Ignite version: 2.5
This is on a 2 node cluster in AWS with 32GB RAM and 8 cores each.
When executing the following SQL query, the query fills the JVM heap of each
node (specified at 10 GB each node) and exits afterwards. On restart, the
USER cache is removed from persistence (the cache on which the SELECT
statement is being performed)
Any tips? Is this expected behavior?
SQL query:
INSERT INTO user_email
SELECT
email,max(sex),year(max(birthdate)),max(country),max(province),substr(email,position('@',email)+1)
FROM USER WHERE email IS NOT NULL GROUP BY email;
JVM options (excluding ignite libs and ignite home): -Xms18g -Xmx18g -server
-XX:+AggressiveOpts -XX:MaxMetaspaceSize=2048m
-Djava.net.preferIPv4Stack=true -XX:+UseParNewGC -XX:+UseConcMarkSweepGC
-XX:+UseTLAB -XX:NewSize=128m -XX:MaxNewSize=128m -XX:MaxTenuringThreshold=0
-XX:SurvivorRatio=1024 -XX:+UseCMSInitiatingOccupancyOnly
-XX:CMSInitiatingOccupancyFraction=40 -XX:MaxGCPauseMillis=1000
-XX:InitiatingHeapOccupancyPercent=50 -XX:+UseCompressedOops
-XX:ParallelGCThreads=8 -XX:ConcGCThreads=8 -XX:+DisableExplicitGC
-XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath=/usr/share/apache-ignite
-XX:+ExitOnOutOfMemoryError -DIGNITE_QUIET=false
-DIGNITE_SUCCESS_FILE=/usr/share/apache-ignite/work/ignite_success_ccedb5ad-5e37-4fee-aeae-19456d7da609
-Dcom.sun.management.jmxremote -Dcom.sun.management.jmxremote.port=49124
-Dcom.sun.management.jmxremote.authenticate=false
-Dcom.sun.management.jmxremote.ssl=false
Config (AWS credentials removed):
Logs:
Restart log:
--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
