Ignite reads are slower with getAll

[KR Kumar <ig...@gmail.com>]

Hi guys - I have a ignite cluster with persistence enabled that has 200
million events in it. Right now read throughput is around 3000 events per
second. I have increased the IOPS to 10000 and even then I have the same
performance. Am I doing something really wrong or this is how it perform
with large amounts of data.

I am using getAll with a batch of 300 keys in one read. The cache is
basically a string key and json message, so its String,String type of cache. 

Any help/pointers ??

Thanx and Regards,
KR Kumar



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Re: Ignite reads are slower with getAll

[Stanislav Lukyanov <st...@gmail.com>]

Hi,

Increasing IOPS (I'm assuming we're talking about disk IOPS) is not always
the solution since disk performance is not always the bottleneck. It may
also be network, CPU, etc. Usually one needs to use profiling and monitoring
tools, e.g. dstat, to measure the usage of the system resources. If dstat
shows, say, 100% CPU usage you're probably not going to see a performance
gain from switching to a faster disk.

Stan


KR Kumar wrote
> Hi guys - I have a ignite cluster with persistence enabled that has 200
> million events in it. Right now read throughput is around 3000 events per
> second. I have increased the IOPS to 10000 and even then I have the same
> performance. Am I doing something really wrong or this is how it perform
> with large amounts of data.
> 
> I am using getAll with a batch of 300 keys in one read. The cache is
> basically a string key and json message, so its String,String type of
> cache. 
> 
> Any help/pointers ??
> 
> Thanx and Regards,
> KR Kumar
> 
> 
> 
> --
> Sent from: http://apache-ignite-users.70518.x6.nabble.com/





--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/