This week The apache Ignite book becomes one of the top books of leanpub

This week The apache Ignite book becomes one of the top books of leanpub.


Centralize logs with logstash

Now a days logging is the essential part of the any application. Logging useful piece of information can easily help to find errors, fixing bug and much more. Modern application now scaling up in hundreds servers in cloud. Managing and monitoring logs of heterogeneous system is very challenging for any system administrator even more challenging for developers to fixing bugs. Last Friday evening we stared our testing with 3rd party products and stuck with a few bugs. As usual we first looked for the log and tried to find some hints to reproduce the bug. Here we got serious problems, our application scaled on few application servers such as Oracle GlassFish, Apache Tomcat. It was not a pleasant moments to search all over the server to find a few pieces of information. Here we have understand that we have to use any tools to manage, monitor and search logs. With my experience we have a few options:
1) Flume, Hadoop hdfs and ElasticSearch.
2) Kafka, Storm and Slor.
3) Logstash and Graylog2.
First option with Hadoop we have implemented in few cases but for the current project it seems very big gun. Second option also need some configuration and coding experience to build up the log managements tools from the scratch. My aim was to use something new and elegant which we can configured with less effort and easy to use. A few times i heard about logstash and decided to make a try. In the rest of the post i will describe how to install and configure logstash for centralizing log, i.e. collecting, aggregating and searching log. Most of the features of logstash is as follows:
1) Collecting log through agents
2) Aggregating logs
3) Shipping the logs in ElasticSearch
4) Web interface for searching logs
5) Open source
6) Everything in an one jar, nothing more.
7) Very well documented with examples

Take a look at the high level architecture of logstash:
For centralizing you have to need the followings components:
1) ElasticSearch
2) Redis
I have go through the getting started page and everything runs fines as a charm. Only one error i have got when tried to install Redis.
$ make
clang: error: no such file or directory: '../deps/hiredis/libhiredis.a'
clang: error: no such file or directory: '../deps/lua/src/liblua.a'
make[1]: *** [redis-server] Error 1
make: *** [all] Error 2
By googling in internet i have found the solution very easily as follows:
$ make
cd deps
make lua hiredis linenoise
and finalized the installation
$ make
cd $REDIS_CODE/src
In my cases i wanted to collect log from the Glassfish server.log and use the basic configuration for agent
input {
    file {
    type => "server"

    # Wildcards work, here :)
    path => [ "$DOMAIN_HOME/logs/*.log" ]

output {
  #stdout { codec => rubydebug }
  redis { host => "" data_type => "list" key => "crm" }
That's all. Happy coding and bloging.

1 comment :

elizabeth thomas said...

"Very Nice post..well written..

Here i would like to give a plug to my website..which your reader might find is as below

My site links. - Agile Certification .PMP Certification"