"query" node runs the query engine, which can execute the SQL queries; "job" node will execute the cube build jobs, like run hive command, submit MR, update cube status, etc. "all" node will run both query engine and job engine.
You can submit a job on a "query" node, but the job's execution only happens on the "job" or "all" node; they share the same meta store, "query" node will write the job task to meta store, "job" node will read it and then execute. Hope this clarifies. 2017-11-16 10:29 GMT+08:00 jxs <[email protected]>: > Hi, ShaoFeng, what the configuration "kylin.server.mode=query" means? > I saw the behaviour too yesterday that the "query" node submitted cube and > job operations. > > > > 在2017年11月16 09时13分, "ShaoFeng Shi"<[email protected]>写道: > > > Even in a Hadoop client node (or "edge node"), Kylin can run the job > engine; The behavior of submitting a job to Hadoop can happen on any node; > you only need properly install/configure the Hadoop client and > configurations on it. > > > 2017-11-15 17:27 GMT+08:00 Roberto Tardío <[email protected]>: > >> Hi Jxs, >> >> >> In our case we have not not detectted that kylin.server.mode affects >> correct operating of Kylin in dedicated EC2 using an EMR cluster. >> >> Kind regards, >> >> >> >> El 15/11/2017 a las 3:21, jxs escribió: >> >> >> When using dedicated client node on EC2, should I configure >> kylin with the same: >> kylin.env.hdfs-working-dir = s3://our-emr/kylin/ >> kylin.storage.hbase.cluster-fs = s3://our-emr >> >> >> Should the kylin.server.mode be "query" ? >> >> >> Thanks. >> >> >> >> >> >> >> >> >> >> -- >> >> >> *Roberto Tardío Olmos* >> *Senior Big Data & Business Intelligence Consultant* >> Avenida de Brasil, 17 >> <https://maps.google.com/?q=Avenida+de+Brasil,+17&entry=gmail&source=g>, >> Planta 16.28020 Madrid >> >> Fijo: 91.788.34.10 >> >> > > > > -- > Best regards, > > Shaofeng Shi 史少锋 > > -- Best regards, Shaofeng Shi 史少锋
