geronimo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Matt Hogstrom <m...@hogstrom.org>
Subject Re: Java Service Wrapper for startup scripts
Date Wed, 10 Aug 2005 15:22:31 GMT
07@apache.org>
In-Reply-To: <42FA16C8.4050107@apache.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-MMS-Smtp-Program: Macallan-Mail-Solution; Version 4.6.0.1
X-MMS-Smtp-Auth: Authenticated As matt@hogstrom.org
X-MMS-Smtp-Mailer-Program: Macallan-Mail-Solution; Version 4.6.0.1

Jeff Genender wrote:

>
>
> Matt Hogstrom wrote:
>
>> Jeff Genender wrote:
>>
>>> Thanks Bruce and John.
>>>
>>> I would like to hear from more people on this as well as from David 
>>> Jencks and Dain too.
>>>
>>> Bruce Snyder wrote:
>>>
>>>> On 8/10/05, sissonj@insession.com <sissonj@insession.com> wrote:
>>>>
>>>>
>>>>> I noticed that Tomcat 5.5 uses the commons-daemon project and for
>>>>> non-Windows platforms the source (included with the Tomcat 
>>>>> binaries) must be
>>>>> compiled.
>>>>> http://jakarta.apache.org/tomcat/tomcat-5.5-doc/setup.html
>>>>>
>>>>> Maybe we could follow Tomcat's approach of not providing the 
>>>>> binaries for
>>>>> all platforms and leaving it up to the end user.  It is not as 
>>>>> though they
>>>>> can't start Geronimo without them, since the shell scripts should 
>>>>> always be
>>>>> available. 
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> After thinking further about this, I now agree with what Jeff
>>>> suggested and what you say above, John. I think we should simply
>>>> develop a set of generic shell scripts and init scripts and nothing
>>>> more (beyond a Windoze service, that is). We also need to encourage
>>>> users to contribute their customizations to these scripts for
>>>> additional platforms.
>>>
>>>
>>>
>> On a related topic, I've been experimenting with native libraries to 
>> extract information on a per request basis like CPU time.  For most 
>> platforms I should be able to get User / Kernel time for a request 
>> (I've been experimenting on Linux).  As we look at the potential 
>> platforms that can be supported the binary issue has been nagging at 
>> me as I'm not sure how to support that.  One part of me says that we 
>> should provide a binary for a given platform (as donated by folks) as 
>> well as source for those that may be ahead of the game / want to 
>> experiment and tweak further.  I don't have the answers but I think 
>> as Geronimo matures this will be an issue that we'll have to deal 
>> with.  In theory Java should be portable but portability is an 
>> illusion that we need to archive by taking on the responsibility by 
>> making moving around easy.
>>
>
> I don't necessarily agree here.  In the instance of this thread, we 
> are talking about start/stop daemons, which easily can be done outside 
> of the binary world (except for Windows).  I don't see the binary as a 
> critical path, and it is more of a restriction to what platforms will 
> ultimately use our product w/o our potential intervention.  IMHO, lets 
> let Sun (or other JVM vendor) deal with the binary issues, and we deal 
> with getting our stuff to run in a cross-platform perspective (i.e. 
> Java only).
>
> If we have a critical path (a function Geronimo absolutely cannot live 
> without), then I would be willing to acquiesce on this point.
>
Right, my point is off topic but related insofar as binaries and 
multiple-platforms are concerned so I thought I would throw it in to 
paint the landscape.  At least from my point of view its not an issue 
today but maybe in the future.

>>>>  
>>>>
>>>>> Is anyone familiar with the pros and cons of commons-daemon over Java
>>>>> Service Wrapper and why the Tomcat project chose to use 
>>>>> commons-daemon? 
>>>>
>>>>
>>>>
>>>>
>> Not yet.
>>
>>>>
>>>> Not I.
>>>> Bruce 
>>>
>>>
>>>
>>>
>>>
>>>
>>
>>
>




Mime
View raw message