Jan:
Jan Luehe wrote:
>
>
> Peter Fabian wrote On 04/12/07 10:22 AM,:
>
>> Jan Luehe wrote:
>>
>>>
>>>
>>> Peter Fabian wrote On 04/11/07 08:05 AM,:
>>>
>>>> Jan Luehe wrote:
>>>>
>>>>>
>>>>>
>>>>> Ashish Sahni wrote On 04/06/07 11:18 AM,:
>>>>>
>>>>>> Hi Jan/Larry,
>>>>>> So, the one-machine scenario should be good enough for me now but
>>>>>> for some reason http session data does not seem to be replicated
>>>>>> across instances
>>>>>> for either my custom-app of for the sample app - clusterjsp.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> To summarize:
>>>>>
>>>>> - custom-app was not marked as distributable, i.e., was missing
>>>>> <distributable/> in its web.xml
>>>>>
>>>>> - The clusterjsp sample app specifies "ha" instead of "replicated"
>>>>> as its
>>>>> persistence type in sun-web.xml.
>>>>>
>>>>> After adding <distributable/> to one, and changing "ha" to
>>>>> "replicated"
>>>>> in the other app, memory replication has been enabled.
>>>>>
>>>> changing ha to replicated gives this error when running ant all-app.
>>>> this is odd..
>>>>
>>>>
>>>> verify_common:
>>>> [echo] Verifying Process for ./build/assemble/ear/clusterjsp.ear
>>>> [exec] SEVERE: "DPL8015: Invalid Deployment Descriptors in
>>>> Deployment descriptor file WEB-INF/sun-web.xml in archive
>>>> [clusterjsp.war].
>>>> [exec] Line 11 Column 52 -- *Attribute "persistence-type" with
>>>> value "replicated" must have a value from the list "memory file
>>>> custom ha s1ws60 mmap "."*
>>>> [exec] SEVERE: "DPL8005: Deployment Descriptor parsing
>>>> failure : Attribute "persistence-type" with value "replicated" must
>>>> have a value from the list "memory file custom ha s1ws60 mmap "."
>>>> [exec] INFO:
>>>> [exec] # of Failures : 0
>>>> [exec] # of Warnings : 0
>>>> [exec] # of Errors : 2
>>>> [exec] INFO: Look in file "clusterjsp.ear.txt" for detailed
>>>> results.
>>>> [exec] Result: 2
>>>
>>>
>>>
>>> You need to upgrade clusterjsp's sun-web.xml to conform to
>>> sun-web-app_2_5-0.dtd,
>>> which is where we added "replicated" to the list of valid
>>> persistence-type
>>> values, i.e., use this DOCTYPE declaration:
>>>
>>> <!DOCTYPE sun-web-app PUBLIC "-//Sun Microsystems, Inc.//DTD
>>> Application Server 9.0 Servlet 2.5//EN"
>>> "http://www.sun.com/software/appserver/dtds/sun-web-app_2_5-0.dtd">
>>>
>>>
>> so this is actually bug in samples, right?
>
>
> If the sample is supposed to use in-memory replication as opposed to
> HADB,
> then yes, that would be a bug in samples (for not using latest DOCTYPE
> in bundled
> sun-web.xml).
>
> An alternative to upgrading the DOCTYPE in the sun-web.xml would be to
> remove its
> persistence-type attribute, in which case the persistence-type
> specified in the domain's
> <web-container-availability> element will be picked up, which defaults
> to "replicated"
> in the case of the "cluster" profile.
>
That's right, but I am not sure if the "validation"
of this attribute/element in sun-web.xml itself
is valid :)
This is because AFAIK, there can be any name to the
persistence type because we do have the SPI for
memory replication that one can provide an
implementation for and name it the way implementor
likes.
We made the change to domain's DTD to make this
value a CDATA and not an enumeration. Shouldn't
sun-web-dtd do a corresponding change?
Am I missing something?
Kedar
> Jan
>
>
>> P
>>
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe_at_glassfish.dev.java.net
> For additional commands, e-mail: dev-help_at_glassfish.dev.java.net
>