Submitting a job

Graph and Search create and restore calls accept ?async=true. The server returns HTTP 202 with a job ID before the operation finishes. Without this parameter, the request waits for the operation result.

For example, submit a full Graph backup:

curl -s -X POST -H "Authorization: ${TC_OIDC_TOKEN}" \
  'https://telicent.core.domain/api/sparql/backups/create/?async=true'

Or submit a Search backup of knowledge:

curl -s -X POST -H "Authorization: ${TC_OIDC_TOKEN}" \
  -H "Content-Type: application/json" \
  'https://telicent.core.domain/api/search/backups/create?async=true' \
  -d '{"indexName":"knowledge","repositoryName":"scs-snapshots-repo"}'

To restore asynchronously, append ?async=true to the restore URL and use the same backup ID or JSON request body as in the Graph or Search restore guide.

Monitoring the result

Graph submission responses use job-id and status-path; Search uses jobId and statusPath. Both return an operation, status and message. Submission confirms acceptance, not successful completion.

Replace JOB_ID below with the returned ID:

curl -s -H "Authorization: ${TC_OIDC_TOKEN}" \
  https://telicent.core.domain/api/sparql/backups/jobs/JOB_ID

curl -s -H "Authorization: ${TC_OIDC_TOKEN}" \
  https://telicent.core.domain/api/search/backups/jobs/JOB_ID

Omit /JOB_ID to list jobs known to that server instance. The status path in a submission response is service-relative: use /$/backups/jobs/... directly with Graph, or /backups/jobs/... directly with Search. Through the gateway, use the corresponding /api/sparql/backups/jobs/... or /api/search/backups/jobs/... URL above.

Jobs progress through PENDING and RUNNING to SUCCEEDED or FAILED. Poll until a terminal status is returned, then inspect result, including the operation’s success fields and any per-dataset failures. In Graph, a successful HTTP status can still carry unsuccessful component results, so the job status alone is insufficient to establish a complete restore or backup.

Job history is held in memory by the service instance and is lost on restart. In a deployment with multiple instances, direct polling to the instance that accepted the job.