Re: [I] [SUPPORT] Hudi 0.13.1 on EMR, MOR table writer hangs intermittently with S3 read timeout error for column stats index [hudi]
CTTY commented on issue #10415: URL: https://github.com/apache/hudi/issues/10415#issuecomment-1977660648 This looks similar to this issue: https://github.com/apache/hudi/issues/7487 where user ran into S3 throttling issue due to too many S3 calls. Was wondering if you can check if there are too many 503 error code from your S3 bucket? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@hudi.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
Re: [I] [SUPPORT] Hudi 0.13.1 on EMR, MOR table writer hangs intermittently with S3 read timeout error for column stats index [hudi]
ad1happy2go commented on issue #10415: URL: https://github.com/apache/hudi/issues/10415#issuecomment-1919021038 Thanks for trying @ergophobiac. @CTTY any insights here ? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@hudi.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
Re: [I] [SUPPORT] Hudi 0.13.1 on EMR, MOR table writer hangs intermittently with S3 read timeout error for column stats index [hudi]
ergophobiac commented on issue #10415: URL: https://github.com/apache/hudi/issues/10415#issuecomment-1886257556 Hello @ad1happy2go , We ran a test with the same configurations, just one addition: spark.hadoop.fs.s3a.connection.maximum=2000. (We found a resource saying the default on EMR is 50). Ran into the same error, application failed 3 days after starting. We have disabled multi-modal indexing for now. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@hudi.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
Re: [I] [SUPPORT] Hudi 0.13.1 on EMR, MOR table writer hangs intermittently with S3 read timeout error for column stats index [hudi]
ergophobiac commented on issue #10415: URL: https://github.com/apache/hudi/issues/10415#issuecomment-1874167209 Hey @ad1happy2go, we have a test case running, we'll observe till we're sure it's stable and let you know how it turns out. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@hudi.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
Re: [I] [SUPPORT] Hudi 0.13.1 on EMR, MOR table writer hangs intermittently with S3 read timeout error for column stats index [hudi]
ad1happy2go commented on issue #10415: URL: https://github.com/apache/hudi/issues/10415#issuecomment-1874092759 @ergophobiac Did you got a chance to try this out? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@hudi.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
Re: [I] [SUPPORT] Hudi 0.13.1 on EMR, MOR table writer hangs intermittently with S3 read timeout error for column stats index [hudi]
ad1happy2go commented on issue #10415: URL: https://github.com/apache/hudi/issues/10415#issuecomment-1870277355 @ergophobiac Are you setting fs.s3a.connection.maximum to a higher value? Can you try increasing its value and try? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@hudi.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[I] [SUPPORT] Hudi 0.13.1 on EMR, MOR table writer hangs intermittently with S3 read timeout error for column stats index [hudi]
ergophobiac opened a new issue, #10415: URL: https://github.com/apache/hudi/issues/10415 **Describe the problem you faced** Stack: Hudi 0.13.1, EMR 6.13.0, Spark 3.4.1 We are writing to an MOR table in S3, using Spark Structured Streaming job on EMR. Once this job has run for a while (like 12+ hours) we notice that at random times, there are long periods of time with no operations/commits on the timeline. Usually, the offending executor is removed and the task retried, but eventually, it stalls and the application exits with fatal errors on the driver. We are using deployment model B: Single writer with async services + OCC for metadata table. We also tried using DynamoDB based lock provider (in case OCC was the culprit) and the same thing happened. Sometimes, a deltacommit, clean or compaction is inflight but never completed, and our logs are reporting the same error over and over: (on Spark driver stderr)  The timeout always occurs while trying to read column stats. Meanwhile, the timeline shows no progress:  Notice that the deltacommit inflight at 10:20:01 AM is stuck (all prior deltacommits were finished within 45 seconds). The micro-batch interval is 5 mins. We can see no active tasks on the Spark Web UI:  In some instances, the tasks will succeed eventually and the job will progress ahead, until it gets stuck again with the same errors. Here are all our Hudi configs: hoodie.table.version -> 5 hoodie.datasource.write.hive_style_partitioning -> True hoodie.datasource.hive_sync.enable -> True hoodie.datasource.hive_sync.auto_create_database -> True hoodie.datasource.hive_sync.skip_ro_suffix-> True hoodie.parquet.small.file.limit -> 104857600 hoodie.parquet.max.file.size -> 125829120 hoodie.compact.inline.trigger.strategy -> NUM_OR_TIME hoodie.compact.inline.max.delta.commits-> 3 hoodie.compact.inline.max.delta.seconds-> 600 hoodie.parquet.compression.codec -> snappy hoodie.clean.automatic -> True hoodie.index.type -> BLOOM hoodie.bloom.index.use.metadata-> True hoodie.metadata.enable -> True hoodie.metadata.index.bloom.filter.enable -> True hoodie.metadata.index.column.stats.enable-> True hoodie.keep.max.commits-> 50 hoodie.archive.automatic -> True hoodie.archive.beyond.savepoint -> True hoodie.metrics.on -> True hoodie.metadata.metrics.enable -> True hoodie.metrics.executor.enable-> True hoodie.metrics.reporter.type -> GRAPHITE hoodie.metrics.graphite.host -> hoodie.metrics.graphite.port -> 2003 hoodie.metrics.graphite.report.period.seconds -> 30 hoodie.metrics.graphite.metric.prefix-> test_prefix_demo_mor hoodie.cleaner.policy.failed.writes -> LAZY hoodie.write.concurrency.mode -> OPTIMISTIC_CONCURRENCY_CONTROL hoodie.write.lock.provider -> org.apache.hudi.client.transaction.lock.InProcessLockProvider hoodie.metrics.lock.enable-> True hoodie.clean.async-> True hoodie.archive.async -> True hoodie.metadata.index.async-> False hoodie.metadata.clean.async -> False hoodie.cleaner.policy -> KEEP_LATEST_BY_HOURS hoodie.cleaner.hours.retained-> 1 hoodie.datasource.write.table.name -> table_0
