
│ Mode                 │ User Service                           │
│ Data Directory       │ /root/.local/share/edgedb/data/default │
│ Database Server Port │ 5656                                   │
│ EdgeDB Version       │ 1-alpha4                               │
└──────────────────────┴────────────────────────────────────────┘
=================================================== Exception occurred: pg_ctl init exited with status 1:
pg_ctl: cannot be run as root
Please log in (using, e.g., "su") as the (unprivileged) user that will
own the server process.
 ====================================================

1. edb.server.pgcluster.ClusterError: pg_ctl init exited with status 1:
pg_ctl: cannot be run as root
Please log in (using, e.g., "su") as the (unprivileged) user that will
own the server process.


---------------------------------------------------------------------------------------------------------------------------------------- Traceback -----------------------------------------------------------------------------------------------------------------------------------------

    /usr/bin/edgedb-server-1-alpha4, line 10, in <module>
        > sys.exit(main())
    /usr/lib/x86_64-linux-gnu/edgedb-server-1-alpha4/lib/python3.8/site-packages/click/core.py, line 722, in __call__
        > return self.main(*args, **kwargs)
    /usr/lib/x86_64-linux-gnu/edgedb-server-1-alpha4/lib/python3.8/site-packages/click/core.py, line 697, in main
        > rv = self.invoke(ctx)
    /usr/lib/x86_64-linux-gnu/edgedb-server-1-alpha4/lib/python3.8/site-packages/click/core.py, line 895, in invoke
        > return ctx.invoke(self.callback, **ctx.params)
    /usr/lib/x86_64-linux-gnu/edgedb-server-1-alpha4/lib/python3.8/site-packages/click/core.py, line 535, in invoke
        > return callback(*args, **kwargs)
    /usr/lib/x86_64-linux-gnu/edgedb-server-1-alpha4/lib/python3.8/site-packages/edb/server/main.py, line 597, in main
        > server_main(**kwargs)
    /usr/lib/x86_64-linux-gnu/edgedb-server-1-alpha4/lib/python3.8/site-packages/edb/server/main.py, line 586, in server_main
        > run_server(ServerConfig(**kwargs))
    /usr/lib/x86_64-linux-gnu/edgedb-server-1-alpha4/lib/python3.8/site-packages/edb/server/main.py, line 256, in run_server
        > pg_cluster_init_by_us = cluster.ensure_initialized()
    /usr/lib/x86_64-linux-gnu/edgedb-server-1-alpha4/lib/python3.8/site-packages/edb/server/pgcluster.py, line 327, in ensure_initialized
        > initdb_output = self.init(
    /usr/lib/x86_64-linux-gnu/edgedb-server-1-alpha4/lib/python3.8/site-packages/edb/server/pgcluster.py, line 374, in init
        > raise ClusterError(

edb.server.pgcluster.ClusterError: pg_ctl init exited with status 1:
pg_ctl: cannot be run as root
Please log in (using, e.g., "su") as the (unprivileged) user that will
own the server process.

[2020-07-27T15:04:01Z ERROR edgedb::server::init] Bootstrap error, cleaning up...
Error: Error bootstrapping /root/.local/share/edgedb/data/default: Command "/usr/bin/edgedb-server-1-alpha4" "--bootstrap" "--log-level=warn" "--data-dir" "/root/.local/share/edgedb/data/default" exit code: 1

Note: I have a running server already inside container

Is what I am doing correct way of upgrading or these commands should not be used with docker version?
I couldn't find any documentation on this.

[elprans]

The intended way is to have edgedb server manage the docker container for you, while running on the host (via --method=docker). Support for that hasn't landed yet, but is in progress (geldata/gel-cli#67, geldata/gel-cli#72).

Meanwhile, for manually-managed docker containers you have to do the upgrade manually:

1. do edgedb dump on the old version,
2. pull the new image and run it with a new data directory,
3. create the database in the new instance,
4. restore the database content using the previously created dump with edgedb restore.

[Fogapod]

My problem was outdated docker image. So server commands are for migrating data?

[elprans]

edgedb server commands are intended to install, update, start, stop, and upgrade server instances. edgedb server upgrade will migrate data automatically to the new version. It is supposed to be running on the host, it will not work inside the EdgeDB docker image. Like I said, the Docker installation method isn't supported by edgedb server yet, we're working on that.

[Fogapod]

If I need a newer docker image, should I clone and use https://github.com/edgedb/edgedb-pkg ?
Or building my own image with dev environment for edgedb will be easier?

[elprans]

Which image is outdated? Last night's nightly contains the latest commit:

$ docker run -it --rm edgedb/edgedb:nightly edgedb-server --version
edgedb-server, version 1.0-alpha.4+dev.98.g4ab945ae

If I need a newer docker image, should I clone and use https://github.com/edgedb/edgedb-pkg ?

Yes, you can use edgedb-pkg to build native Debian or RHEL packages for any version of EdgeDB.

[Fogapod]

Oh, I pulled the wrong tag and didn't notice recent image on docker hub somehow.
Thanks for the quick and detailed responses as always!