service gives 500 when installing on macOS, start seems to work but doesn't ask for permisions

[sammcj]

✓ Starting...
 ✓ Ready in 2.2s
(node:78173) [DEP0040] DeprecationWarning: The `punycode` module is deprecated. Please use a userland alternative instead.
(Use `node --trace-deprecation ...` to show where the warning was created)
 ○ Compiling / ...
 ✓ Compiled / in 5s (3527 modules)
 GET / 200 in 5441ms
 ✓ Compiled in 1009ms (1759 modules)
 ○ Compiling /api/tags ...
 ✓ Compiled /api/tags in 3s (3514 modules)
 GET / 200 in 601ms
 GET /favicon.ico 200 in 17ms
 GET /api/tags 200 in 4ms
✓ Compiled /api/install-service in 244ms (1792 modules)
 GET /api/install-service 500 in 30957ms

But - when I click 'start service' I do notice a new icon in the system tray, when I click it - it has two "quit" options and no other information:

It hasn't prompted me to allow it to screen record etc... so I don't think it's actually working.

[sammcj]

Ooo somehow I have lots of new tray icons now!

[sammcj]

and asking any questions with any of my ollama models seems to fail:

 ✓ Starting...
 ✓ Ready in 1170ms
 ○ Compiling / ...
 ✓ Compiled / in 2.2s (3518 modules)
 GET / 200 in 2595ms
 ✓ Compiled in 392ms (1759 modules)
 ○ Compiling /api/tags ...
 ✓ Compiled /api/tags in 715ms (3514 modules)
 GET /api/tags 200 in 443ms
 GET /api/tags 200 in 442ms
 GET /favicon.ico 200 in 1094ms
 GET / 200 in 28ms
 ✓ Compiled /api/chat in 334ms (134 modules)
Error: [TypeError: fetch failed] { cause: [Error: AggregateError] }
 POST /api/chat 500 in 408ms
 ✓ Compiled /api/start-service in 287ms (1796 modules)
Error: [TypeError: fetch failed] { cause: [Error: AggregateError] }
 POST /api/chat 500 in 20ms
 GET / 200 in 78ms
Error: [TypeError: fetch failed] { cause: [Error: AggregateError] }
 POST /api/chat 500 in 8ms
 GET / 200 in 21ms

[lucacri]

Same problems here:

 ✓ Starting...
 ✓ Ready in 5.4s
(node:65920) [DEP0040] DeprecationWarning: The `punycode` module is deprecated. Please use a userland alternative instead.
(Use `node --trace-deprecation ...` to show where the warning was created)
 ○ Compiling / ...
 ✓ Compiled / in 5.6s (3528 modules)
 GET / 200 in 6034ms
 ✓ Compiled in 532ms (1760 modules)
 ○ Compiling /api/tags ...
 ✓ Compiled /api/tags in 1955ms (1763 modules)
 GET /api/tags 200 in 1999ms
 GET /api/tags 200 in 1999ms
 ✓ Compiled /api/start-service in 309ms (1792 modules)
 GET / 200 in 226ms
 ✓ Compiled /api/chat in 438ms (134 modules)
Server response: {"message":"An error occurred while processing the query.","status":"error"}

Error: [Error: HTTP error! status: 500, message: {"message":"An error occurred while processing the query.","status":"error"}
]
 POST /api/chat 500 in 585ms

[DonTizi]

Hey @lucacri,

Can you run this command in the terminal when the services are on?

curl -X POST http://localhost:8005/query \
     -H "Content-Type: application/json" \
     -d '{"query": "hey!"}
Let me know what it tells you.

And @sammcj, could you go to /users/{your username}/library/Application support/RemindEnchanted and check if you have any screenshots in the screenshots folder? I need to confirm this first.

Also, the quit feature in the icon works for the service in the repo. I arranged it so that it closes everything when we click on it. i'll upload the repo soon. I will also make sure that when we click on "Start Service", if a service is already running, it doesn't start another one. Thanks for letting me know!

For now to close everything , go to the activity monitor , and close all pythons and remindOCR process.

[lucacri]

Hi! I just checked and indeed I do have some screenshots, but i turned it off after not long. If it helps, I have 4 monitors but the screenshot is only of the main one. Maybe for a future issue! Thank you, Luca

…

On Sep 5, 2024 at 1:19 PM -0400, DonTizi ***@***.***>, wrote: Hey @lucacri, Can you run this command in the terminal when the services are on? curl -X POST http://localhost:8005/query \ -H "Content-Type: application/json" \ -d '{"query": "hey!"} Let me know what it tells you. And @sammcj, could you go to /users/{your username}/library/Application support/RemindEnchanted and check if you have any screenshots in the screenshots folder? I need to confirm this first. Also, the quit feature in the icon works for the service in the repo. I arranged it so that it closes everything when we click on it. i'll upload the repo soon. I will also make sure that when we click on "Start Service", if a service is already running, it doesn't start another one. Thanks for letting me know! For now to close everything , go to the activity monitor , and close all pythons and remindOCR process. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[sammcj]

Just checked mine as well, yes there's screenshots but they're very low resolution 4:3 of a single monitor so I'm not sure if they'd actually be useful.

I had a look through the code and saw this: https://github.com/DonTizi/ReMind/blob/main/installer.py#L10 😬, I'd suggest moving to using standard launchd jobs to register, start and stop the background service so that it's managed in a standard, repeatable way in the same place as all other background services.

For example, my Ollama launchd job:

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
	<key>AssociatedBundleIdentifiers</key>
	<string>com.electron.ollama</string>
	<key>Disabled</key>
	<false/>
	<key>EnvironmentVariables</key>
	<dict>
		<key>OLLAMA_CACHE_TYPE_K</key>
		<string>q8_0</string>
		<key>OLLAMA_CACHE_TYPE_V</key>
		<string>q8_0</string>
		<key>OLLAMA_FLASH_ATTENTION</key>
		<string>1</string>
		<key>OLLAMA_KEEP_ALIVE</key>
		<string>6h</string>
		<key>OLLAMA_MAX_LOADED_MODELS</key>
		<string>3</string>
		<key>OLLAMA_ORIGINS</key>
		<string>'http://localhost:*,https://localhost:*,app://obsidian.md*,app://*'</string>
	</dict>
	<key>KeepAlive</key>
	<false/>
	<key>Label</key>
	<string>ollama.set.envs.on.boot</string>
	<key>ProgramArguments</key>
	<array>
		<string>/opt/homebrew/bin/bash</string>
		<string>-c</string>
		<string>(open -a /Applications/Ollama.app; sudo /usr/sbin/sysctl iogpu.wired_limit_mb=84000)</string>
	</array>
	<key>RunAtLoad</key>
	<true/>
	<key>WatchPaths</key>
	<array>
		<string>/Applications/Ollama.app</string>
	</array>
</dict>
</plist>

 curl -X POST http://localhost:8005/query -H "Content-Type: application/json" -d '{"query": "hey!"}
Let me know what it tells you.'
curl: (7) Failed to connect to localhost port 8005 after 0 ms: Couldn't connect to server

[kjarnot]

Having the same issue with /api/chat throwing Error: [TypeError: fetch failed] { cause: [Error: AggregateError] }.

@DonTizi - what is supposed to be listening on port 8005?

[Saidshire21]

y