@kapa.ai what can i do if my disk has no free space left?

When Airbyte reports “no space left on device”, there are a few common culprits and fixes, depending on your deployment. ### 1. MinIO / S3-compatible storage is full If you use the bundled

airbyte-minio

(Kubernetes/OSS Helm): - MinIO mainly stores job logs and related outputs. [[Minio cleanup](https://discuss.airbyte.io/t/3101); [S3 Minio full](https://discuss.airbyte.io/t/1857)] - You can safely delete old logs inside the MinIO volume; examples from running pods: - Older path name:

/storage/airbyte-dev-logs/job-logging/workspace/

[[Minio cleanup](https://discuss.airbyte.io/t/3101)] - Newer path name:

/storage/airbyte-storage/job-logging/workspace

and possibly

/storage/airbyte-storage/workload/output

[[Syncs hanging](https://github.com/airbytehq/airbyte/issues/55812)] After deleting logs/outputs, MinIO frees disk and syncs can run again. To avoid recurrence, reduce how long Temporal keeps workflow history (this directly affects log retention in workspace):

yaml env_vars:   TEMPORAL_HISTORY_RETENTION_IN_DAYS: 7

Setting this env var (e.g., in your Helm values) shortens history and reduces MinIO usage. [[Syncs hanging](https://github.com/airbytehq/airbyte/issues/55812); [Log sweeping](https://discuss.airbyte.io/t/2714)] If you see an error like “Storage backend has reached its minimum free disk threshold” or

XMinioStorageFull

, that also indicates your MinIO disk is full and you must delete objects or increase the PVC size. [[S3 Minio full](https://discuss.airbyte.io/t/1857); [S3 + Minio issue](https://github.com/airbytehq/airbyte/issues/59197)] --- ### 2. Airbyte workspace logs on the host are too large (Docker / VM) Airbyte’s workspace (logs, job artifacts) can fill the host disk over time and cause: - Sync jobs failing - UI not loading - Commands like

docker logs

failing with “no space left on device” [[Syncs suddenly failing](https://discuss.airbyte.io/t/3303); [Airbyte rotation logs](https://discuss.airbyte.io/t/332)] Current behavior: - From v0.40.5+, Airbyte periodically cleans files from the workspace (default: every 30 days). [[Log sweeping](https://discuss.airbyte.io/t/2714)] - You can configure retention with:

bash   TEMPORAL_HISTORY_RETENTION_IN_DAYS=1   # example: keep 1 day

This replaces the older

MINIMUM_WORKSPACE_RETENTION_DAYS

/

MAXIMUM_WORKSPACE_RETENTION_DAYS

/

MAXIMUM_WORKSPACE_SIZE_MB

mechanism, which no longer applies with the Temporal-based scheduler. [[Log sweeping](https://discuss.airbyte.io/t/2714)] If your disk is already full: - Manually delete old log folders under the Airbyte workspace directory (on many Docker installs this is something like

./data

where you cloned Airbyte), or, if using MinIO as above, clear logs from MinIO. --- ### 3. Destination database temp or tmp tables Sometimes the “no space left on device” is in the destination database, not the Airbyte host: - Example: Postgres / Aurora error

could not write to file "base/pgsql_tmp/..." : No space left on device

appears during normalization. [[No space normalization](https://discuss.airbyte.io/t/1572); [No space root cause](https://discuss.airbyte.io/t/1572)] In that case, you must free or increase storage on the destination DB side (e.g., clean up large temp files, increase storage allocated to the DB instance). [[No space root cause](https://discuss.airbyte.io/t/1572)] Also, failed syncs may leave large

_airbyte_tmp_*

tables in the destination; those are not auto-deleted when a sync fails and can consume space. You can drop these

_airbyte_tmp_*

tables manually in the destination DB. [[S3 source tmp tables](https://discuss.airbyte.io/t/2067)] --- ### 4. Jobs database or other containers filling disk A very large jobs database (Airbyte internal DB) can also fill the disk and bring

airbyte-db

down, which in turn breaks the whole instance. [[Syncs suddenly failing](https://discuss.airbyte.io/t/3303)] In that case you may need to: - Increase disk allocated to the host/volume used by `airbyte-db`; and/or - Manually prune job records in the jobs DB (not directly documented here, but this is what the referenced user did by truncating jobs history). [[Syncs suddenly failing](https://discuss.airbyte.io/t/3303)] --- ### 5. Summary of concrete actions Depending on where your space is exhausted: 1. MinIO PVC full - SSH/exec into MinIO pod and delete contents in: -

/storage/airbyte-storage/job-logging/workspace

-

/storage/airbyte-storage/workload/output

(or

/storage/airbyte-dev-logs/job-logging/workspace

on older setups). [[Minio cleanup](https://discuss.airbyte.io/t/3101); [Syncs hanging](https://github.com/airbytehq/airbyte/issues/55812)] - Optionally increase MinIO PVC size. - Set

TEMPORAL_HISTORY_RETENTION_IN_DAYS

to a lower value. 2. Host workspace full (Docker/VM) - Manually remove old log/job folders under the Airbyte workspace directory. - Upgrade to ≥0.40.5 and set

TEMPORAL_HISTORY_RETENTION_IN_DAYS

to control retention. [[Log sweeping](https://discuss.airbyte.io/t/2714)] 3. Destination DB out of space - Increase DB storage or clean up space (including temp files or

_airbyte_tmp_*

tables). [[No space root cause](https://discuss.airbyte.io/t/1572); [S3 source tmp tables](https://discuss.airbyte.io/t/2067)] If you share what deployment you’re using (Docker vs Kubernetes, and whether you see MinIO-related errors), I can point to the most relevant of these steps.

@kapa.ai is safe to a do a image prune?

A message from kapa.ai