dejanmm at fastmail
Sep 15, 2009, 5:06 AM
Post #7 of 7
(644 views)
Permalink
|
Hi Kazutomo-san,
On Tue, Sep 15, 2009 at 07:08:53PM +0900, NAKAHIRA Kazutomo wrote:
> Hi, Dejan
>
> I'm sorry to have kept you waiting.
> The operation test was completed and the latest JBoss RA is attached.
>
> The only difference from your comment is the stop procedure.
>
> I was misunderstanding about sending SIGQUIT to the JVM process.
> It means "Do JVM thread dump" and never stops the JVM process.
Ah, OK. Now I understand. QUIT by default means kill process and
do coredump.
> And I redefined the stop procedure of JBoss as follows.
>
> 1. Execute shutdown.sh script
> 2. Waiting a JBoss process down until $SHUTDOWN_TIMEOUT seconds pass
> 3-1. If a JBoss process stops normally, then return $OCF_SUCCESS
> 3-2. If a JBoss process dose not stops, then goto step 4
> 4. Output JVM thread dump to $CONSOLE
> 5. Execute "pkill -TERM $PSTRING" command $KILL_TIMEOUT times.
> 6-1. If a JBoss process stops, then return $OCF_SUCCESS
> 6-2. If a JBoss process dose not stops, then goto step 7
> 7. Execute "pkill -KILL $PSTRING" command until a JBoss process is stopped
> 8-1. If a JBoss process stops, then return $OCF_SUCCESS
> 8-2. If a JBoss process dose not stops, then JBoss RA stop action will be
> timed-out
OK. This script is now in the repository. I just replace '..' by
'.' to source ${OCF_ROOT}/resource.d/heartbeat/.ocf-shellfuncs
properly and modified some metadata content.
Many thanks for the contribution.
Cheers,
Dejan
>
> Best Regards,
> NAKAHIRA Kazutomo
>
> NAKAHIRA Kazutomo wrote:
>> Hi, Dejan
>>
>> Thank you for the swift response.
>>
>> I corrected a JBoss RA referring to your comments.
>> And I am testing a JBoss RA now.
>>
>> On completion of the operation test, I post a latest RA to the list.
>> Please wait for a while...
>>
>> Best Regards,
>> NAKAHIRA Kazutomo
>>
>> Dejan Muhamedagic wrote:
>>> Hi Kazutomo-san,
>>>
>>> On Wed, Sep 09, 2009 at 02:46:52PM +0900, NAKAHIRA Kazutomo wrote:
>>>> Hi, Dejan
>>>>
>>>> I'm sorry I didn't get back to you sooner.
>>>> I took over takenaka-san's work.
>>>>
>>>> I corrected a JBoss RA referring to your comments.
>>>> The modification and my comments is written in the attached patch.
>>> Many thanks for your work.
>>>
>>> I made some changes and the result is attached. The changes are
>>> mainly in the documentation (meta-data). I also renamed the two
>>> timeout parameters: this way your intention should be clearer.
>>> There are other minor changes, please let me know if you find
>>> anything wrong or not clear.
>>>
>>> The stop procedure should be simplified by splitting the pkill
>>> -TERM and pkill -QUIT parts into two sequential processes (not
>>> interleaved as it is now). Then you could also make a short
>>> function with signal and timeout as input.
>>>
>>> The JAVA_HOME variable is sometimes already available in the
>>> environment, perhaps that could be used instead of requiring that
>>> the user sets that parameter.
>>>
>>> I find the "jboss_" prefix for parameters superfluous, for
>>> instance it should be just "name" and not "jboss_name" since it's
>>> obvious that the parameter is in the jboss primitive, but I leave
>>> that to you to decide.
>>>
>>> Cheers,
>>>
>>> Dejan
>>>
>>>> Best Regards,
>>>> NAKAHIRA Kazutomo
>>>>
>>>> Dejan Muhamedagic wrote:
>>>>> Hi Keisuke-san,
>>>>>
>>>>> On Thu, Jun 11, 2009 at 07:06:42PM +0900, Keisuke MORI wrote:
>>>>>> Hi,
>>>>>>
>>>>>> I'm posting an OCF RA for JBoss, which was originally posted by Stefan
>>>>>> to the users list, and includes some modifications as suggested by
>>>>>> Takenaka-san:
>>>>>>
>>>>>> http://www.gossamer-threads.com/lists/linuxha/users/53969
>>>>>>
>>>>>> Stefan,
>>>>>> Do you have any comment on this modification?
>>>>>>
>>>>>> Dejan,
>>>>>> Would you please review this RA if you have any chance?
>>>>> Script with comments attached. And a bit to reiterate:
>>>>>
>>>>> Why use resource specific logs? Shouldn't it be better to have
>>>>> all in one place, i.e. wherever lrmd is logging (lrmd will catch
>>>>> all stdin/stderr output). I don't know anything about jboss.
>>>>> Perhaps it is common to log to a separate file. In particular if
>>>>> the logging goes on while the process is running. It's just that
>>>>> start/stop should probably leave some trace in lrmd logs.
>>>>>
>>>>> This RA doesn't do enough validation for binaries, scripts, etc.
>>>>>
>>>>> Cheers,
>>>>>
>>>>> Dejan
>>>>>
>>>>>> If you are all OK, I will commit the RA to the -dev repository.
>>>>>>
>>>>>> Thanks,
>>>>>>
>>>>>> --
>>>>>> Keisuke MORI
>>>>>> _______________________________________________________
>>>>>> Linux-HA-Dev: Linux-HA-Dev[at]lists.linux-ha.org
>>>>>> http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
>>>>>> Home Page: http://linux-ha.org/
>>>>> ------------------------------------------------------------------------
>>>>>
>>>>> _______________________________________________________
>>>>> Linux-HA-Dev: Linux-HA-Dev[at]lists.linux-ha.org
>>>>> http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
>>>>> Home Page: http://linux-ha.org/
>>>> --
>>>> ----------------------------------------
>>>> NAKAHIRA Kazutomo
>>>> NTT DATA INTELLILINK CORPORATION
>>>> Open Source Business Unit
>>>> Software Services Integration Business Division
>>>>
>>>> ------------------------------------------------------------------------
>>>>
>>>> _______________________________________________________
>>>> Linux-HA-Dev: Linux-HA-Dev[at]lists.linux-ha.org
>>>> http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
>>>> Home Page: http://linux-ha.org/
>>
>>
>
>
> --
> ----------------------------------------
> NAKAHIRA Kazutomo
> NTT DATA INTELLILINK CORPORATION
> Open Source Business Unit
> Software Services Integration Business Division
_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev[at]lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
|
jboss
