Immich•2y ago

Clean install of 1.60, no object/face detection

I finally got my new hardware, specially for Immich 🙂 (with avx2 this time) Everything is up and running, except for the face recognition / object detection. The ml container doesn't log much, the microservices container logs the following:

immich_microservices     | [Nest] 1  - 06/06/2023, 7:06:35 AM   ERROR [JobService] Unable to run job handler: Error: connect EADDRNOTAVAIL ::1:80 - Local (:::0)
immich_microservices     | [Nest] 1  - 06/06/2023, 7:06:35 AM   ERROR [JobService] Error: connect EADDRNOTAVAIL ::1:80 - Local (:::0)
immich_microservices     |     at internalConnect (node:net:1066:16)
immich_microservices     |     at defaultTriggerAsyncIdScope (node:internal/async_hooks:464:18)
immich_microservices     |     at node:net:1250:9
immich_microservices     |     at process.processTicksAndRejections (node:internal/process/task_queues:77:11)
immich_microservices     | [Nest] 1  - 06/06/2023, 7:06:35 AM   ERROR [JobService] Object:
immich_microservices     | {
immich_microservices     |   "id": "2fd47f0b-1b15-491e-838e-429c616f43e4"
immich_microservices     | }
immich_microservices     |

immich_microservices     | [Nest] 1  - 06/06/2023, 7:06:35 AM   ERROR [JobService] Unable to run job handler: Error: connect EADDRNOTAVAIL ::1:80 - Local (:::0)
immich_microservices     | [Nest] 1  - 06/06/2023, 7:06:35 AM   ERROR [JobService] Error: connect EADDRNOTAVAIL ::1:80 - Local (:::0)
immich_microservices     |     at internalConnect (node:net:1066:16)
immich_microservices     |     at defaultTriggerAsyncIdScope (node:internal/async_hooks:464:18)
immich_microservices     |     at node:net:1250:9
immich_microservices     |     at process.processTicksAndRejections (node:internal/process/task_queues:77:11)
immich_microservices     | [Nest] 1  - 06/06/2023, 7:06:35 AM   ERROR [JobService] Object:
immich_microservices     | {
immich_microservices     |   "id": "2fd47f0b-1b15-491e-838e-429c616f43e4"
immich_microservices     | }
immich_microservices     |

The ML container log is here: https://pastebin.com/raw/WAKyDysS Any idea what is preventing the object detection? I restarted the job, it get queued and then is gone. Thanks in advance!

10 Replies

bo0tzz•2y ago

Can you post your .env and docker-compose.yml?

JohanOP•2y ago

My .env is:

DB_DATABASE_NAME=immich
DB_HOSTNAME=immich_postgres
DB_PASSWORD=[....]
DB_USERNAME=[....]
DISABLE_REVERSE_GEOCODING=true
IMMICH_MACHINE_LEARNING_URL=true
IMMICH_SERVER_URL=http://immich-server:3001
IMMICH_WEB_URL=http://immich-web:3000
PUBLIC_LOGIN_PAGE_MESSAGE="Welkom!"
REDIS_HOSTNAME=immich_redis
TYPESENSE_API_KEY=some-random-text
TYPESENSE_ENABLED=true
UPLOAD_LOCATION=/data/photos/immich/

DB_DATABASE_NAME=immich
DB_HOSTNAME=immich_postgres
DB_PASSWORD=[....]
DB_USERNAME=[....]
DISABLE_REVERSE_GEOCODING=true
IMMICH_MACHINE_LEARNING_URL=true
IMMICH_SERVER_URL=http://immich-server:3001
IMMICH_WEB_URL=http://immich-web:3000
PUBLIC_LOGIN_PAGE_MESSAGE="Welkom!"
REDIS_HOSTNAME=immich_redis
TYPESENSE_API_KEY=some-random-text
TYPESENSE_ENABLED=true
UPLOAD_LOCATION=/data/photos/immich/

The compose is exact copy of the release one (I pull that with every update), with the addition of:

networks:
  default:
    external: true
    name: nginx

networks:
  default:
    external: true
    name: nginx

bo0tzz•2y ago

IMMICH_MACHINE_LEARNING_URL=true won't work ;) you probably just want to leave that default

JohanOP•2y ago

Ah! So, comment/remove it?

bo0tzz•2y ago

Probably set it to whatever the default .env uses

JohanOP•2y ago

Man, I feel stupid 🙈 URL !=boolean Hmm, after adding the default user to the variable, doing a compose restart, it gives a 502. The log shows

immich_proxy             | 2023/06/06 08:03:17 [error] 42#42: *105 connect() failed (111: Connection refused) while connecting to upstream, client: 172.20.0.1, server: , request: "GET /favicon.ico HTTP/1.1", upstream: "http://172.20.0.6:3000/favicon.ico", host: "immich.2411.duckdns.org", referrer: "https://<myurl>/"

immich_proxy             | 2023/06/06 08:03:17 [error] 42#42: *105 connect() failed (111: Connection refused) while connecting to upstream, client: 172.20.0.1, server: , request: "GET /favicon.ico HTTP/1.1", upstream: "http://172.20.0.6:3000/favicon.ico", host: "immich.2411.duckdns.org", referrer: "https://<myurl>/"

Used the example.env, with my DB user/pw and an custom TYPESENSE_API_KEY, but still no object processing. The micro processing container still logs this:

immich_microservices  | [Nest] 1  - 06/06/2023, 8:33:16 AM   ERROR [JobService] Unable to run job handler: Error: connect EADDRNOTAVAIL ::1:80 - Local (:::0)
immich_microservices  | [Nest] 1  - 06/06/2023, 8:33:16 AM   ERROR [JobService] Error: connect EADDRNOTAVAIL ::1:80 - Local (:::0)
immich_microservices  |     at internalConnect (node:net:1066:16)
immich_microservices  |     at defaultTriggerAsyncIdScope (node:internal/async_hooks:464:18)
immich_microservices  |     at node:net:1250:9
immich_microservices  |     at process.processTicksAndRejections (node:internal/process/task_queues:77:11)
immich_microservices  | [Nest] 1  - 06/06/2023, 8:33:16 AM   ERROR [JobService] Object:
immich_microservices  | {
immich_microservices  |   "id": "2fd47f0b-1b15-491e-838e-429c616f43e4"
immich_microservices  | }

immich_microservices  | [Nest] 1  - 06/06/2023, 8:33:16 AM   ERROR [JobService] Unable to run job handler: Error: connect EADDRNOTAVAIL ::1:80 - Local (:::0)
immich_microservices  | [Nest] 1  - 06/06/2023, 8:33:16 AM   ERROR [JobService] Error: connect EADDRNOTAVAIL ::1:80 - Local (:::0)
immich_microservices  |     at internalConnect (node:net:1066:16)
immich_microservices  |     at defaultTriggerAsyncIdScope (node:internal/async_hooks:464:18)
immich_microservices  |     at node:net:1250:9
immich_microservices  |     at process.processTicksAndRejections (node:internal/process/task_queues:77:11)
immich_microservices  | [Nest] 1  - 06/06/2023, 8:33:16 AM   ERROR [JobService] Object:
immich_microservices  | {
immich_microservices  |   "id": "2fd47f0b-1b15-491e-838e-429c616f43e4"
immich_microservices  | }

Is that normal?

bo0tzz•2y ago

It's still using the wrong address for ML. did you restart everything after changing the env?

JohanOP•2y ago

I did via docker compose restart, or should I remove the containers and start new ones?

bo0tzz•2y ago

Restart just restarts in place, you need to use up

JohanOP•2y ago

Thanks @bo0tzz for helping! 💪 Its frying my CPU right now 🙂

Gaming

Programming

Clean install of 1.60, no object/face detection

Did you find this page helpful?