Thanks for this detailed explanation, Hugo! That was really useful info.

On Thu, Feb 20, 2014 at 3:02 AM, Hugo Trippaers <h...@trippaers.nl> wrote:

> OK, the systemvm status is making more sense now.
>
>
> For everybody just as confused as me by the current discussion let me try
> to summarize.
>
> SystemVM template build process:
>   Currently there is a build process for the systemvm template in the
> CloudStack source code. This build process is using veewee and virtual box
> and is started by running the command build.sh located in the
> tools/appliance directory. (parameter systemvmtemplate builds the 32bits
> template and command systemvm64template builds the 64bit image). This build
> is periodically executed by jenkins.buildacloud.org for each branch, so
> that is the authoritative source to get the latest version of the systemvm
> templates (both 32 and 64 bit).
>   However due to incompatibilities between vmware OVA import and the
> virtual box OVA export there is manual workaround to make the vmware
> template work (see CLOUDSTACK-4864 and the wiki page at
> https://cwiki.apache.org/confluence/display/CLOUDSTACK/Build+Your+Own+SystemVM+Templates).
> There is a fix for this (CLOUDSTACK-5883) that makes the automagically
> build template usable with VMware, this is confirmed to work with vSphere
> 5.1 for the 32bit template, but Sateesh reports there still are issues with
> the 64bit template. I've reopened the ticket 5883 to track this, but moved
> the priority from Blocker to Major as we have a workaround in a manual
> procedure.
>
> SystemVM template release process:
>   The systemvm template is generated using scripts in the source code and
> as part of the procedure parts of the CloudStack sources are copied onto
> the systemvm template when it is generated. The systemvm template is also
> an integral part of a release of CloudStack as CloudStack will not work as
> advertised if we do not make the systemvm template available to our users.
>   When a release is created a particular state of the source tree is fixed
> an a release artifact is put up for vote. However in the current process
> part of the release (the systemvm template) contains source code that might
> not be the same as the artifact that is supplied in the vote. The main
> reason for this is that the process of releasing the systemvm template is
> different from releasing the source code for CloudStack. The system vm
> template is released in advance to cloud.com and the process to trigger
> this is not completely clear to the community yet.
>   Combine this with the need for manual steps in the build process of the
> template and we have a situation where it is not transparent to the
> community what exact is in the released systemvm template at any given time.
>
> Current status of the systemvm:
>   The systemvm uploaded to cloud.com works with the current release under
> vote and workaround have been manually applied to ensure correct operation
> with VMware.
>   The current date field in the download URL of the systemvm template is
> confusing, but everyone agrees that there have been no changes to the
> relevant parts of the source code between the moment the current version of
> the systemvm-template on cloud.com was created and the source code in
> current release artifact.
>
> Recommedations/proposal for a next release (4.3.1 or 4.4):
>   Fix remaining issues in CLOUDSTACK-5883 so we have a fully automated
> build of the systemvm without manual steps.
>   With the next release generate a systemvm using the release artifact and
> upload that one to cloud.com for testing.
>   Master and release branches should be switched to use a download url
> that points to the latest successful build of the systemvm template for
> that branch to make sure all developments and tests are done using the
> latest sources available.
>
>
> With this i'm rescinding my -1 vote on this release candidate.
>
>
>
> Cheers,
>
> Hugo
>
>
>
> On 20 feb. 2014, at 09:23, Sateesh Chodapuneedi <
> sateesh.chodapune...@citrix.com> wrote:
>
> > Hugo,
> >
> > I manually prepared the 64-bit systemvm template via
> https://issues.apache.org/jira/browse/CLOUDSTACK-4864.
> >
> > In the JIRA bug I explained the issues with the default template and
> also attached the modified OVF file as well. I can add a wiki page
> explaining the exact changes done.  I looked at the changes you made as
> part of 5883 and observe the following issues which need to be resolved.
> Let me know.
> >
> > From the autogenerated template, the operating system section is
> referring to guest os id for vbox environment which is not correctly
> mapping to the right guest OS on ESXi hypervisor. This eventually resulting
> in the Debian 64bit system VM to be deployed as 32-bit otherLinux VM on
> ESXi.
> > ---
> >    <OperatingSystemSection ovf:id="96">
> >      <Info>The kind of installed guest operating system</Info>
> >      <Description>Debian_64</Description>
> >      <vbox:OSType ovf:required="false">Debian_64</vbox:OSType>
> > ---
> > Also it still has Vbox specific sections which may not make sense to
> vCenter. Ofcourse this would not block deployment of the template in
> itself. Removing this section ( <vbox:Machine>...</vbox:Machine>) would
> make it cleaner.
> >
> > Regards,
> > Sateesh
> >
> >
> >> -----Original Message-----
> >> From: Trippie [mailto:trip...@gmail.com] On Behalf Of Hugo Trippaers
> >> Sent: 20 February 2014 13:18
> >> To: <dev@cloudstack.apache.org>
> >> Subject: Re: [VOTE] Apache CloudStack 4.3.0 (sixth round)
> >>
> >> Animesh,
> >>
> >> Before we agree to use these templates, please detail how these are
> build. The current process is not transparent to the developer
> >> community so for now i'm -1 on using these templates. The process
> should be clear end-to-end before we can reasonably be expected to
> >> vote on a release as the templates are an integral part of the release
> and should include scripts and artifacts from the release being voted
> >> on.
> >>
> >> As demonstrated by the bugs found in CLOUDSTACK-5883 the procedure that
> is currently part of our community template build process is
> >> not the actual procedure used to build the systemVMs that are made
> available to the community. Before we proceed with the release we
> >> need to make sure that the systemvm build is part of the community
> process and thus part of the vote we are having. How can we
> >> reasonably expect somebody to vote on a release when part of the
> release is not done in the open?
> >>
> >> Cheers,
> >>
> >> Hugo
> >>
> >>
> >>
> >>
> >> On 20 feb. 2014, at 01:10, Animesh Chaturvedi <
> animesh.chaturv...@citrix.com> wrote:
> >>
> >>> Folks
> >>>
> >>> I want to clear up the confusion on system templates for 4.3
> >>>
> >>>
> >>> We had fixed the templates to be used in 4.3 and reflected the links
> in cloudstack database after uploading to download.cloud.com
> >>>
> >>>
> >>>
> >>> Here are the links:
> >>>
> >>>
> >>>
> >>> HyperV:
> http://download.cloud.com/templates/4.3/systemvm64template-2013-12-23-hyperv.vhd.bz2
> >>>
> >>> KVM:
> http://download.cloud.com/templates/4.3/systemvm64template-2014-01-14-master-kvm.qcow2.bz2
> >>>
> >>> VMWare:
> http://download.cloud.com/templates/4.3/systemvm64template-2014-01-14-master-vmware.ova
> >>>
> >>> Xen:
> http://download.cloud.com/templates/4.3/systemvm64template-2014-01-14-master-xen.vhd.bz2
> >>>
> >>>
> >>>
> >>> Unless there is a discussed change that requires updating the
> templates we use these published templates for 4.3.
> >>>
> >>>
> >>>
> >>> I am still waiting on feedback for this RC.
> >>>
> >>>
> >>>
> >>> Thanks
> >>>
> >>> Animesh
> >>>
> >>>
> >>>
> >>>
> >>> From: Animesh Chaturvedi
> >>> Sent: Tuesday, February 18, 2014 8:25 PM
> >>> To: dev@cloudstack.apache.org
> >>> Subject: [VOTE] Apache CloudStack 4.3.0 (sixth round)
> >>>
> >>>
> >>>
> >>> Hi All,
> >>>
> >>>
> >>>
> >>> I've created a 4.3.0 release, with the following artifacts up for a
> >>>
> >>> vote:
> >>>
> >>>
> >>>
> >>> Given that we have had multiple RC rounds with community testing a few
> times, and this VOTE has just one isolated fix in Nicira from
> >> previous RC, I would like to close the VOTE sooner by Thursday morning
> PST (36 hours). Please call out if you want to stick to 72 hour (3
> >> working day) period.
> >>>
> >>>
> >>>
> >>> Git Branch and Commit SH:
> >>>
> >>>
> https://git-wip-us.apache.org/repos/asf?p=cloudstack.git;a=shortlog;h=refs/heads/4.3
> >>> Commit: 307ad15bb68179129b8eadeaed115f5d088adfd9
> >>>
> >>>
> >>>
> >>> List of changes:
> >>>
> >>> New Features in 4.3:
> https://issues.apache.org/jira/issues/?filter=12325248
> >>>
> >>> Improvement in 4.3:
> https://issues.apache.org/jira/issues/?filter=12325249
> >>>
> >>> Issues fixed in 4.3
> https://issues.apache.org/jira/issues/?filter=12326161
> >>>
> >>> Known Issues in 4.3:
> https://issues.apache.org/jira/issues/?filter=12326162
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> Source release (checksums and signatures are available at the same
> >>>
> >>> location):
> >>>
> >>> https://dist.apache.org/repos/dist/dev/cloudstack/4.3.0/
> >>>
> >>>
> >>>
> >>> PGP release keys (signed using 94BE0D7C):
> >>>
> >>> https://dist.apache.org/repos/dist/release/cloudstack/KEYS
> >>>
> >>>
> >>>
> >>> Testing instructions are here:
> >>>
> >>>
> https://cwiki.apache.org/confluence/display/CLOUDSTACK/Release+test+procedure
> >>>
> >>>
> >>>
> >>> Vote will be open for 36 hours (Thursday morning PST)
> >>>
> >>>
> >>>
> >>> For sanity in tallying the vote, can PMC members please be sure to
> indicate "(binding)" with their vote?
> >>>
> >>>
> >>>
> >>> [ ] +1  approve
> >>>
> >>> [ ] +0  no opinion
> >>>
> >>> [ ] -1  disapprove (and reason why)
> >>>
> >
>
>


-- 
*Mike Tutkowski*
*Senior CloudStack Developer, SolidFire Inc.*
e: mike.tutkow...@solidfire.com
o: 303.746.7302
Advancing the way the world uses the
cloud<http://solidfire.com/solution/overview/?video=play>
*(tm)*

Reply via email to