Hi, 1) Caches are created by clients and cache configuration mode is not transactional. It is atomic. 2) Yes you are right, it is a putAll(). And in worst case, there could be 10 clients writing at max .1G of data, so total data written concurrently can be around 1G or slightly more. But in reality only 1 client wrote at 1 time and only 1 out of the 10 clients had .1G of data. Rest had substantially low data. 3) The keys are fairly large in size, they are proper *objects* consisting of a) several 4-6 Strings b) 2-3 Integers c) 2-3 lists of String
Keeping in mind these things, does batch size hold any importance since I am using Atomic mode ? If yes, how can I configure the same ? As I mentioned, I am using putAll (Map x) api, does it make sense that I iterate over the entries instead of putAll ? I will try to immediately clear 'x' after putting it in Ignite. Apologies but increasing heap size is a very difficult option for me although I am using 1G which typically is nothing at all. Also do you think that its important to set the directmemorysize ? -- Sent from: http://apache-ignite-users.70518.x6.nabble.com/
