Open guilhermenoronha opened 4 months ago
@guilhermenoronha did you try to increase resurces used by the source/destination? Exist 137 means mostly OOM
I set up the following variables:
JOB_MAIN_CONTAINER_MEMORY_REQUEST=3Gi JOB_MAIN_CONTAINER_MEMORY_LIMIT=6Gi CHECK_JOB_MAIN_CONTAINER_MEMORY_REQUEST=3Gi CHECK_JOB_MAIN_CONTAINER_MEMORY_LIMIT=6Gi
As I concern, the changes took effect as shown in the log below:
Preparing command: docker run --rm --init -i -w /data/514/0 --log-driver none --name source-declarative-manifest-check-514-0-wfels --network host -v airbyte_workspace:/data -v oss_local_root:/local -e DEPLOYMENT_MODE=OSS -e WORKER_CONNECTOR_IMAGE=airbyte/source-declarative-manifest:0.85.0 -e AUTO_DETECT_SCHEMA=true -e LAUNCHDARKLY_KEY= -e SOCAT_KUBE_CPU_REQUEST=0.1 -e SOCAT_KUBE_CPU_LIMIT=2.0 -e FIELD_SELECTION_WORKSPACES= -e USE_STREAM_CAPABLE_STATE=true -e WORKER_ENVIRONMENT=DOCKER -e AIRBYTE_ROLE=dev -e APPLY_FIELD_SELECTION=false -e WORKER_JOB_ATTEMPT=0 -e OTEL_COLLECTOR_ENDPOINT=http://host.docker.internal:4317 -e FEATURE_FLAG_CLIENT=config -e AIRBYTE_VERSION=0.58.0 -e WORKER_JOB_ID=514 --memory-reservation=3Gi --memory=6Gi airbyte/source-declarative-manifest:0.85.0 check --config source_config.json
However, it didn't work. Checking the memory usage on the Airbyte machine, it remains stable between 3G and 5G of usage (considering the whole Airbyte architectures). See the print below:
Any ideas?
The datalake destination keeps records in memory and automatically flushes them to athena every 10000 records. It is possible that that 10k limit is to high for the type data you are trying to sync (judging from your output 398.44 MB| 2,850
) and that can cause an OOM in the destination processs. I think the best way forward here is to allow for configuring that limit in the config instead of hardcoding it.
Connector Name
source-aws datalake
Connector Version
0.1.7
What step the error happened?
During the sync
Relevant information
My connector is getting this error 137 every time I try to extract a data. Something I noticed is the connector getting error when the extraction comes to approximately 400mb of data in the source. Here is some metadata of the extraction when the error occurred:
11:43AM 06/05/2024 |398.44 MB| 2,850 records extracted |no records loaded| Job id: 214| 17m 21s
Here is the full log:
default_workspace_job_214_attempt_2.txt
Relevant log output
Contribute