@Neha Pawar do we still need that property, looks like its not adding any value

maybe it was added to roll out the feature in a controlled manner

actually i dont know why we have that prop. i can check and remove it if i dont find any use

cool, lets create an issue and remove it

created issue: https://github.com/apache/incubator-pinot/issues/5615

Hi. I'm getting the following error related to my realtime table ingesting events from kafka.

insertions

is just a long metric column. It's not a dimension. I'm not sure why it's being hitting this.

Metrics aggregation cannot be turned ON in presence of dictionary encoded metrics, eg: insertions

For realtime metrics aggregation, all metric columns should be noDictionary

This ^^

Also dimension columns cannot be no dict

we should probably consider adding support for realtime metrics aggregation with dictionary encoded metric columns and remove this restriction unless it is infeasible

we can but it inefficient

I don't think these metrics need to be dictionary encoded. I assumed metric columns by default would not be dictionary encoded.

all columns by default are dictionary encoded AFAIK

Looks like the docs say that. I'll update my config. Thanks!

Separate topic, my

pinot-server-0

log is showing that my params in

streamConfig

were

supplied but isn't a known config

. Any ideas?

2020/06/26 17:26:05.729 WARN [ConsumerConfig] [HelixTaskExecutor-message_handle_thread] The configuration 'stream.kafka.hlc.zk.connect.string' was supplied but isn't a known config.
2020/06/26 17:26:05.729 WARN [ConsumerConfig] [HelixTaskExecutor-message_handle_thread] The configuration 'realtime.segment.flush.threshold.size' was supplied but isn't a known config.
2020/06/26 17:26:05.737 WARN [ConsumerConfig] [HelixTaskExecutor-message_handle_thread] The configuration 'stream.kafka.decoder.class.name' was supplied but isn't a known config.
2020/06/26 17:26:05.737 WARN [ConsumerConfig] [HelixTaskExecutor-message_handle_thread] The configuration 'streamType' was supplied but isn't a known config.
...

Here is what my table config looks like (with some repetitive parts removed)

{
  "tableName": "metrics",
  "tableType": "REALTIME",
  "segmentsConfig": {
    "timeColumnName": "timestamp",
    "timeType": "MILLISECONDS",
    "retentionTimeUnit": "DAYS",
    "retentionTimeValue": "5",
    "segmentPushType": "APPEND",
    "segmentAssignmentStrategy": "BalanceNumSegmentAssignmentStrategy",
    "schemaName": "metrics",
    "replication": "1",
    "replicasPerPartition": "1" 
  },  
  "tableIndexConfig": {
    "loadMode"  : "MMAP",
    "starTreeIndexConfigs": [
      {   
        "dimensionsSplitOrder": [
          "platform_id",
          ... my other dimensions
        ],  
        "skipStarNodeCreationForDimensions": [
        ],  
        "functionColumnPairs": [
          "SUM__insertions",
          ... my other metrics
        ]   
      }   
    ],  
    "streamConfigs": {
      "streamType": "kafka",
      "stream.kafka.consumer.type": "simple",
      "stream.kafka.topic.name": "metrics-realtime",
      ... all of the configs are ignored
    }   
  },  
  "tenants": {}, 
  "metadata": {
    "customConfigs": {}
  }   
}

those are coming from the kafka consumer. because we put the entire streamConfigs map from pinot into the consumer configs. those should be harmless.

a good beginner task to fix

Cool. I was hitting an issue where my Pinot server temporarily stopped ingesting events from Kafka and was trying to figure out if there was an issue. It's working again and still logging those lines.

Hi, I am having trouble querying the recent data from Pinot,

select * from <tablename> order by timestamp limit 10

returns empty. I also see some of the old segments to be in error state from the broker logs

Resource: <tablename>_REALTIME, partition: <tablename>__7__0__20200625T1913Z is in ERROR state
Resource: <tablename>_REALTIME, partition: <tablename>__8__0__20200625T1913Z is in ERROR state
Resource: <tablename>_REALTIME, partition: <tablename>__9__0__20200625T1913Z is in ERROR state

Wondering how do I go ahead about debugging this?

The server log should show you what the segment went into error state.

Above is from server logs, for example:

<tablename>__9__0__20200625T1913Z

is one of the segment that went into error state (though it’s an older segment)

I don’t see any logs on recent segments

Can you grep for one of the segments in error state in all the server logs (if you have separate logs per day)? It should show some error/exeception that caused this to happen.

Alternatively, you can try to re-start the server and see if the segment still gets into ERROR state.

tried restarting, seems like a helix issue let me dig a bit more

Are these old segments still within your retention period?

yeah these are within ttl

for exmple one of the exceptions I see in the server logs

Sleep for 10000ms as service status has not turned GOOD: IdealStateAndCurrentStateMatchServiceStatusCallback:partition=<tablename>__57__0__20200625T191
3Z, expected=ONLINE, found=OFFLINE, creationTime=1593202611703, modifiedTime=1593202650032, version=95, waitingFor=CurrentStateMatch, resource=<tablename>_REALTIME, numResourcesLeft=1, numTotalResources=1, minStartCount=1,;IdealStateAndExternalViewMatchServiceStatusCallback:Init;
Caught exception in state transition from OFFLINE -> ONLINE for resource: <tablename>_REALTIME, partition: <tablename>__57__0__20200625T1913Z
java.lang.RuntimeException: org.apache.pinot.spi.utils.retry.AttemptsExceededException: Operation failed after 3 attempts
        at org.apache.pinot.core.data.manager.realtime.RealtimeTableDataManager.downloadAndReplaceSegment(RealtimeTableDataManager.java:286) ~[pinot-all-0
.4.0-jar-with-dependencies.jar:0.4.0-8355d2e0e489a8d127f2e32793671fba505628a8]
        at org.apache.pinot.core.data.manager.realtime.RealtimeTableDataManager.addSegment(RealtimeTableDataManager.java:252) ~[pinot-all-0.4.0-jar-with-d
ependencies.jar:0.4.0-8355d2e0e489a8d127f2e32793671fba505628a8]
        at org.apache.pinot.server.starter.helix.HelixInstanceDataManager.addRealtimeSegment(HelixInstanceDataManager.java:132) ~[pinot-all-0.4.0-jar-with
-dependencies.jar:0.4.0-8355d2e0e489a8d127f2e32793671fba505628a8]
        at org.apache.pinot.server.starter.helix.SegmentOnlineOfflineStateModelFactory$SegmentOnlineOfflineStateModel.onBecomeOnlineFromOffline(SegmentOnlineOfflineStateModelFactory.java:164) [pinot-all-0.4.0-jar-with-dependencies.jar:0.4.0-8355d2e0e489a8d127f2e32793671fba505628a8]
        at sun.reflect.GeneratedMethodAccessor10.invoke(Unknown Source) ~[?:?]

did you setup deep store for Pinot?