Newlisp tracks jobtracker status

The basic idea is to use newlisp to regularly download the jobtracker page, use regular expressions to parse the table elements in HTML, and then obtain the latest mapreduce status.

Each time the status data is obtained, it is stored in the MySQL database, and tableau is used to present the status of the mapreduce cluster with reports.

This is the data of the jobtracker site.

This is a report drawn by tableau.

In this way, the computing pressure of the hadoop cluster can be displayed in data visualization mode.

The following is the newlisp code. It mainly uses regular expressions to parse HTML and write it into the database using the MySQL module.

#!/usr/bin/newlisp(load "mysql.lsp")(define (check-args)  (print "args: ")  (println (main-args))  (set ‘args-length (length (main-args)))  (if (< args-length 3)      (begin(println "the number of args must be 3 or 4, e.g. ./job.lsp jobtracker.bigdata.cn 8080")(exit))))(define (parse-args)  (set ‘domain (main-args 2))  (if (= 4 args-length)      (set ‘port (main-args 3)))  (if (= 3 args-length)      (set ‘port "80"))  (set ‘url (string "http://" domain ":" port "/jobtracker.jsp"))  (println (string "job tracker site is located at " url)))(define (access-job-tracker-site)  (set ‘page-content (get-url url))  (extract-tables page-content)  )(define (extract-summary-table table)  (if (regex "Running Map Tasks" table)      table)  )(define (get-number td)  (set ‘r ((regex "<td>(.*)</td>" td) 3))  (if (find "<a href" r)      ((regex "<a.*?>(.*)</a>" r) 3)    r))(define (remove-td tds)  (set ‘result ‘())  (dolist (td tds)    (push (get-number td) result -1)    )  result  )(define (parse-summary-table table)  (set ‘all-tds (find-all "<td>[\\s\\S]*?</td>" table))  (set ‘all-summary-values (remove-td all-tds))  )(define (extract-tables html-content)  (set ‘all-tables (find-all "<table[^>]+>[\\s\\S]*?</table>" html-content))  (dolist (table all-tables)    (if (extract-summary-table table)(set ‘summary-table table))        )  (parse-summary-table summary-table)  )(define (write-summary-to-mysql all-summary-values)  (println all-summary-values)  (set ‘mysql-instance (Mysql))  (println "mysql-instance: " mysql-instance)  (set ‘mysql-host  "10.100.10.10")  (set ‘mysql-port 3306)  (set ‘mysql-user "user")  (set ‘mysql-pwd "123456")  (set ‘mysql-db "bigdata_data_market")  (set ‘job-tracker-summary-table "hadoop_job_tracker_summary")  (:connect mysql-instance mysql-host mysql-user mysql-pwd mysql-db mysql-port)  (:query mysql-instance "SET character_set_client = utf8;")  (set ‘insert-summary-sql (format "insert into %s (collect_time,running_map_tasks,running_reduce_tasks,total_submissions,nodes,occupied_map_slots,occupied_reduce_slots,reserved_map_slots,reserved_reduce_slots,map_task_capacity,reduce_task_capacity,average_tasks_per_node,blacked_listed_nodes,gray_listed_nodes,exclueded_nodes) values (now(),%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s)"   job-tracker-summary-table   (all-summary-values 0)   (all-summary-values 1)   (all-summary-values 2)   (all-summary-values 3)   (all-summary-values 4)   (all-summary-values 5)   (all-summary-values 6)   (all-summary-values 7)   (all-summary-values 8)   (all-summary-values 9)   (all-summary-values 10)   (all-summary-values 11)   (all-summary-values 12)   (all-summary-values 13)))  (println insert-summary-sql)  (:query mysql-instance insert-summary-sql)  );; main logic starts now(check-args)(parse-args)(access-job-tracker-site)(write-summary-to-mysql all-summary-values)(exit)

Newlisp tracks jobtracker status