kudu-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Janne Keskitalo <janne.keskit...@paf.com>
Subject Re: Low ingestion rate from Kafka
Date Thu, 02 Nov 2017 07:31:21 GMT
2017-11-01 21:40 GMT+01:00 Todd Lipcon <todd@cloudera.com>:

> Great. Keep in mind that, since you have a UUID component at the front of
> your key, you are doing something like a random-write workload. So, as your
> data grows, if your PK column (and its bloom filters) ends up being larger
> than the available RAM for caching, each write may generate a disk seek
> which will make throughput plummet. This is unlike some other storage
> options like HBase which does "blind puts".

 Is this cache size configurable or just dependent on the available RAM on
the host? And how could I check the current sizes of the PK bloom filters
or detect when some of them don't fit the cache anymore?

Janne Keskitalo,

View raw message