Problem: submission stuck in “Metadata Validating”
Description:
- wranglers have encountered a few cases of submission being stuck in the “metadata validating” state for hours blocking them to export the submission.
Steps to troubleshoot:
- Find which entities are in validating
- in ingest UI, filter each of the biomaterials, processes, protocols and data tabs by “validating”
- this should pin down which specific entities are stuck
- in this case, one or more file entities are stuck
- Check the file validation job status
- Click the “ingest api url” in Ingest UI
- the
validationJob
property will tell if the file validation has completed and successful or not - If the file is validating it is likley that the
jobCompleted: false
- Check the AWS batch job for the file
- check the queue
dcp-upload-validation-q-<env>
- use the Advanced filters to locate the job:
- Filter type
Job name
, Filter valuevalidation-<env>-<sub-uuid>-<validationId>
- for e.g.
validation-prod-9b984263-527f-4fcd-a5a1-c59cd1b91aa3-d8ffa2a0-8747-40cd-8c8a-ff83c3eaa969
- check the job validation log for any issue
- check the queue
If the AWS Batch Job Succeeded
- The file is validating
- The AWS batch job succeeded
- The
validationJob
status is not updated
This suggests that the upload service call to ingest is failing / ingest is not receiving and processing the status and set the document to Valid. Since the file validation succeeded, it should be safe to manually set the file to Valid to unblock the submission
curl -X PUT -H "Authorization: Bearer $TOKEN" https://api.ingest.archive.data.humancellatlas.org/files/<fileDocumentId>/validEvent
If the AWS Batch Job Failed
Use this script to retrigger File Validation.
Further investigation needed
- to understand what is happening
- places to look: ingest-validator, ingest-state-tracker and ingest-core logs
- upload service - ingest communication
- see state tracker README for the submission state diagram