You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@druid.apache.org by GitBox <gi...@apache.org> on 2019/06/18 12:46:49 UTC
[GitHub] [incubator-druid] litao91 opened a new issue #7921: [Performance]
will a decompressed bytebuffer cache help?
litao91 opened a new issue #7921: [Performance] will a decompressed bytebuffer cache help?
URL: https://github.com/apache/incubator-druid/issues/7921
The flame-graph sampled from our production cluster shows that a large portion of CPU time has spent on LZ4Decompression (probably 60%+ of all query related methods), as shown in the following picture.
I wonder would a (probably on disk) LRUCache on de-compressed data help?
It seems it would not be hard to implement. A, say, `BufferedGenericIndexd` can be created with a generated UUID, and then we can use `<uuid>-bufferNumber` as the cache key.
But I'm not sure about the effectiveness. It does introduce extra cost of cache miss penalty, random disk IO, etc.
That's why I would like to raise a discussion here and find out is it a feasible solution.
Thanks
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org