Hi Ilya, >>Generally a reducer node will need to connect to all map nodes while doing SQL query.
>>You will not see this message if connection was already up. *Here are examples of 'outgoing communication' when the count was incorrect : *Example 1 : 2021-01-15T17:04:35,905 INFO o.a.i.s.c.t.TcpCommunicationSpi [grid-nio-worker-tcp-comm-0-#60%Instance%]: Established outgoing communication connection [locAddr=/a.b.c.153:59010, rmtAddr=/a.b.c.199:47130] 2021-01-15T17:04:51,682 INFO o.a.i.s.c.t.TcpCommunicationSpi [grid-nio-worker-tcp-comm-1-#61%Instance%]: Established outgoing communication connection [locAddr=/a.b.c.153:61374, rmtAddr=/a.b.c.201:47130] /2021-01-15T17:04:51,840 INFO o.a.i.s.c.t.TcpCommunicationSpi [grid-nio-worker-tcp-comm-2-#62%Instance%]: Established outgoing communication connection [locAddr=/x.y.z.153:52558, rmtAddr=machine003.cmc.local/x.y.z.202:47130]/ Example 2 : 2021-01-15T17:17:25,876 INFO o.a.i.s.c.t.TcpCommunicationSpi [grid-nio-worker-tcp-comm-0-#60%Instance%]: Established outgoing communication connection [locAddr=/a.b.c.153:59786, rmtAddr=/a.b.c.199:47130] 2021-01-15T17:17:41,720 INFO o.a.i.s.c.t.TcpCommunicationSpi [grid-nio-worker-tcp-comm-1-#61%Instance%]: Established outgoing communication connection [locAddr=/a.b.c.153:62150, rmtAddr=/a.b.c.201:47130] /2021-01-15T17:17:41,878 INFO o.a.i.s.c.t.TcpCommunicationSpi [grid-nio-worker-tcp-comm-2-#62%Instance%]: Established outgoing communication connection [locAddr=/x.y.z.153:53334, rmtAddr=machine003.cmc.local/x.y.z.202:47130]/ *Here is an example of 'outgoing communication' when the count was correct : *Example 3 : 2021-01-15T17:30:12,638 INFO o.a.i.s.c.t.TcpCommunicationSpi [grid-nio-worker-tcp-comm-0-#60%Instance%]: Established outgoing communication connection [locAddr=/a.b.c.153:60534, rmtAddr=/a.b.c.199:47130] 2021-01-15T17:30:28,403 INFO o.a.i.s.c.t.TcpCommunicationSpi [grid-nio-worker-tcp-comm-1-#61%Instance%]: Established outgoing communication connection [locAddr=/a.b.c.153:62898, rmtAddr=/a.b.c.201:47130] 1. We are on a 3 node cluster ( a.b.c.199, a.b.c.201, x.y.z.202 ). When the count is right it doesn't even try to make a connection to x.y.z.202 . It is happy with the connections made to a.b.c.199, a.b.c.201. As mentioned in the cases where the count is incorrect the connection to x.y.z.202 is made between cache.query and cursor.getall. 2. The cache that is used in the query here is a REPLICATED cache. As per this documentation here - the query on a REPLICATED CACHE would not be run against multiple server nodes - https://apacheignite-sql.readme.io/docs/select Would the reducer still be used ? regards, Veena. -- Sent from: http://apache-ignite-users.70518.x6.nabble.com/