By reading document we know it need read object from off-heap to on-heap when do some reads on server node. We do some timer job that would query cache(igniteCache.withKeepBinary().query(new ScanQuery())) , Does this operation need more on-heap memory? we setup CacheConfiguration.setOnheapCacheEnabled (false) And we also want to know using EntryProcessor(withKeepBinary or not) need more on-heap memory?
c c <[email protected]> 于2019年12月8日周日 上午10:24写道： > thank you very much. > > Mikael <[email protected]> 于2019年12月8日周日 上午1:24写道： > >> Hi! >> >> The data regions are always off-heap, you just configure the Java heap >> for on-heap usage with -Xmx and so on as usual, have a look in the >> ignite.sh/ignite.bat, it depends on how you run your application, just >> configure this any way you like if you use embedded Ignite instance, also >> read the section about capacity planning. >> >> The java heap is just for java objects, services and any on-heap data, >> all caches are stored in data regions and they are off heap and have >> nothing to do with -Xmx except when you use on-heap caching. >> >> The Ignite documentation is very good and explains all you need to know >> on how to configure it. >> >> >> https://apacheignite.readme.io/docs/memory-configuration#section-on-heap-caching >> >> https://apacheignite.readme.io/docs/memory-configuration >> >> Mikael >> Den 2019-12-07 kl. 17:41, skrev c c: >> >> HI, >> According to document we can setup memory size by >> org.apache.ignite.configuration.DataStorageConfiguration. >> But we do not know this works for off-heap or on-heap memory. We want to >> know how to setup ignite jvm startup option(xms, xmx). Shoud jvm heap >> memory be great than maxSixe in DataStorageConfiguration. We know some >> hot data would be deserialized from off-heap to on-heap. Would you mind >> giving me some advice? thanks very much! >> >>
