Apache Cassandra - Tutorial 17 - Cassandra Write Path

[@FloatingSunfish]

Awesome tutorial series on Cassandra! Please keep going! We're learning so much!

[@jumpstartCS]

Thanks Floating Sunfish, planing on making some more videos on Cassandra as well as some other topics in the future. Stay tuned!

[@FabricioDestro]

This is the best Cassandra tutorial on KZbin

[@mateustabaldi2080]

Não posso discordar kk

[@arkumar]

Thanks man. Beautifully explained the inner working behind the curtains. This is how muh in depth most other tutorials should be. Looking forward for more such Tech Series. Thanks again.

[@Paarth2000]

Excellent videos - looking forward to your next series.

[@juliajin4732]

Awesome material...Thanks a lot, looking forward to the advanced ones....

[@saikatcse03]

As you explian the read and write path. Could you explain how deletes works or what is tombstones. How this has impact on performance .

[@nitinkarve3608]

Simply Superb!!

[@saikatcse03]

Thanks . Also it would great if you could explain is the JVM GC could cause STW problem in cassandra

[@cantwaittowatch]

great video. question: at frame 5:22, when the flush happens, I see that the data is removed from the mem table, and you say, reads are done via sstable. Reads in mem are faster than on disk, like cache, so am not clear why memtable is not updated correctly?

[@rodrigomageste145]

Awesome..

[@atul211]

superb

[@paulfunigga]

Not sure about earlier versions of cassandra, but in the latest versions, commit log is not durable. It is only flushed to disk every 10 seconds by default. If you want to make it durable, you have to switch the commitlog_sync setting to "batch" instead of "periodic", note, however, that it will drastically decrease write performance, even if you use the fastest SSD's.