<@U04S8PLJWMT> why would sync show success even th...
# ask-ai
a
@kapa.ai why would sync show success even though pod was never allocated? 2025-11-26 095322 platform INFO APPLY Stage: BUILD — (workloadId=d88dfecc-6847-4445-8733-c2d0ab51ff67_67_0_sync) 2025-11-26 095322 platform INFO APPLY Stage: CLAIM — (workloadId=d88dfecc-6847-4445-8733-c2d0ab51ff67_67_0_sync) 2025-11-26 095322 platform INFO Claimed: true for workload d88dfecc-6847-4445-8733-c2d0ab51ff67_67_0_sync via API in dataplane 2894b95b-f14c-446c-9c04-14eb422092fa (acf550f2-6155-4451-8db9-b62443826c92) 2025-11-26 095322 platform INFO APPLY Stage: LOAD_SHED — (workloadId=d88dfecc-6847-4445-8733-c2d0ab51ff67_67_0_sync) 2025-11-26 095322 platform INFO APPLY Stage: CHECK_STATUS — (workloadId=d88dfecc-6847-4445-8733-c2d0ab51ff67_67_0_sync) 2025-11-26 095322 platform INFO No pod found running for workload d88dfecc-6847-4445-8733-c2d0ab51ff67_67_0_sync 2025-11-26 095322 platform INFO APPLY Stage: MUTEX — (workloadId=d88dfecc-6847-4445-8733-c2d0ab51ff67_67_0_sync) 2025-11-26 095322 platform INFO Mutex key: d88dfecc-6847-4445-8733-c2d0ab51ff67 specified for workload: d88dfecc-6847-4445-8733-c2d0ab51ff67_67_0_sync. Attempting to delete existing pods... 2025-11-26 095322 platform INFO Existing pods for mutex key: d88dfecc-6847-4445-8733-c2d0ab51ff67 deleted. 2025-11-26 095322 platform INFO APPLY Stage: ARCHITECTURE — (workloadId=d88dfecc-6847-4445-8733-c2d0ab51ff67_67_0_sync) 2025-11-26 095322 platform INFO APPLY Stage: LAUNCH — (workloadId=d88dfecc-6847-4445-8733-c2d0ab51ff67_67_0_sync) 2025-11-26 095322 platform INFO [initContainer] image: airbyte/workload-init-container:2.0.1 resources: ResourceRequirements(claims=[], limits={memory=7530Mi, cpu=3.6723}, requests={memory=6348Mi, cpu=1.845}, additionalProperties={}) 2025-11-26 095322 platform INFO Launching replication pod: replication-job-67-attempt-0 (selectors = {kube/nodetype=airbyte-job}) with containers: 2025-11-26 095322 platform INFO [source] image: 454115844779.dkr.ecr.us-east-1.amazonaws.com/conn/teams-dev:e3395cc.4059 resources: ResourceRequirements(claims=[], limits={memory=7530Mi, cpu=3.6723}, requests={memory=6348Mi, cpu=1.845}, additionalProperties={}) 2025-11-26 095322 platform INFO [destination] image: 454115844779.dkr.ecr.us-east-1.amazonaws.com/conn/efsdest-dev:fe5e292.3541 resources: ResourceRequirements(claims=[], limits={memory=7530Mi, cpu=3.6723}, requests={memory=6348Mi, cpu=1.845}, additionalProperties={}) 2025-11-26 095322 platform INFO [orchestrator] image: airbyte/container-orchestrator:2.0.1 resources: ResourceRequirements(claims=[], limits={memory=7530Mi, cpu=3.6723}, requests={memory=6348Mi, cpu=1.845}, additionalProperties={}) 2025-11-26 102408 platform INFO ----- START POST REPLICATION OPERATIONS ----- 2025-11-26 102408 platform INFO No post-replication operation(s) to perform. 2025-11-26 102408 platform INFO ----- END POST REPLICATION OPERATIONS -----
k
Based on the provided logs and the available knowledge sources, a sync job in Airbyte may show as "success" even though a pod was never actually allocated or ran. This can happen due to a flaw in how Airbyte tracks and updates workload status: - The launcher goes through its stages (BUILD, CLAIM, LOAD_SHED, CHECK_STATUS, etc.). - If the pod is never actually created or started (as indicated by repeated "No pod found running for workload" messages), the system may still proceed through the stages and eventually mark the job as completed or successful. - This is often due to the fact that the status update from "launched" to "success" is supposed to happen when the CHECK_STATUS stage detects the pod has completed. However, if the pod was never created, the system may not properly detect this and can incorrectly mark the job as successful. - There are known issues where the workload status in the database is not updated correctly, especially if the pod never runs or if there is a failure in the status-checking logic. This can result in jobs being stuck in a "launched" state or being marked as "success" without actual execution, as described in the issues about jobs hanging or not launching pods, and the design flaw where status updates depend on successful claims and pod checks, creating a catch-22 scenario if the pod never exists or is not detected properly [[airbyte services stuck where no new job pods are launched](https://github.com/airbytehq/airbyte/issues/68115)]. In summary, this is a known design flaw in Airbyte's workload and status management, where the absence of a running pod is not always correctly handled, leading to false positives in job status reporting. If you are experiencing this, it is likely related to the issues described above.