whirr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrea Turli <andrea.tu...@gmail.com>
Subject Re: Virtualbox implementation
Date Mon, 08 Oct 2012 17:00:14 GMT
Thanks Patrick,

the ssh trick can totally help.

Parsing is not the only problem:
without the stub, all the domain classes are gone, so I need to recreate
all the pojo classes by hand (IMachine, INetworkAdapter etc)

Thoughts?

Andrea
Il giorno 08/ott/2012 17:46, "Patrick Debois" <patrick.debois@gmail.com> ha
scritto:

> ssh overhead can be minimized by using something like ssh contro session,
> where all your subsequent ssh connection happen over an established ssh
> tunnel. see ssh controlpath
>
> many of the ssh command like status have a way to have the output in
> machinereadable output, so parsing is pretty easy.
>
> On 08 Oct 2012, at 17:23, Andrea Turli <andrea.turli@gmail.com> wrote:
>
> > Thanks Brian for your help!
> >
> > Even if ssh approach looks reasonable to me as well I've 2 main concerns:
> > 1. the overhead introduced by each ssh call
> > 2. ssh invoking is fine, but retrieving (parsing) the responses can be
> > a a very hard and painful.
> >
> > Thoughts?
> >
> > Andrea
> >
> > On Mon, Oct 8, 2012 at 5:11 PM, Brian McCallister <brianm@skife.org>
> wrote:
> >> On Sun, Oct 7, 2012 at 6:40 AM, Andrea Turli <andrea.turli@gmail.com>
> wrote:
> >>> Sure I'll add this information to the jclouds-virtualbox documentation.
> >>>
> >>> Yesterday Ioannis was adding the vritualbox api to the jclouds-karaf
> >>> features that are used by jclouds-cli, so that we can use virtualbox
> >>> from the cli.
> >>> He noticed that vboxjws is LGPL2 licensed which is not compatible with
> >>> the ASL license. So he thinks that it is not possible package it
> >>> inside the cli itself.
> >>>
> >>> This is a big problem for jclouds and whirr in general, I guess, so we
> >>> need to find a solution.
> >>>
> >>> Ioannis was suggesting to ask Oracle to relicense the jar
> >>> Technically it is also possible to avoid the usage of vbox web service
> >>> and either ssh to host machine and run the vbox command on it
> >>> (vagrant/veewee approach) or use java native approach to run the C++
> >>> vbox commands.
> >>>
> >>> Thoughts?
> >>
> >> SSH into the remote machine and run vbox sounds pretty reasonable. The
> >> main alternative would be to write your own wrapper which invokes the
> >> command line tool, and listens on http(s) or something. In that case
> >> you still need to connect to a remote machine, and now you have
> >> another daemon process to monitor, etc.
> >>
> >> There are many decent ssh options available, I have found sshj (
> >> https://github.com/shikhar/sshj ) to be pleasant to use if you want
> >> pure ssh, or I have a little convenience wrapper around shelling out
> >> to command line ssh ( https://github.com/brianm/ssh-java-wrapper ) if
> >> you prefer that model.
> >>
> >> -Brian
> >>
> >>>
> >>> Andrea
> >>>
> >>> On Sat, Oct 6, 2012 at 9:35 PM, Adrian Cole <adrian.f.cole@gmail.com>
> wrote:
> >>>> Thanks for going through the process of getting the jar on the way to
> >>>> maven, Andrea!  Can you summarize what that was in case it needs to
be
> >>>> done again?
> >>>>
> >>>> -A
> >>>>
> >>>> On Sat, Oct 6, 2012 at 12:06 PM, Andrea Turli <andrea.turli@gmail.com>
> wrote:
> >>>>> I've uploaded the vboxjws-4.2.0 to
> >>>>>
> https://oss.sonatype.org/content/repositories/central-sync/org/virtualbox/vboxjws/4.2.0/
> >>>>>
> >>>>> So now jclouds-1.6.0-SNAPSHOT is using the last virtualbox version
> >>>>> which seems much better in terms of session management.
> >>>>>
> >>>>> Best,
> >>>>> Andrea
> >>>>>
> >>>>> On Tue, Sep 25, 2012 at 12:10 AM, Andrea Turli <
> andrea.turli@gmail.com> wrote:
> >>>>>> Hi all,
> >>>>>>
> >>>>>> I've updated the jclouds-virtualbox implementation to the latest
> release 4.2.0
> >>>>>>
> >>>>>> Here you can find more details about the status
> >>>>>> http://code.google.com/p/jclouds/issues/detail?id=384
> >>>>>>
> >>>>>> I'm now waiting for an answer from virtualbox developers about
> making
> >>>>>> available vboxjws-4.2.0.jar available on maven central,
> >>>>>> to push this new version.
> >>>>>>
> >>>>>> Cheers,
> >>>>>> Andrea
> >>>>>>
> >>>>>> On Wed, Sep 19, 2012 at 11:23 AM, Andrea Turli <
> andrea.turli@gmail.com> wrote:
> >>>>>>> Hi all,
> >>>>>>>
> >>>>>>> quick update on jclouds-virtualbox.
> >>>>>>>
> >>>>>>> The provider is still on labs, but after the last improvements,
> I've
> >>>>>>> been able to successfully start up to 30 guests on my laptop
> without
> >>>>>>> problems.
> >>>>>>>
> >>>>>>> I've also tested the jclouds-virtualbox provider with Apache
Whirr:
> >>>>>>>
> >>>>>>> bin/whirr launch-cluster --config recipes/zookeeper.properties
> >>>>>>> --private-key-file ~/.ssh/id_rsa_whirr --template
> >>>>>>> osFamily=UBUNTU,osVersionMatches=12.04
> >>>>>>>
> >>>>>>> where $ cat recipes/zookeeper.properties
> >>>>>>>
> >>>>>>> #
> >>>>>>> # Licensed to the Apache Software Foundation (ASF) under
one or
> more
> >>>>>>> # contributor license agreements.  See the NOTICE file distributed
> with
> >>>>>>> # this work for additional information regarding copyright
> ownership.
> >>>>>>> # The ASF licenses this file to You under the Apache License,
> Version 2.0
> >>>>>>> # (the "License"); you may not use this file except in compliance
> with
> >>>>>>> # the License.  You may obtain a copy of the License at
> >>>>>>> #
> >>>>>>> #     http://www.apache.org/licenses/LICENSE-2.0
> >>>>>>> #
> >>>>>>> # Unless required by applicable law or agreed to in writing,
> software
> >>>>>>> # distributed under the License is distributed on an "AS
IS" BASIS,
> >>>>>>> # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express
or
> implied.
> >>>>>>> # See the License for the specific language governing permissions
> and
> >>>>>>> # limitations under the License.
> >>>>>>> #
> >>>>>>>
> >>>>>>> #
> >>>>>>> # Deploy an Apache ZooKeeper cluster
> >>>>>>> #
> >>>>>>> whirr.cluster-name=zookeeper
> >>>>>>> whirr.instance-templates=3 zookeeper
> >>>>>>> whirr.bootstrap-user=toor:password
> >>>>>>>
> >>>>>>> with default ram size for the guests 1024 Mb, and everything
seems
> to
> >>>>>>> be in good shape.
> >>>>>>>
> >>>>>>> I think it is a good result :)
> >>>>>>>
> >>>>>>> If you are interested in having more details, feel free
to ping me.
> >>>>>>>
> >>>>>>> Thanks everyone for great team work particularly
> >>>>>>> Adrian Cole, David Alves, Mattias Holmqvist, Andrei Savu,
Tony
> >>>>>>> Batchelli, Andrew Bayer and Patrick Debois.
> >>>>>>>
> >>>>>>> Ciao,
> >>>>>>> Andrea
> >>>>>>>
> >>>>>>>
> >>>>>>> On Mon, May 21, 2012 at 5:42 PM, Adrian Cole <
> adrian.f.cole@gmail.com> wrote:
> >>>>>>>> haha baby steps.
> >>>>>>>>
> >>>>>>>> -A
> >>>>>>>>
> >>>>>>>> On Mon, May 21, 2012 at 8:38 AM, Andrea Turli <
> andrea.turli@gmail.com> wrote:
> >>>>>>>>>> I'd try using that
> >>>>>>>>>> in a different package.  Ex. a LOT of the code
is the same,
> regardless
> >>>>>>>>>> of which api is used (ex. image construction,
guest additions,
> etc),
> >>>>>>>>>> so maybe make a package org.jclouds.virtualbox.xpcom
and put
> things
> >>>>>>>>>> like a ComputeServiceAdapter that uses xpcom?
> >>>>>>>>>>
> >>>>>>>>>> wdyt?
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>> I agree with the idea of having a different package:
at the
> moment I'm using
> >>>>>>>>> a simple eclipse project to test things.
> >>>>>>>>> Unfortunately even if xpcom seems faster than SOAP
and doesn't
> require a
> >>>>>>>>> running vbox web server (good)
> >>>>>>>>> the session reliability is still a problem with
xpcom bridge as
> well.
> >>>>>>>>>
> >>>>>>>>> Maybe we could start a brave implementation of a
third bridge:
> native :)
> >>>>>>>>>
> >>>>>>>>> wdyt?
> >>>>>>>>>
> >>>>>>>>> --
> >>>>>>>>> You received this message because you are subscribed
to the
> Google Groups
> >>>>>>>>> "jclouds-dev" group.
> >>>>>>>>> To post to this group, send email to
> jclouds-dev@googlegroups.com.
> >>>>>>>>> To unsubscribe from this group, send email to
> >>>>>>>>> jclouds-dev+unsubscribe@googlegroups.com.
> >>>>>>>>> For more options, visit this group at
> >>>>>>>>> http://groups.google.com/group/jclouds-dev?hl=en.
> >>>>>>>>
> >>>>>>>> --
> >>>>>>>> You received this message because you are subscribed
to the
> Google Groups "jclouds-dev" group.
> >>>>>>>> To post to this group, send email to jclouds-dev@googlegroups.com
> .
> >>>>>>>> To unsubscribe from this group, send email to
> jclouds-dev+unsubscribe@googlegroups.com.
> >>>>>>>> For more options, visit this group at
> http://groups.google.com/group/jclouds-dev?hl=en.
> >>>>>
> >>>>> --
> >>>>> You received this message because you are subscribed to the Google
> Groups "jclouds-dev" group.
> >>>>> To post to this group, send email to jclouds-dev@googlegroups.com.
> >>>>> To unsubscribe from this group, send email to
> jclouds-dev+unsubscribe@googlegroups.com.
> >>>>> For more options, visit this group at
> http://groups.google.com/group/jclouds-dev?hl=en.
> >>>>
> >>>> --
> >>>> You received this message because you are subscribed to the Google
> Groups "jclouds-dev" group.
> >>>> To post to this group, send email to jclouds-dev@googlegroups.com.
> >>>> To unsubscribe from this group, send email to
> jclouds-dev+unsubscribe@googlegroups.com.
> >>>> For more options, visit this group at
> http://groups.google.com/group/jclouds-dev?hl=en.
> >>>
> >>> --
> >>> You received this message because you are subscribed to the Google
> Groups "jclouds-dev" group.
> >>> To post to this group, send email to jclouds-dev@googlegroups.com.
> >>> To unsubscribe from this group, send email to
> jclouds-dev+unsubscribe@googlegroups.com.
> >>> For more options, visit this group at
> http://groups.google.com/group/jclouds-dev?hl=en.
> >>
> >> --
> >> You received this message because you are subscribed to the Google
> Groups "jclouds-dev" group.
> >> To post to this group, send email to jclouds-dev@googlegroups.com.
> >> To unsubscribe from this group, send email to
> jclouds-dev+unsubscribe@googlegroups.com.
> >> For more options, visit this group at
> http://groups.google.com/group/jclouds-dev?hl=en.
> >>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message