Hello! You should keep pageSize the default, and checkpointBufferSize should be tuned up until you only see checkpoints with reason 'timeout', i.e. when you will not be running out of it.
You should be aiming for an optimal maxSize/checkpointBufferSize mix. Regards, -- Ilya Kasnacheev чт, 29 нояб. 2018 г. в 15:58, Viraj Rathod <vir9...@gmail.com>: > Actual operations contain high write operations and lesser read ones. > > Okay distributed storage noted. Thank you. > > On Thu, 29 Nov 2018 at 6:16 PM, Ilya Kasnacheev <ilya.kasnach...@gmail.com> > wrote: > >> Hello! >> >> What are actual operations that you want best TPS for? For different >> operations the the answer will be different. With regards to maxSize you >> should try to load your data, see how much memory it will take. >> >> Note that Apache Ignite will distribute cache between all four nodes by >> default. >> >> Regards, >> >> -- >> Ilya Kasnacheev >> >> >> чт, 29 нояб. 2018 г. в 15:32, Viraj Rathod <vir9...@gmail.com>: >> >>> I am setting up a cluster of 4 nodes. 2 are partitioned and 2 are >>> backups of the respective nodes. >>> The dataset contains 1 million rows of 150 columns having VARCHAR >>> values. >>> What would be the configurations needed for the same? In regards of >>> pagesize, maxsize, checkpointbuffersize. What would be ideal values of >>> these parameters for fast and efficient TPS? >>> -- >>> Regards, >>> Viraj Rathod >>> >> -- > Regards, > Viraj Rathod >
