Hi Maxim,
unfortunately no. We decided to stay at 3.11.6 for production until we found a solution because our system is latency sensitive too. >From what I can see the most impact on read-latency is on the tables using time windowed compaction strategy. Tables with e.g. leveled compaction strategy are not affected that much. I'm suspecting some change in the behavior of the bloom filter. Bloom filter false positive rate hast increased too with 3.11.7, probably causing more SSTables to be read on request unnecessarily. Reading more SSTables probably causes a higher read-latency. However, the way leveled compaction strategy is organizing the SSTables bloom filter changes might not have such a big impact on the number of SSTables to be read on request. In case you have a test environment you could try the version update and compare bloom filter and SSTable reads on your tables. Would be interested to see your results too. Regards Nico -----"Maxim Parkachov" <lazy.gop...@gmail.com> schrieb: ----- An: user@cassandra.apache.org Von: "Maxim Parkachov" <lazy.gop...@gmail.com> Datum: 23.09.2020 08:15 Betreff: Re: Re: Increased read latency with Cassandra >= 3.11.7 Hi Nico, we wanted to upgrade to 3.11.8 (from .6), but now I'm very concerned as our load is mostly read-only and very latency sensitive. Did you figure out the reason for such behaviour with the new version ? Regards, Maxim. LANCOM Systems GmbH Adenauerstr. 20 / B2 52146 Würselen Deutschland Tel: +49 2405 49936-0 Fax: +49 2405 49936-99 Web: https://www.lancom-systems.de Geschäftsführer: Ralf Koenzen, Stefan Herrlich Sitz der Gesellschaft: Aachen, Amtsgericht Aachen, HRB 16976
