https://pinot.apache.org/ logo
Join Slack
Powered by
# general
  • n
    oh you started with 
    hybrid
    quickstart
  • d
    Yea
  • x
    Hi Pinot Community, I’m calling for a vote to the release Apache Pinot (incubating) 0.3.0: [VOTE] Apache Pinot (incubating) 0.3.0 RC1 Please reply to pinot dev mailing list (dev@pinot.apache.org) to vote.
    ➕ 3
  • s
    <!here> 👆👆
  • d
    Has anyone gotten the docker streaming QuickStart to work recently? When I try, docker container dies after a couple seconds with an exit code of 1.
  • d
    The batch version works fine for me.
  • x
    do you have the logs ?
  • x
    the command should be
  • x
    Copy code
    # stop previous container, if any, or use different network
docker run \
    --network=pinot-demo \
    --name pinot-quickstart \
    -p 9000:9000 \
    -d apachepinot/pinot:latest QuickStart \
    -type stream
  • x
    not streaming
  • x
    I updated the doc, could you try again
  • d
    Yes, that worked. Thanks!
  • y
    folks, per https://readthedocs.org/projects/pinot/downloads/pdf/latest/
    Copy code
    FASTHLL (WARN: will be deprecated soon. FASTHLL stores serialized HyperLogLog in String format, which
performs worse than DISTINCTCOUNTHLL, which supports serialized HyperLogLog in BYTES (byte array)
format)
    is there documents about the difference between 
    fasthll
    and 
    distinctcounthll
    ?
  • x
    fasthll is used before bytes supported, so we use string to represent hyperloglog object
  • x
    distinctCountHll is on bytes column
  • y
    other than the input formats, are there other impl differences?
  • x
    i dont think so
  • y
    cool. for porting from fasthll to distintcounthll, whats the suggested function for converting the input from string to bytes?
  • y
    btw, is there a plan to add some docs like https://prestodb.io/docs/current/functions/hyperloglog.html that clearly describes the input/output and the semantics?
  • x
    Will do !
    👍 1
  • k
  • m
    Will validate and +1
  • k
  • n
    Sharing a design proposal for supporting Tiered Storage in Pinot:  https://docs.google.com/document/d/1Z4FLg3ezHpqvc6zhy0jR6Wi2OL8wLO_lRC6aLkskFgs/edit?usp=sharing
  • e
    Question: can incomingGranularitySpec convert string columns in the format of 
    2020-03-13T11:51:46.075-07:00[America/Los_Angeles]
    ? Or would you recommend just using epoch? For the primary time column.
  • d
    Has anyone here gotta an AWS Lambda to send messages to a Kafka started using the AWS EKS Pinot/Kafka instructions? I followed Kafka's NodePort and they didn't work. I figured someone else has gotten this to work and can forward advice.
  • x
    I haven’t tried out lambda, please let me know if you make this pipeline working, it will be a good story to put up.
    👍 1
  • i
    Hello. We are trying to choose OLAP solution for our project and we are considering Pinot. First question that we have: Is it possible/efficient to aggregate only on LATEST metric value for targets = "Give me current average CPU usage across all targets"? (maybe with some workarounds, additional tables etc.) I read pretty interesting article regarding workarounds for Clickhouse here: https://www.altinity.com/blog/clickhouse-continues-to-crush-time-series and "lastpoint"+"groupby-orderby-limit" cases of TSBS Time series benchmark and I am wondering how Pinot could deal with those cases.
  • k
    Reading...
  • k
    Pinot does not support it. We can add a special operator to achieve this
1...119120121...160Latest