[jboss-jira] [JBoss JIRA] Commented: (JBAS-5163) Add JDK 6 version of ServerInfo.listThreadDump() taking advantage of new management classes

[Galder Zamarreno (JIRA)]

    [ http://jira.jboss.com/jira/browse/JBAS-5163?page=comments#action_12414405 ] 
            
Galder Zamarreno commented on JBAS-5163:
----------------------------------------

FAO Dimitris and anyone else interested in this JIRA:

In the next few minutes, I'll be uploading the following:

- a JBDS project called thread-deadlock containing source and 
deployable sar that generates two type of deadlocks: one based 
on synchronized statements and the other one based on 
java.util.concurrent.* classes. The aim of this project is to show how 
we can now find deadlocks and show them via the jmx console 
output. Both types of deadlocks will only be visible with JDK 6.

- Two PDFs containing a thread dump taken with JDK 5 showing standard 
thread dump and how deadlocks can be shown that are based on 
synchronized statements. Bear in mind that waiting locks as now shown 
as "- waiting on ..." so that they're closer to the signal based thread dumps 
that we're all used to.

- Two PDFs containing a thread dump taken with JDK 6 showing standard 
thread dump and how deadlocks based on synchronized statements and 
java.util.concurrent.* classes can be shown. JDK 6, *only for monitors 
acquired via synchronized statements* shows where these locks were 
acquired. You should be able to spot them by "- locked...". However, there's no 
API to be able to know where a java.util.concurrent.* lock class was locked, 
hence we only list the locked java.util.concurrent.* locks.

If anyone is interested, I can upload a source patch for people to have a look at.

> Add JDK 6 version of ServerInfo.listThreadDump() taking advantage of new management classes
> -------------------------------------------------------------------------------------------
>
>                 Key: JBAS-5163
>                 URL: http://jira.jboss.com/jira/browse/JBAS-5163
>             Project: JBoss Application Server
>          Issue Type: Feature Request
>      Security Level: Public(Everyone can see) 
>          Components: JMX/Web Console
>    Affects Versions:  JBossAS-5.0.0.Beta3, JBossAS-4.2.2.GA
>            Reporter: Galder Zamarreno
>         Assigned To: Galder Zamarreno
>             Fix For: JBossAS-5.0.1.CR1
>
>         Attachments: jmx-serverinfo.dump, kill-3.dump
>
>
> There're 4 ways to generate a thread dump. I have taken the output from the same run with 
> different methods:
> 1.- kill -3. Can be parsed by tools like https://tda.dev.java.net/ (simply, one of the most useful 
> tools I've found out there!). Example:
> "JBossLifeThread" prio=1 tid=0x0a690cc0 nid=0x1741 in Object.wait() [0x867f2000..0x867f30b0]
>         at java.lang.Object.wait(Native Method)
>         - waiting on <0x9247e200> (a java.lang.Object)
>         at java.lang.Object.wait(Object.java:474)
>         at org.jboss.system.server.ServerImpl$LifeThread.run(ServerImpl.java:940)
>         - locked <0x9247e200> (a java.lang.Object)	
> 2.- JMX via org.jboss.system.server.ServerInfo.listThreadDump(). Here's some output:
> Thread: JBossLifeThread : priority:5, demon:false, threadId:58, threadState:WAITING, lockName:java.l
> ang.Object at 10cff6b
> java.lang.Object.wait(Native Method)
> java.lang.Object.wait(Object.java:474)
> org.jboss.system.server.ServerImpl$LifeThread.run(ServerImpl.java:940)	
> Lock name is different to lock monitor achieved via kill -3. Doesn't looks like it prints previously locked 
> monitors. Cannot be parsed by TDA tools because it's in JBoss specific format. 
> 3.- Thread.getAllStackTraces().
> Haven't tested the output from this. What would this print? Couldn't find an MBean that uses/parses 
> the output from this, but I suspect it's the same as method 2, specially since 
> http://jira.jboss.com/jira/browse/JBAS-1448 has already been fixed.
> 4.- JConsole Threads view:
> Code:
> Name: JBossLifeThread
> State: WAITING on java.lang.Object at 10cff6b
> Total blocked: 0  Total waited: 1
> Stack trace: 
> java.lang.Object.wait(Native Method)
> java.lang.Object.wait(Object.java:474)
> org.jboss.system.server.ServerImpl$LifeThread.run(ServerImpl.java:940)	
> Seems to print same info as JBoss' ServerInfo.
> Bottom line: we should try generating thread dumps that look like kill -3 
> output to make them more readable than in their current format and to 
> benefit from existing tools that can parse standard kill -3 format.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira