Java Utilities

Here are some Java utilities (with common parameters) we use for troubleshooting Production issues.

Display a list of all Java processes.


Display a list of all Java processes with full command line information.

>jps -lmv

Print available jstat options.

>jstat -options

Display GC information for the given process id.

>jstat -gc {pid}

Create a thread dump, displayed on the console, for the given process id.

>jstack {pid}

Display heap summary for the given process id.

>jmap -heap {pid}

Display heap histogram for the given process id.

>jmap -histo:live {pid}

Create a heap dump in hprof format.

>jmap -dump:live,format=b,file=dump.hprof {pid}

Display available commands that can be sent to the Java process with the given id.

>jcmd {pid} help

Force a garbage collection.

>jcmd {pid}

Display uptime information.

>jcmd {pid} VM.uptime

Display performance information.

>jcmd {pid} PerfCounter.print

These utilities are great for scripting and headless environments. If you wish to work with graphical interface, check out VisualVM. Here’s a list of all the JDK Tools and Utilities.

Testing cron Jobs

1. Use crontab -e when possible when editing your cron table. It validates the entries on save.

2. Redirect output to a log file. For example:

 * * * * *  /path/ &> /path/script.log    

3. Run your script manually under the same user used in crontab.

4. If you are using run-parts (cron.daily, cron.hourly, etc), you can manually run run-parts.

run-parts /etc/cron.daily    

5. If you are not using run-parts and putting cron jobs directly in crontab, you can change the cron schedule. For example, instead of running the job daily, run it every minute for testing purposes.

6. Monitor the cron log file. Here is tailing the log in CentOS.

tail -f /var/log/cron    

7. Here are many possible reasons why crontab does not work.

Installing ElasticSearch with River JDBC on RedHat/CentOS

1. Install OpenJDK if it doesn’t already exist. Here we are running as root, so sudo is not required.

> yum install java-1.7.0-openjdk.x86_64

2. Install ElasticSearch using yum. The latest ElasticSearch version is 1.4 at the time of this writing.

> rpm --import
> vi /etc/yum.repos.d/elasticsearch.repo

# contents of elasticsearch.repo
name=Elasticsearch repository for 1.4.x packages

> yum install elasticsearch

3. Start ElasticSearch as a service, and tail the log.

> chkconfig --add elasticsearch
> service elasticsearch start
> tail -f /var/log/elasticsearch/elasticsearch.log

4. Install River JDBC. We are using the latest Rivers JDBC version, which is at the time of this writing.

> cd /usr/share/elasticsearch
> bin/plugin --install jdbc --url
cd plugins/jdbc/
> wget {url-to-jdbc-jar}/ojdbc6.jar
> service elasticsearch restart

5. Configure River. Do the initial, oneshot data load from Oracle. Replace the variables which are surrounded by {}. If you are unfamiliar with ElasticSearch indexes and types, they are similar to relational databases and tables.

        "type" : "jdbc",
        "jdbc" : {
            "url" : "jdbc:oracle:thin:@//{IP}:1521/{SERVICE}",
            "user" : "{USERNAME}",
            "password" : "{PASSWORD}",
            "sql" : [{
                "statement" : "{QUERY}"
            "index" : "{INDEX}",
            "type" :  "{TYPE}"

6. Load new or updated data every 10 minutes.

        "type" : "jdbc",
        "schedule" : "0 */10 * * * ?",
        "jdbc" : {
            "url" : "jdbc:oracle:thin:@//{IP}:1521/{SERVICE}",
            "user" : "{USERNAME}",
            "password" : "{PASSWORD}",
            "sql" : [{
                "statement" : "{QUERY where created_on > ? or modified_on > ?}",
                "parameter" : [ "$river.state.last_active_begin", "$river.state.last_active_begin" ]
            "index" : "{INDEX}",
            "type" :  "{TYPE}"