We have two scenes that need to be used on our production env. 1、We have lots of middle datas that need to be saved.
We use yarn ignite integration to save these datas. We have already deployed this successfully. our idea code like this to start ignite yarn code as below *val igniteContext = new IgniteContext[Integer, Integer](sc, "/u01/apache-ignite-fabric-1.6.0-bin/config/default-config.xml")* We have no issues for this 1) 2、Another scene that we also have embedded ignite spark integration, all the ignite process need to run with spark job. we can't shared these data. val igniteContext = new IgniteContext[Integer, Integer](sc, "/u01/apache-ignite-fabric-1.6.0-bin/config/default-config.xml",*false*) We use this code to start ignite, is this right? we get error message when we use this way, error issues is 16/08/26 16:23:52 ERROR YarnScheduler: Lost executor 1 on xxxx: Container marked as failed: container_1455892346017_5494_01_000002 on host: vmsecdomain010194054060.cm10. Exit status: -100. Diagnostics: Container released on a *lost* node sometime this mode will clean *yarn-ignite-contain existed proces*s , I think that they are separated. right? -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Embedded-mode-ignite-on-spark-tp6942p7359.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.
