Hello! 1. Normally, all server nodes has some primary partitions. So client node will be getting updates from all servers. 2. It is done eagerly in case FULL_SYNC mode is used for cache. If PRIMARY_SYNC is used it will happen somewhere in the future after the operation is complete.
Regards, -- Ilya Kasnacheev вт, 4 дек. 2018 г. в 11:08, rom_we <rom.weinst...@sizmek.com>: > Hello, > I'm exploring ignite as a distributed in memory cache. > The cache configuration contains a dedicated with ignite server nodes > (partitioned) and application bootstrapping as a client with near cache. > > I would like to better understand the implications of running ignite near > cache in my application regarding network overload. I know near cache is > automatically synchronized with remote server cache > 1. Does it mean client nodes register to all servers events in order to > invalid cache entry ? register only to > relevant primary partition nodes ? > 2. Invalidation of entry is done eagerly or check remote cache for each > get() operation ? > > I would appreciate if you could also refer me to more detailed near cache > documentation, I only found high level description of it. > > > > > > > > > > > > > -- > Sent from: http://apache-ignite-users.70518.x6.nabble.com/ >