[Question]: parsing file failed

[lilu6301]

Describe your problem

[ERROR]Insert chunk error, detail info please check ragflow-logs/api/cron_logger.log. Please also check ES

[mjtechguy]

Same. I am trying to ingest a .csv file using the latest docker-compose and Ollama (nomic-embed-text)

Any ideas?

[lilu6301]

my os is ubuntu 20.04. fixed by updated to 22.04

[PuppyMeng]

my os is ubuntu 20.04. fixed by updated to 22.04

I’m having the same problem, too. using ubuntu 24.04.1

[cfenglv]

same here. Running with local build docker image on Mac.
Looking into cron_logger.log gives:
{'type': 'document_parsing_exception', 'reason': '[1:90157] failed to parse: The [cosine] similarity does not support vectors with zero magnitude. Preview of invalid vector: [0.0, 0.0, 0.0, 0.0, 0.0, ...]', 'caused_by': {'type': 'illegal_argument_exception', 'reason': 'The [cosine] similarity does not support vectors with zero magnitude. Preview of invalid vector: [0.0, 0.0, 0.0, 0.0, 0.0, ...]'}}"

By the way, looking into database.log gives:
ERROR (21) Can't update token usage for bc2bb8c3862b11ef8466f966ff2a9065/CHAT

I am using local Ollama, chat model local llama3.2:1b, embedding model local Ollama "snowflake-arctic-embed"

[aalboori]

I had this error due to low storage space. Please check the status of ES through ragflow user setting-->system. If it is red, please follow the link https://www.elastic.co/guide/en/elasticsearch/reference/8.15/red-yellow-cluster-status.html#fix-red-yellow-cluster-status to diagnose the reason.

[EzeLLM]

g the same problem, too. using ubuntu 24.04.1

this solved my problem. the following command in particular, as it fits with my logs:

curl -X PUT "localhost:9200/_cluster/settings?pretty" -H 'Content-Type: application/json' -d'
{
  "persistent" : {
    "cluster.routing.allocation.enable" : null
  }
}
'

[WaterDrop-EarthDivision]

I'm happy to share my experience here.

The reason I have this is because of lack of space, strictly because elasticsearch monitors your storage usage and stops sharding when it's greater than 85 percent.

I will describe below how I discovered and solved this problem.

1. Find the username and password for elasticsearch by using docker logs -f ragflow-server.

'username': 'elastic', 'password': 'infini_rag_flow'

2. Diagnosis via link shared by @aalboori

2.1 Use curl -X GET “localhost:1200/_cluster/health?pretty” -u elastic:infini_rag_flow to view cluster status.

Use curl -X GET “localhost:1200/_cluster/health?pretty” -u elastic:infini_rag_flow to view cluster status. Note that in the original code it was “localhost:7200”. The -u stands for the login and password.

2.2 Use curl -X GET “localhost:1200/_cat/allocation?v=true&h=node,shards,disk.*&pretty” -u elastic:infini_rag_flow to view usage.And then I realized that it was using up to 94% of the memory.Then I realized that a whopping 94% of the memory was being used. But I have almost 2T of RAM.

3. Replace using the specified memory instead of a percentage.

curl -X PUT "localhost:1200/_cluster/settings?pretty" -H 'Content-Type: application/json' -u elastic:infini_rag_flow -d'
{
  "persistent": {
    "cluster.routing.allocation.disk.watermark.low": "50gb",
    "cluster.routing.allocation.disk.watermark.high": "30gb",
    "cluster.routing.allocation.disk.watermark.flood_stage": "20gb"
  }
}'

You can't change just one alone. Must all three be percentages or memory capacity ! ! !

I hope my experience helps you.