The conclusion regarding this thread from the admin iteam meeting is
that we will be looking into the ability to store the initial state of
an instance appropriately, so that if an instance has never been
started, such that we know that it is down, then replication will be
skipped for that instance, so there will be no message for the instance.
Created issue 14733:
https://glassfish.dev.java.net/issues/show_bug.cgi?id=14733
Tom
On 11/16/2010 3:47 PM, Tom Mueller wrote:
> The code isn't case by case. The message is generated by the
> InstanceCommandExecutor class - the same class is used for all
> commands. We would need to talk about using parameters on the
> @ExecuteOn or @Target annotations or something like that.
>
> Tom
>
> On 11/16/2010 3:25 PM, Chris Kasso wrote:
>>
>>
>> Tom Mueller wrote:
>>> Note that with issue 14284, which points out that sometimes this
>>> warning message is repeated multiple times by some commands (such a
>>> deploy), the text of the message is going to be changed thusly:
>>>
>>> WARNING : Instance c1i1_aha seems to be offline; ${command name} was
>>> not replicated to that instance
>> This is more of a nit but I would rather it say it could not contact
>> the instance rather than stating the instance is offline.
>>
>>>
>>> where ${command name} is the subcommand that was not replicated.
>>> This way, a deploy with --force would have messages with _disable,
>>> _deploy, etc. in them, rather than duplicates. In the case of
>>> creating an instance, the subcommand would be
>>> _register-instance-at-instance.
>>
>> Great, so now we start exposing a bunch of hidden commands the users
>> will know even less about.
>>>
>>> Also, note that it isn't just cluster instances that result in this
>>> message. It is output for stand-alone instances too when a
>>> domain-level resource is created, for example.
>> I think we need to look at these on a case by case basis and
>> determine if the message can be optimized or even needed for certain
>> commands.
>>>
>>> Tom
>>>
>>>
>>>
>>> On 11/16/2010 2:46 PM, Chris Kasso wrote:
>>>> Tom (and Admin team),
>>>>
>>>> As I mentioned last week I have a conflict this week so I can not
>>>> attend the admin meeting. In regards to the topic concerning
>>>> {create, delete}-instance and the offline instance message:
>>>>
>>>> 1) Do we feel strongly that the existence of this message is more
>>>> useful to users than confusing? I assume most users do not
>>>> understand the design of the product or the need for these commands
>>>> to be replicated behind the scenes.
>>>>
>>>> My reaction when I first saw this message was concern about what
>>>> went wrong, why it went wrong and what I now need to do to correct
>>>> the situation. (The reality was I did not need to be concerned and
>>>> I did not need to take any action)
>>>>
>>>> 2) Is this going to be a common scenario repeated and seen by many
>>>> - e.g. create a bunch of instances prior to starting them?
>>>> 3) What do we expect most users to do with this information? Why
>>>> is it useful to them? If it really is just an info message
>>>> concerning the overall health of the cluster (e.g. hey buddy, FYI,
>>>> not all instances are up - just thought you should know) then
>>>> shouldn't it say that instead of saying the command could not be
>>>> replicated (since the user isn't expected to replicate the command
>>>> in this case)?
>>>>
>>>> 4) If I can't convince folks to drop the message all together...
>>>>
>>>> WARNING : Instance c1i1_aha seems to be offline; Command was not
>>>> replicated to that instance
>>>>
>>>> ...then we need to improve this message so that it eases any
>>>> concerns a user may have that they now need to do something to
>>>> resolve the fact that the command was not replicated. Maybe
>>>> something like:
>>>>
>>>> INFO: The following cluster instance could not be contacted during
>>>> this operation. The instance will be notified when it becomes
>>>> available: c1i1_aha
>>>>
>>>> Chris
>>>>
>>>> Tom Mueller wrote:
>>>>> Here is the agenda for today's admin iteam meeting. Just one
>>>>> topic from last week, plus status reports.
>>>>> Please let me know if you have additional topics.
>>>>> http://wikis.sun.com/display/GlassFish/AdminIteam
>>>>> Tom
>>>>>
>>>>>
>>>>> Meeting Details:
>>>>>
>>>>> Tuesdays 3PM Pacific
>>>>> Room: PEZ, SCA 22, 1st floor, 1305
>>>>> US toll free: 866-682-4770
>>>>> Code: 1485398#
>>>>> Security Code: 4321#
>>>>> Email alias: admin_at_glassfish.dev.java.net
>>>>>
>>>>>
>>>>> Agenda - Nov 9, 2010
>>>>>
>>>>> * warning message from create-instance (about down instances)
>>>>>
>>>>> * General Topics
>>>>> o Quality update
>>>>> + Monitoring
>>>>> + Admin Infrastructure
>>>>> o Documentation update
>>>>>
>>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: admin-unsubscribe_at_glassfish.dev.java.net
>>>> For additional commands, e-mail: admin-help_at_glassfish.dev.java.net
>>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: admin-unsubscribe_at_glassfish.dev.java.net
>>> For additional commands, e-mail: admin-help_at_glassfish.dev.java.net
>>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: admin-unsubscribe_at_glassfish.dev.java.net
>> For additional commands, e-mail: admin-help_at_glassfish.dev.java.net
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: admin-unsubscribe_at_glassfish.dev.java.net
> For additional commands, e-mail: admin-help_at_glassfish.dev.java.net
>