Hi <@U04S8PLJWMT> seeing this error in the server ...
# ask-ai
i
Hi @kapa.ai seeing this error in the server logs 2025-12-01T205646.590357331Z stdout F 2025-12-01 205646,590 [io-executor-thread-6] ESC[1;31mERRORESC[0;39m i.a.c.s.e.h.UncaughtExceptionHandler(handle):33 - Uncaught exception 2025-12-01T205646.590365878Z stdout F java.lang.IllegalStateException: Cannot create an attempt for a job id: 60048 that has a running attempt: RUNNING for connection id: 6a6a0c43-9ade-4ea0-8979-d28115ff5df3
k
That error means Airbyte’s job engine is trying to start a new attempt on a sync job (job id
60048
) while there is already an attempt in
RUNNING
status for the same connection (
6a6a0c43-9ade-4ea0-8979-d28115ff5df3
). Airbyte enforces at most one non‑terminal job per connection at a time via the jobs/attempts tables and associated indexes. [[Jobs DB](https://docs.airbyte.com/platform/understanding-airbyte/database-data-catalog#jobs-database)] Concretely: - In the
jobs
table there is a non‑terminal
sync
job for that connection (
status
is still
running
). - In the
attempts
table there is at least one attempt for that job whose
status
is
running
. [[Jobs DB](https://docs.airbyte.com/platform/understanding-airbyte/database-data-catalog#jobs-database)] - When Airbyte’s workflow tries to create another attempt (e.g., retry / new manual sync), it hits this guard and throws
IllegalStateException
instead of double‑running. This is similar in spirit to the “A sync is already running for: <connection_id>” error when you click “Sync now” on a connection that already has a running job. [[Stuck sync already running](https://discuss.airbyte.io/t/3336)] What you can do using only documented behavior: 1. Confirm job / attempt state in the DB Query the Jobs DB (Postgres) for that job and connection:
Copy code
sql    -- job record    SELECT id, config_type, scope, status, created_at, updated_at    FROM jobs    WHERE id = 60048;     -- attempts for that job    SELECT id, attempt_number, status, created_at, updated_at, ended_at    FROM attempts    WHERE job_id = 60048    ORDER BY attempt_number;
You’re looking for any attempt still marked
running
with no
ended_at
. [[Jobs DB](https://docs.airbyte.com/platform/understanding-airbyte/database-data-catalog#jobs-database)] 2. Check whether the job is really stuck - In the Airbyte UI, open the connection and check the job history page. - If the job shows as
running
but logs are no longer progressing (e.g., the worker pod/container died, similar to the “sync attempt with killed pod hangs eternally” issue), it’s effectively stuck. [[Killed pod hangs](https://github.com/airbytehq/airbyte/issues/48879)] - If logs do show an ongoing sync, then the error is expected: something tried to trigger another attempt while the current one is still running. 3. If it is stuck and you just want to clear it The knowledge sources don’t document an official DB‑level manual fix for stuck running attempts, so I cannot give a supported SQL update. The documented behaviors are: - Airbyte will normally retry attempts according to its retry rules (number of attempts and backoff) when an attempt fails. [[Sync jobs](https://docs.airbyte.com/platform/understanding-airbyte/jobs#sync-jobs)] - Only one non‑terminal job per connection can exist at a time (enforced via
scope_non_terminal_status_idx
). [[Jobs DB](https://docs.airbyte.com/platform/understanding-airbyte/database-data-catalog#jobs-database)] If the attempt is stuck in
running
and the underlying worker has died (so no automatic failure/heartbeat), this matches known reports where jobs “hang eternally.” [[Sync attempt killed pod](https://github.com/airbytehq/airbyte/issues/48879)] In that case, the documented/safe options from the UI are: - Cancel the job from the Jobs page (if the UI still allows it), then - Trigger a new manual sync once the job is in a terminal state (`failed`/`cancelled`). 4. If you keep seeing this on many connections Given there are known issues where pods dying externally can leave jobs marked as running indefinitely, you may be hitting a similar class of bug in your deployment. [[Sync attempt killed pod](https://github.com/airbytehq/airbyte/issues/48879)] Since the knowledge sources don’t describe a configuration switch or fix for this specific
IllegalStateException
, the next step would be: - Collect: - Airbyte version - Deployment type (Docker / Kubernetes, Helm chart version if any) - Screenshot / export of the job history for that connection - Open a GitHub issue or forum post referencing the stuck job and this exception so the team can correlate with existing Temporal/heartbeat behavior. Because the docs and linked issues don’t provide a fully documented remediation for this exact stack trace beyond the general job/attempt semantics above, I can’t give more specific corrective steps without going beyond the provided information.