I agree.
If one of us developers become confused then imagine what will happen to
a user?
Why not file a bug to change the message? Create-instance itself knows
what the problem is. It can jettison the boiler-plate canned message
and replace with something like:
Whenever a clustered-instance is created all other instances in the
cluster are notified. We were unable to notify all the other instances
in the cluster. They will learn about the new instance the next time
they are started. Here is the actual message with the details:
Instance c1n1i1 seems to be offline; Command was not replicated to that
instance
On 10/29/2010 2:57 PM, Chris Kasso wrote:
>
> There are explicit command and implicit commands. I guess when I run
> an explicit command like deploy and my app can't be deployed to an
> instance I would expect to see a warning. I know I'm deploying an
> app and I've targeted the cluster so I know what the expected outcome is.
>
> Then there are implicit commands - in this case I'm running the
> explicit command to create an instance on a node. Unbeknownst to me
> an implicit command is run which tries to talk to all the other
> instances. I don't expect nor do I probably care if that happens or
> not. So the warning creates confusion.
>
> To create a more usable product maybe we should develop some
> heuristics about when a warning is necessary and when it would just
> create confusion.
>
> Chris
>
> Tom Mueller wrote:
>> On 10/29/2010 4:39 PM, Chris Kasso wrote:
>>> But do I really want to see the warning? What am I supposed to do
>>> when I see it?
>>> If the system fixes itself later when the instance is started maybe
>>> this message should only appear in a log file?
>> This is the case with every "not replicated" message for any
>> command. The system fixes itself when the instance is started.
>>
>> However, if the instance is really up, but you get the warning
>> because the DAS can't communicate with it, then you might be
>> interested in the warning.
>>
>> Tom
>>
>>>
>>> Chris
>>>
>>> Tom Mueller wrote:
>>>> Each instance in a cluster needs to know about every other
>>>> instance (for GMS).
>>>> So when an instance is created, the register-instance is replicated
>>>> to all instances in the cluster.
>>>>
>>>> You should see the warning messages when you create a 50 instance
>>>> cluster:
>>>>
>>>> http://hudson.sfbay.sun.com/job/cluster-devtests-v3.1/241/consoleFull
>>>>
>>>> Tom
>>>>
>>>> On 10/29/2010 4:22 PM, Chris Kasso wrote:
>>>>> I've created a two node cluster where each node supports one
>>>>> instance.
>>>>>
>>>>> On node1 I created instance: c1n1i1
>>>>>
>>>>> When I create my second instance (c1n2i1) on node2 I see a warning:
>>>>>
>>>>> ouch: ./asadmin create-instance --node node2 --cluster c1 c1n2i1
>>>>> Command _create-instance-filesystem executed successfully.
>>>>> Port Assignments for server instance c1n2i1:
>>>>> JMX_SYSTEM_CONNECTOR_PORT=28686
>>>>> JMS_PROVIDER_PORT=27676
>>>>> HTTP_LISTENER_PORT=28080
>>>>> ASADMIN_LISTENER_PORT=24848
>>>>> JAVA_DEBUGGER_PORT=29009
>>>>> IIOP_SSL_LISTENER_PORT=23820
>>>>> IIOP_LISTENER_PORT=23700
>>>>> OSGI_SHELL_TELNET_PORT=26666
>>>>> HTTP_SSL_LISTENER_PORT=28181
>>>>> IIOP_SSL_MUTUALAUTH_PORT=23920
>>>>> The instance, c1n2i1, was created on host
>>>>> dhcp-santaclara22-1fl-west-10-132-181-70.usdhcp.oraclecorp.com
>>>>> WARNING : Instance c1n1i1 seems to be offline; Command was not
>>>>> replicated to that instance
>>>>>
>>>>> While this is only a warning I was still a bit confused by it. I was
>>>>> expecting the create-instance to complete without error. A message
>>>>> about not being able to replicate my create-instance command to my
>>>>> first instance:
>>>>>
>>>>> 1) Seems like it is coming out of left field.
>>>>> 2) Makes little sense to me.
>>>>>
>>>>> If I was deploying an app and that app could not be deployed to
>>>>> my first instance I would expect a message like that but it is odd
>>>>> to get it for create-instance.
>>>>>
>>>>> What's purpose does the message serve?
>>>>>
>>>>> Chris
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> 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
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: admin-unsubscribe_at_glassfish.dev.java.net
> For additional commands, e-mail: admin-help_at_glassfish.dev.java.net
>
--
Byron Nevins - Oracle Corporation
Home: 650-359-1290
Cell: 650-784-4123
Sierra: 209-295-2188