Sync failure: cursor id xxxxxxxxxx not found
- Dominant language
- Python
- Stars
- 133
- Forks
- 205
- Avg merge
- 16h 3m
- Merged PRs (30d)
- 102
Description
When running MongoDB connector with a ML training model (e.g. ELSER), the added time spent on running the data against the training model appears to be causing the MongoDB cursor to timeout:
This results in a failure to sync with a singular log entry showing a critical error:
If I add sync rules to limit the amount of data that gets processed by the training model, the job runs fine then.
We are looking into improving observability in this area via https://github.com/elastic/connectors-python/issues/961. Just filing in case others are running into this. If this ends up being a cursor timeout issue, it can be helpful to remove/extend the cursor timeout.
Contributor guide
Research direction
Start by reviewing the MongoDB connector sync path and issue #961, which covers observability for this area. Reproduce the failure with an ML training model and compare it with a sync constrained by rules. Done should establish whether cursor timeout is the cause and define the resulting handling or observability change.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- mongodb, python
- Domain
- backend, databases
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100