Not being able to execute the operation is a bug. (I believe there's a
JIRA filed by you for that that's scheduled for beta3.)
As for the expected behavior once the bug is fixed, the deployment list
is part of the model controlled by the DC. When the host restarts, it
gets that part of the model from the DC and picks up the current state.
On 4/13/11 2:57 PM, Heiko W.Rupp wrote:
Hi,
I had a .war deployed on a server-group that spans two boxes.
Switched the slave box off.
Tried to remove the deployed war, got
{
"outcome" => "failed",
"result" => {"host-failure-descriptions" =>
[("pintsize" => [
("java.lang.RuntimeException" => "Failed to execute operation
"),
("java.net.ConnectException" => "Operation timed out")
])]}
}
My understanding was so far, that this request is somehow queued on the DC and
delivered to the (slave) HC when this comes online again.
Is such a queueing planned for the future? What other ways to get rid of the
deployment do I have in a scenario where the slave will never come up again
(e.g. hardware defect without a good backup)
Heiko
--
Brian Stansberry
Principal Software Engineer
JBoss by Red Hat