CVAT installed on a server - Could not fetch models meta information

[DP1701]

Hello all,

we have a speed problem with CVAT when we have installed it on a server.

The following constellation:

• Docker is on a Linux server (Docker 20.10.21, Ubuntu 20.04)
• The virtual machine is assigned 8 cores and 32 GB Ram
• Everything is installed on an SSD
• Images (CVAT data files) are stored on a NAS with a 1 GBit connection

Login works fine but it takes a long time to load the projects or tasks. This message also appears frequently:

Could not fetch models meta information - Open the Browser Console to get details

Console output:

Does anyone have a tip on what the problem might be, or instructions on how best to install CVAT on a server?

[azhavoro]

please attach logs from cvat_server container: docker logs cvat_server > cvat.log

[DP1701]

@azhavoro
Here are the logs:

cvat_logs.txt

I had the message today 18.01. at 9:15 a.m. and yesterday at 12 p.m.

[DP1701]

@azhavoro Is there anything noted in the log file that indicates the problem?

[DP1701]

Maybe for info, which CVAT version we have installed on the server.

[azhavoro]

Hi, sorry for the late reply

This is the same issue #4720

As workarround in you case you can define new volume and just mount to /home/django/data/cache path like this:

services:
  cvat_server:
    volumes:
      - cvat_cache:/home/django/data/cache
  cvat_worker_export:
    volumes:
      - cvat_cache:/home/django/data/cache

  cvat_worker_annotation:
    volumes:
      - cvat_cache:/home/django/data/cache

volumes:
  cvat_cache:

[DP1701]

Thanks for the help. The server is working fine now.

[315386775]

@azhavoro i have some issue. My version :

Server version: 2.5
Core version: 9.1.0
Canvas version: 2.16.5
UI version: 1.51.0

My cvat.log

cvat.log

Startup comand:

docker-compose -f docker-compose.yml -f components/analytics/docker-compose.analytics.yml up -d --build

how to slove that. tks.