[
https://issues.apache.org/jira/browse/TRAFODION-1609?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15353036#comment-15353036
]
ASF GitHub Bot commented on TRAFODION-1609:
---
Github user selvaganesang commented on a diff in the pull request:
https://github.com/apache/incubator-trafodion/pull/562#discussion_r68762512
--- Diff: core/sqf/sql/scripts/install_local_hadoop ---
@@ -847,20 +847,31 @@ EOF
cat <$MY_SW_SCRIPTS_DIR/swstatus
#!/bin/sh
-cd ${MY_SW_ROOT}
-. $MY_SW_SCRIPTS_DIR/sw_env.sh
-NUM_JAVA_PROCS=\`ps -aef | grep \$USER | grep java | grep -v grep | wc -l\`
-NUM_MYSQLD_PROCS=\`ps -aef | grep \$USER | grep mysqld | grep -v grep | wc
-l\`
+cd \${MY_SW_ROOT}
+. \$MY_SW_SCRIPTS_DIR/sw_env.sh
+JPS_OUTPUT=\`jps\`
+SERVICES='HMaster NodeManager ResourceManager NameNode DataNode
SecondaryNameNode'
+for s in \$SERVICES; do
+if [[ ! \$JPS_OUTPUT =~ \$s ]]; then
+MISS_SERVICE="\$s \$MISS_SERVICE"
+fi
+done
-if [ "\$1" == "-v" ]; then
- ps -aef | grep \$USER | grep java | grep -v grep
- ps -aef | grep \$USER | grep mysqld | grep -v grep
+if [[ \$MISS_SERVICE != '' ]]; then
+echo "ERROR: Service \"\$MISS_SERVICE\" are not up!"
+exit 1
+else
+echo "The local hadoop services are up!"
fi
-echo "\$NUM_JAVA_PROCS java servers and \$NUM_MYSQLD_PROCS mysqld
processes are running"
-
-jps | grep -v Jps
-
+NUM_MYSQLD_PROCS=\`ps -aef | grep \$USER | grep mysqld | grep -v grep | wc
-l\`
+if [[ \$NUM_MYSQLD_PROCS -ne 0 ]]; then
+echo "\$NUM_MYSQLD_PROCS mysqld processes are running"
+exit 0
+else
+echo "ERROR: mysqld process is not running!"
+exit 1
+fi
EOF
--- End diff --
See Line numbers 777 and above of this file. It has scripts like
swstartall, swstopall, swstarthbase, swstophbase etc.
> Change swstatus to give a simple "up" or "down" answer, instead of the
> current, confusing display
> -
>
> Key: TRAFODION-1609
> URL: https://issues.apache.org/jira/browse/TRAFODION-1609
> Project: Apache Trafodion
> Issue Type: Improvement
> Components: sql-general
>Reporter: Hans Zeller
>Assignee: Eason Zhang
>Priority: Minor
>
> The "swstatus" script used in cases where we did install_local_hadoop counts
> the MySQL processes and then simply invokes jps.
> It would be better if the script would check specifically for HMaster,
> NameNode, DataNode and maybe others, like NodeManager and ResourceManager. If
> any are missing, it should print out which ones are missing, otherwise it
> should just say something like "processes are up".
> It should also return an exit code of 0 if everything is up and 1 if some
> processes are missing.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)