Closed alberttwong closed 1 month ago
This doesn't work. You can see from the logs that it picked up the setting.
export HOODIE_ENV_FS_ATOMIC_CREATION_SUPPORT=s3a
export HOODIE_ENV_HOODIE_FS_ATOMIC_CREATION_SUPPORT=s3a
Doesn't work either
export HOODIE_ENV_hoodie_DOT_fs_DOT_atomic_creation_DOT_support=s3a
It seems that your settings are not working? If you don't need to use locks, you can try setting hoodie.write.lock.provider to org.apache.hudi.client.transaction.lock.InProcessLockProvider.
I'm not sure if I should change https://hudi.apache.org/docs/configurations/#hoodiewritelockprovider since it's org.apache.hudi.client.transaction.lock.ZookeeperBasedLockProvider as the default. Anyhow I tried.
export HOODIE_ENV_hoodie_DOT_write_DOT_lock_DOT_provider=org.apache.hudi.client.transaction.lock.InProcessLockProvider
21748 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO TransactionManager: Transaction starting for Option{val=[==>20240905041525278__compaction__REQUESTED]} with latest completed transaction instant Optional.empty
21748 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO LockManager: LockProvider org.apache.hudi.client.transaction.lock.FileSystemBasedLockProvider
21749 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO HadoopFSUtils: Picking up value for hoodie env var : HOODIE_ENV_hoodie_DOT_write_DOT_lock_DOT_provider
21749 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO HadoopFSUtils: Picking up value for hoodie env var : HOODIE_ENV_fs_DOT_s3a_DOT_endpoint
21749 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO HadoopFSUtils: Picking up value for hoodie env var : HOODIE_ENV_FS_ATOMIC_CREATION_SUPPORT
21749 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO HadoopFSUtils: Picking up value for hoodie env var : HOODIE_ENV_fs_DOT_s3a_DOT_access_DOT_key
21749 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO HadoopFSUtils: Picking up value for hoodie env var : HOODIE_ENV_fs_DOT_s3a_DOT_aws_DOT_credentials_DOT_provider
21749 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO HadoopFSUtils: Picking up value for hoodie env var : HOODIE_ENV_fs_DOT_s3a_DOT_secret_DOT_key
21750 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO HadoopFSUtils: Picking up value for hoodie env var : HOODIE_ENV_hoodie_DOT_fs_DOT_atomic_creation_DOT_support
21750 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO HadoopFSUtils: Picking up value for hoodie env var : HOODIE_ENV_HOODIE_FS_ATOMIC_CREATION_SUPPORT
21751 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO TransactionManager: Transaction ending with transaction owner Option{val=[==>20240905041525278__compaction__REQUESTED]}
21751 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO LockManager: LockProvider org.apache.hudi.client.transaction.lock.FileSystemBasedLockProvider
21751 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO HadoopFSUtils: Picking up value for hoodie env var : HOODIE_ENV_hoodie_DOT_write_DOT_lock_DOT_provider
21751 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO HadoopFSUtils: Picking up value for hoodie env var : HOODIE_ENV_fs_DOT_s3a_DOT_endpoint
21751 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO HadoopFSUtils: Picking up value for hoodie env var : HOODIE_ENV_FS_ATOMIC_CREATION_SUPPORT
21751 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO HadoopFSUtils: Picking up value for hoodie env var : HOODIE_ENV_fs_DOT_s3a_DOT_access_DOT_key
21751 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO HadoopFSUtils: Picking up value for hoodie env var : HOODIE_ENV_fs_DOT_s3a_DOT_aws_DOT_credentials_DOT_provider
21751 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO HadoopFSUtils: Picking up value for hoodie env var : HOODIE_ENV_fs_DOT_s3a_DOT_secret_DOT_key
21751 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO HadoopFSUtils: Picking up value for hoodie env var : HOODIE_ENV_hoodie_DOT_fs_DOT_atomic_creation_DOT_support
21751 [Thread-5] INFO org.apache.hudi.cli.utils.InputStreamConsumer [] - 24/09/05 04:15:27 INFO HadoopFSUtils: Picking up value for hoodie env var : HOODIE_ENV_HOODIE_FS_ATOMIC_CREATION_SUPPORT
logs show it don'est pick up... I don't know why.
work around
export HUDI_CONF_DIR=/opt/hudi/packaging/hudi-cli-bundle/conf/
added to hudi-defaults.conf
hoodie.fs.atomic_creation.support s3a
possible other issue is that you run this on jdk 11 and not jdk 8.
Running compaction on a s3a bucket gives an HoodieLockException
To Reproduce
Steps to reproduce the behavior:
get this error by running compaction
Expected behavior
There shouldn't be any error
Environment Description
Hudi version : 0.15
Spark version : 3.4
Hive version : 2.3.9
Hadoop version : 3.4
Storage (HDFS/S3/GCS..) : minio s3
Running on Docker? (yes/no) : yes
Additional context
Add any other context about the problem here.
Stacktrace
Add the stacktrace of the error.