I would add that it’s possible to allow reads from backups if a read happens on a backup side. Refer to CacheConfiguration.isReadFromBackup() property. However, keep in mind, that if the property is enabled you may have stale reads (a value has already been updated on a primary but not yet on a backup).
— Denis > On Oct 30, 2016, at 10:03 AM, srecon <[email protected]> wrote: > > Hello Devis! > > You have to count a few things as follows: > 1) Ignite always reads the data from the Primary node. In the case of > failure of the primary node, Ignite will read from the backup copy (if > configured). > 2) If you want to store entries in both Node A and B, you have to configure > the cache in Replicated mode. In replicated mode, every node contains the > replica (a redundant copy) of data from another node. > cacheCfg.setCacheMode(REPLICATED); > cacheCfg.setBackups(1); > 3) Whenever you use partitioned mode, a portion (partition) of data will be > available on another node and a backup copy (if configured). > > P.S. You can read an architecture overview of Ignite here > <https://leanpub.com/ignite/read> . https://leanpub.com/ignite/read > <https://leanpub.com/ignite/read> > > Regards > Shamim > > > > -- > View this message in context: > http://apache-ignite-users.70518.x6.nabble.com/About-Cache-tp8611p8613.html > Sent from the Apache Ignite Users mailing list archive at Nabble.com.
