Thanks @nacx.
I think I understand one part of your concern -- Azure \[1\] has a concept of
compute service, which can contain multiple virtual machines. I think that this
`azurecompute` provider tries really hard to hide that; I think that it even
makes an assumption that a compute service alw
Closed #257.
---
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/jclouds/jclouds-labs/pull/257#event-636857076
I've pushed this partial fix to master as
[76f90fc9](http://git-wip-us.apache.org/repos/asf/jclouds-labs/commit/76f90fc9).
I'm closing this PR as the proper fix will require a completely different
patch, but let's continue the discussion here or in the JIRA issue!
---
You are receiving this bec
The `listNodes` method [ignores `null`
deployments](https://github.com/jclouds/jclouds-labs/blob/master/azurecompute/src/main/java/org/jclouds/azurecompute/compute/AzureComputeServiceAdapter.java#L407)
too, so it presents the same issue.
The main problem here is that we map a jclouds `Node` to a
@nacx That sounds good to me. Shall I add a comment like "this is a temporary
workaround for JCLOUDS-1092 and should be removed once the issue is resolved
properly"?
I'm very much interested in a proper fix and I'm willing to test any proposed
fixes with our code. I can submit another PR with a
@Ladicek There is a design issue here regarding the list and get node methods,
but I think we could merge this as-is (without marking the JIRA issue as fixed)
to minimize its impact while we agree on a proper way to address the "null"
thing. WDYT?
---
You are receiving this because you are subs
It looks like the model currently maps a Deployment instance to a node, and
that one can be `null`.
@andreaturli @fmartelli @ilgrosso do you remember the reasoning behind this? Do
you have any suggestion on how to properly fix the issue that nodes that have
ongoing operations are not returned (n
@nacx Thanks for chiming in. I'm personally inclined to agree with you. But
when I was digging through the Git history, I found that the `getNode` method
was changed from a very simple implementation that makes sense on the first
sight to the current monstrosity in
https://github.com/jclouds/jc
Thanks for starting the discussion.
I'd rather go for the second option approach, and try to return the node even
when there are ongoing operations on it. Users should assume that if a node
exists, calling the `ComputeService.getNodemetadata` won't return null.
Otherwise there is not a clear co
@andreaturli What's your opinion on the JIRA in question, please?
---
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/jclouds/jclouds-labs/pull/257#issuecomment-209392952
I'm submitting this mainly to start a discussion around
https://issues.apache.org/jira/browse/JCLOUDS-1092
I'm by no means insisting on applying this exact fix, in fact I'm suggesting
another possible approach in the JIRA. I also didn't run the live tests (yet --
there seems to be quite a lot o
11 matches
Mail list logo