If I understand the proposal correctly, a user supplied resource name
submitted with a resource creation request will be added to a list and
kept for a period of 10 minutes? Is it expected that a create request
will result in an immediate discovery request?
I see that the resource key is part of the cache. Is it safe to assume
that this is the valid key that will be created upon discovery and thus
allow the discovered resource to be correlated with the created
resource?
Why can't the created resource be added to inventory if the create
request is successful and then allow discovery to merge the discovered
resource and the user created resource? This will prevent the need for
the cache.
--
Larry O'Leary
https://plus.google.com/u/0/112645929986009801513
On Mon, 2013-09-23 at 17:38 +0200, Thomas Segismont wrote:
Hi everyone,
I've created BZ1011095 [1] - Give created resources the name supplied by
the user even if the managed object has no "name" attribute
This bug follows BZ886119 [2] - "JON is using JNDI when referring to
child Datasource resource instead of specified Resource name".
As a reminder, when a created resource is discovered, it is given a name
by the agent plugin. If the underlying managed object has no "name"
attribute, the agent plugin gives a name to the resource which is
completely different than the user supplied name.
The proposal is to maintain a cache of the latest (10 minutes) resource
creation requests (CreateResourceReport holds the user supplied name),
and to look-up for an entry in this cache before sending the inventory
report to the server, in order to update the newly discovered resource
name, if needed.
The changes are in a bug branch [3] and I'd like to hear your feedback
(if any) before merging to master.
Thanks,
Thomas
[1]
https://bugzilla.redhat.com/show_bug.cgi?id=1011095
[2]
https://bugzilla.redhat.com/show_bug.cgi?id=886119
[3]
https://git.fedorahosted.org/cgit/rhq/rhq.git/commit/?h=bug/1011095&i...