Hi, In FULL_ASYNC mode the API call returns before the update message is sent to a remote node, let alone the response receipt from the remote node. This means that in FULL_ASYNC mode you can stop your client even before the grid knows that you wanted to put something in the cache. You need to use some of the mechanisms Ignite provides to make sure your updates are written to the grid.
I suggest using IgniteDataStreamer in order to populate caches, it will give you required guarantees and good performance. Please refer to org.apache.ignite.examples.datagrid.CacheDataStreamerExample. --AG 2016-08-29 11:28 GMT+03:00 seyeony <[email protected]>: > hi Val, > thank you very much for your reply and recommendations on my problem. > i changed my id due to the problem with mailing list... > i use the option FULL_ASYNC dut to the insert performance. > > i ran 3 nodes of empty ignite in each server. i tried to put 100000 Person > object as client mode. > i got the following messages (maybe the count of this messages are same as > the count of data loss) > [20:20:43,055][WARNING][sys-#26%null%][atomic] Skip near update request, > node originated update request left [futId=GridCacheVersion > [topVer=83688908, time=1472209800520, order=1472209780296, nodeOrder=19], > node=3f236a2d-63e4-4146-b105-ef42975c8de2] > > if i sleep about 10 secs before the client program quits after completion > of > 100000 inserts, i can see the 100000 keys are inserted in the cache. > > thanks. > > > > -- > View this message in context: http://apache-ignite-users. > 70518.x6.nabble.com/data-loss-occurred-sometimes-when- > configured-primary-backups-with-full-async-tp7297p7365.html > Sent from the Apache Ignite Users mailing list archive at Nabble.com. >
