v1 - Githubissues

mikejgray commented 1 week ago

The initial commit is so that I can pull this down and test on arbitrary VMs. Once it's in a distributable and usable form, I'll close this issue. 🙂

NeonDaniel commented 1 week ago

Some minor readme feedback:

I'd bump the disk requirement to 100GB minimum and 256 or 512GB recommended for round numbers and so we have the option for pulling a bunch more models/images in the future without worrying about disk recommendations
If libretranslate isn't working, we could drop it until we have a suitable replacement ready
I think the hosts file changes on client images will work, though it helps to document for anyone who might want to access the websat from an arbitrary browser
Somewhere near the top, it should tell you how to run the installer

NeonDaniel commented 1 week ago

Feedback from running the installer:

git needs to be installed to clone the repository (details.. we could distribute the script differently; previous Neon installers had a setup script as a simple download that managed pulling things + git package installation)
Neon.AI should be replaced with: Neon AI®
Consider writing script stdout/stderr to a log file that may be shared for troubleshooting
URL listed on the success page doesn't appear to point to anything (nothing listening on ports 80 or 443)
Consider a default password for the neon user, or prompt in the installer to set one if the installing user isn't "neon"
Consider printing the default password for Yacht (I see there's already a related TODO for this in the readme)

I seem to have ended up with a non-working install..

docker logs -f neon-iris:

Connecting to neon-rabbitmq:5672
Traceback (most recent call last):
  File "/usr/local/bin/iris", line 8, in <module>
    sys.exit(neon_iris_cli())
  File "/usr/local/lib/python3.8/site-packages/click/core.py", line 1157, in __call__
    return self.main(*args, **kwargs)
  File "/usr/local/lib/python3.8/site-packages/click/core.py", line 1078, in main
    rv = self.invoke(ctx)
  File "/usr/local/lib/python3.8/site-packages/click/core.py", line 1688, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/usr/local/lib/python3.8/site-packages/click/core.py", line 1434, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/usr/local/lib/python3.8/site-packages/click/core.py", line 783, in invoke
    return __callback(*args, **kwargs)
  File "/usr/local/lib/python3.8/site-packages/neon_iris/cli.py", line 135, in start_gradio
    chat = GradIOClient()
  File "/usr/local/lib/python3.8/site-packages/neon_iris/web_client.py", line 51, in __init__
    NeonAIClient.__init__(self, config.get("MQ"))
  File "/usr/local/lib/python3.8/site-packages/neon_iris/client.py", line 63, in __init__
    self._connection = self._init_mq_connection()
  File "/usr/local/lib/python3.8/site-packages/neon_iris/client.py", line 384, in _init_mq_connection
    mq_connection = NeonMQHandler(mq_config, "mq_handler", self._vhost)
  File "/usr/local/lib/python3.8/site-packages/neon_utils/mq_utils.py", line 66, in __init__
    self.connection = pika.BlockingConnection(
  File "/usr/local/lib/python3.8/site-packages/pika/adapters/blocking_connection.py", line 360, in __init__
    self._impl = self._create_connection(parameters, _impl_class)
  File "/usr/local/lib/python3.8/site-packages/pika/adapters/blocking_connection.py", line 451, in _create_connection
    raise self._reap_last_connection_workflow_error(error)
pika.exceptions.AMQPConnectionError

NeonDaniel commented 1 week ago

RMQ docker logs:

=INFO REPORT==== 15-Oct-2024::15:50:29.976655 ===
    alarm_handler: {set,{system_memory_high_watermark,[]}}
2024-10-15 15:50:30.690982+00:00 [error] <0.253.0> Failed to update enabled plugins file "/etc/rabbitmq/enabled_plugins" from $RABBITMQ_ENABLED_PLUGINS: permission denied

BOOT FAILED
===========
Error during startup: {error,failed_to_update_enabled_plugins_file}

2024-10-15 15:50:30.702343+00:00 [error] <0.253.0>
2024-10-15 15:50:30.702343+00:00 [error] <0.253.0> BOOT FAILED
2024-10-15 15:50:30.702343+00:00 [error] <0.253.0> ===========
2024-10-15 15:50:30.702343+00:00 [error] <0.253.0> Error during startup: {error,failed_to_update_enabled_plugins_file}
2024-10-15 15:50:30.702343+00:00 [error] <0.253.0>
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>   crasher:
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>     initial call: application_master:init/4
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>     pid: <0.252.0>
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>     registered_name: []
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>     exception exit: {failed_to_update_enabled_plugins_file,
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>                         {rabbit,start,[normal,[]]}}
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>       in function  application_master:init/4 (application_master.erl, line 142)
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>     ancestors: [<0.251.0>]
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>     message_queue_len: 1
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>     messages: [{'EXIT',<0.253.0>,normal}]
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>     links: [<0.251.0>,<0.44.0>]
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>     dictionary: []
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>     trap_exit: true
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>     status: running
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>     heap_size: 376
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>     stack_size: 28
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>     reductions: 167
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>   neighbours:
2024-10-15 15:50:31.703603+00:00 [error] <0.252.0>
2024-10-15 15:50:31.706181+00:00 [notice] <0.44.0> Application rabbit exited with reason: {failed_to_update_enabled_plugins_file,{rabbit,start,[normal,[]]}}
Runtime terminating during boot (terminating)

Crash dump is being written to: erl_crash.dump...[os_mon] cpu supervisor port (cpu_sup): Erlang has closed
[os_mon] memory supervisor port (memsup): Erlang has closed

Are the .yml files in neon-hub-installer/ansible referenced by containers? This could be a permissions issue as those are in /home/d_mcknight while the service configs are in /home/neon

NeonDaniel commented 1 week ago

For service credentials (Wolfram Alpha, OpenWeatherMap, etc.), consider a couple pages in the installer to setup/skip, OR a post-install web UI (probably overlaps or is the same as the configuration management UI)

mikejgray commented 1 week ago

For service credentials (Wolfram Alpha, OpenWeatherMap, etc.), consider a couple pages in the installer to setup/skip, OR a post-install web UI (probably overlaps or is the same as the configuration management UI)

I figured for phase 1 we can just give instructions on how to do it in neon.yaml. Phase 2 we can make it configurable in the configuration management UI

mikejgray commented 1 week ago

Most feedback addressed in https://github.com/OscillateLabsLLC/neon-hub-installer/pull/4/files

NeonDaniel commented 1 week ago

I figured for phase 1 we can just give instructions on how to do it in neon.yaml. Phase 2 we can make it configurable in the configuration management UI

Just to clarify, I think in the SoW document we had the installer as Phase 1.A and the management UI as Phase 1.B. I'm satisfied with instructions for the first release of this installer though and the management UI to follow

mikejgray commented 6 days ago

Technically this is v0.0.3, but we're ready to close this out 🙂

NeonGeckoCom / neon-hub-installer

v1 #1