deltacloud-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From David Lutterkort <>
Subject Re: jruby support: which drivers?
Date Tue, 04 Jan 2011 00:08:05 GMT
On Fri, 2010-12-31 at 17:49 +0100, Adrian Cole wrote:
> I've found a few issues, noting I'm only testing read-only at this
> point.  A lot of issues revolve around state consistency, which I sent
> an email wrt earlier.  There's also a few cases where fetching
> something by href returns the wrong value.  I think at least the state
> question needs to be cleared up before we look into other issues.

Are you working off the 0.1.0 gems or the head of the svn/git repo ?

>     * I've sent a couple notes to the list, but I also get "Bad
> Request" attempting to start an instance

Sounds like a bug.

> rackspace: INCONSISTENT
>     * states are out of sync with latest mock

As noted in the other mail thread, states still need more cleanup. Part
of the problem here is two conflicting schools of thought: initially,
the drivers were written to expose as much detail about the cloud's
native instance lifecycle as possible, later on there has been some
effort to simplify the set of instance states, and with them the FSM for
the instance lifecycle that is exposed.

I've created an issue in Jira to track this [1]

> opennebula: NOT SURE
>     * got the following error, but I'm not sure my local runtime is
> setup: RuntimeError - CloudClient::Error is not a valid input stream.

Do you have a stacktrace ? From looking at the code, looks like a simple
error ('return' instead of 'raise' in

>     * see rackspace: ec2 has a STOPPING state.  Is this expected?

Part of the state mess.

> Also, fetching an availability zone by href returns the wrong
> response.

A regression introduced by svn commit r1044338 (the capabilities stuff)

>     * uses power shell

We are in the process of rewriting the RHEV-M drive against the new
RHEV-M REST API - I wouldn't worry about RHEV-M until that's done.

>     * see rackspace: terremark has a SHUTTING_DOWN state.  Is this
> expected? also a problem with "undefined local variable or method
> `vdc_id'" for images and instances.

Do you have a stacktrace for that ?

> rimuhosting: BROKE, INCONSISTENT
>     * see rackspace: rimuhosting has a SHUTTING_DOWN state.  Is this
> expected? also everything but realm queries break on doing .each on
> something that is null

We probably need to get Ivan to update that driver.



View raw message