cloudstack-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Chip Childers <chip.child...@sungard.com>
Subject Re: [VOTE] Apache Cloudstack 4.0.0-incubating Release, third round
Date Mon, 29 Oct 2012 17:29:38 GMT
On Mon, Oct 29, 2012 at 10:10 AM, Chip Childers
<chip.childers@sungard.com> wrote:
> On Sun, Oct 28, 2012 at 8:01 PM, Chip Childers
> <chip.childers@sungard.com> wrote:
>> On Sun, Oct 28, 2012 at 1:53 PM, Noah Slater <nslater@apache.org> wrote:
>>> Actually, sorry, I probably gave that +1 a *little* too soon.
>>>
>>> Can someone make the case for "waf" not being a release blocker?
>>>
>>> Everything else can probably slide.
>>
>> Waf needs to go.  No doubt about it.  However, I'm not certain it's a
>> blocker for a release.  Here's my reasoning:
>>
>> Waf is being removed from the project entirely (ongoing work in
>> origin/maven-to-rpm), so we're not doing something that isn't already
>> targeted for being undone as fast as possible.   The important thing
>> to me is understanding that work to remove waf is taking longer than
>> the community desire to get a release completed at the end of the
>> summer.  It's one line of bytecode within a file that is not core to
>> the project itself (it's only a build utility), that has a clearly
>> marked license for file itself (so distribution is OK), that is only
>> used as a quick way to create packages for Linux distros (not to do
>> the actual project compilation), and that will be ignored by any
>> distro packagers that want to create proper RPMs / DEBs anyway.
>>
>> Thoughts?
>>
>> -chip
>
> Noah has raised the question on general@i.a.o, and I'm holding off on
> closing this vote to get some comments back from the incubator.
>
> -chip

One more follow up.  The discussion on general@i.a.o seems to be
reaching the consensus that we will be ok with the waf file (but
really should remove it).  I'm going to wait until tomorrow morning US
ET to close the vote down, giving the folks on general@ time to raise
any additional concerns.

>
>>> On 28 October 2012 17:51, Noah Slater <nslater@apache.org> wrote:
>>>
>>>> It is frustrating because everything is so slow, and my (main) computer is
>>>> practically unusable while it is running. How feasible would it be to
>>>> provide this DevCloud thing as an AMI? I would much prefer to do this in
>>>> the cloud.
>>>>
>>>> Okay, I made it to this step:
>>>>
>>>> Additional Testing
>>>>
>>>> The instance is removed from the web console.
>>>>
>>>> And this is what I see on the terminal:
>>>>
>>>>   [sshexec]      [java] WARN  [cloud.consoleproxy.ConsoleProxyManagerImpl]
>>>> (AgentManager-Handler-6:) VM 3 lost host info, failed authentication request
>>>>   [sshexec]
>>>>   [sshexec]      [java] WARN  [cloud.consoleproxy.ConsoleProxyManagerImpl]
>>>> (AgentManager-Handler-8:) VM 3 lost host info, failed authentication request
>>>>   [sshexec]
>>>>   [sshexec]      [java] WARN  [cloud.consoleproxy.ConsoleProxyManagerImpl]
>>>> (AgentManager-Handler-9:) VM 3 lost host info, failed authentication request
>>>>   [sshexec]
>>>>   [sshexec]      [java] INFO  [cloud.vm.UserVmManagerImpl]
>>>> (UserVm-Scavenger-1:) Found 1 vms to expunge.
>>>>   [sshexec]
>>>>   [sshexec]      [java] INFO  [network.security.SecurityGroupManagerImpl]
>>>> (UserVm-Scavenger-1:) Disassociated 1 network groups  from uservm 3
>>>>   [sshexec]
>>>>   [sshexec]      [java] INFO  [cloud.ha.HighAvailabilityManagerImpl]
>>>> (HA-Worker-1:work-2) Processing HAWork[2-HA-4-Stopped-Scheduled]
>>>>   [sshexec]
>>>>   [sshexec]      [java] INFO  [cloud.ha.HighAvailabilityManagerImpl]
>>>> (HA-Worker-1:work-2) HA on VM[DomainRouter|r-4-TEST]
>>>>   [sshexec]      [java] INFO  [cloud.ha.HighAvailabilityManagerImpl]
>>>> (HA-Worker-1:work-2) VM VM[DomainRouter|r-4-TEST] has been changed.
>>>>  Current State = Running Previous State = Stopped last updated = 10
>>>> previous updated = 7
>>>>   [sshexec]      [java] INFO  [cloud.ha.HighAvailabilityManagerImpl]
>>>> (HA-Worker-1:work-2) Completed HAWork[2-HA-4-Stopped-Scheduled]
>>>>
>>>> However, from devcloud, I see this:
>>>>
>>>> root@devcloud:~# xe vm-list
>>>> uuid ( RO)           : 83fdea03-9d67-d5dc-e090-0acba65e7926
>>>>      name-label ( RW): r-4-TEST
>>>>     power-state ( RO): running
>>>>
>>>>
>>>> uuid ( RO)           : f01b195e-074b-e8f2-f8b1-b2c493070c8a
>>>>      name-label ( RW): Control domain on host: devcloud
>>>>     power-state ( RO): running
>>>>
>>>>
>>>> uuid ( RO)           : 2ab8057a-767e-b700-53e3-861f2e23823a
>>>>      name-label ( RW): s-1-TEST
>>>>     power-state ( RO): running
>>>>
>>>>
>>>> uuid ( RO)           : 8b27e755-4486-e27e-ee7a-6e5bc6ea768f
>>>>      name-label ( RW): v-2-TEST
>>>>     power-state ( RO): running
>>>>
>>>> This looks, to my untrained eyes, like the VM might still be running, and
>>>> in fact that there might be more than one?
>>>>
>>>> Okay, at this point, if someone can clarify that issues (or not?) I have
>>>> run into with the functional tests are not release blockers, I am happy to
>>>> just shut up about them and hope that testing is made easier for the next
>>>> release.
>>>>
>>>> Just moving on now to some more thorough testing.
>>>>
>>>> My manual license check seems satisfactory.
>>>>
>>>> We're still shipping the "waf" binary in the root of our source, and I
>>>> don't understand why. I know I was asked about this during the last round.
>>>> And to be honest, I do not know whether it's a blocker. Can someone clarify
>>>> for me why it is used, and why we can't ship the source instead?
>>>>
>>>> Manual file type test: OK
>>>>
>>>> Okay, going through some files by hand.
>>>>
>>>> The following files do not have copyright headers:
>>>>
>>>> setup/apidoc/generateadmincommands.xsl
>>>> setup/apidoc/generatecommand.xsl
>>>> setup/apidoc/generatedomainadmincommands.xsl
>>>> setup/apidoc/generatetoc_footer.xsl
>>>> setup/apidoc/generateusercommands.xsl
>>>>
>>>> I do not consider this a release blocker, but these should be fixed (where
>>>> possible) for the next release.
>>>>
>>>> I almost raised a blocked with the files I found under tools/marvin,
>>>> before I remembered our discussion about them. Heh!
>>>>
>>>> Okay, manually checked all the other manually flagged files, and they all
>>>> check out in pom.xml or LICENSE.
>>>>
>>>> I am +1 on this release.
>>>>
>>>> For the next release, please can we:
>>>>
>>>> 1) Prepare a DecCloud AMI
>>>> 2) Improve the test procedure
>>>> 3) Fix up the XSL license headers
>>>>
>>>> Thanks!
>>>>
>>>> On 28 October 2012 17:00, Sebastien Goasguen <runseb@gmail.com> wrote:
>>>>
>>>>>
>>>>> On Oct 28, 2012, at 5:48 PM, Noah Slater <nslater@apache.org> wrote:
>>>>>
>>>>> > I was expecting a VM for the storage thing. But that never appeared.
I
>>>>> did
>>>>> > see a line for the VM I created. Either way, this testing step should
>>>>> > indicate *clearly* what I should be looking for before moving on.
>>>>> >
>>>>> > As it stands, CloudStack itself is now refusing to answer HTTP
>>>>> requests, so
>>>>> > I am unable to complete the configuration change step. I am close
to
>>>>> giving
>>>>> > up. This has been an incredibly frustrating experience.
>>>>>
>>>>> It should not be that frustrating.
>>>>>
>>>>> It's not clear to me where you stand and what you are doing.
>>>>> I presume you are following the test procedure pointed by Chip.
>>>>> I also presume you successfully checked the keys etc..and managed to
>>>>> build and deploy and are running CS via rdebug.
>>>>> I also presume that you managed to configure the infra by hand…did
>>>>> everything turn green ? Is the zone enabled ?
>>>>>
>>>>> Can you give us more info or maybe ask on IRC ?
>>>>>
>>>>> -sebastien
>>>>>
>>>>> >
>>>>> > On 28 October 2012 16:45, Sebastien Goasguen <runseb@gmail.com>
wrote:
>>>>> >
>>>>> >>
>>>>> >> On Oct 28, 2012, at 4:41 PM, Noah Slater <nslater@apache.org>
wrote:
>>>>> >>
>>>>> >>> Okay,
>>>>> >>>
>>>>> >>> I did a rebuild and the menu populated. Very strange.
>>>>> >>>
>>>>> >>> Now I have two more problems.
>>>>> >>>
>>>>> >>> From the test procedure:
>>>>> >>>
>>>>> >>> Wait for secondary storage VM coming up.
>>>>> >>>> This may take a bit...
>>>>> >>>
>>>>> >>>
>>>>> >>> No clue about this. How do I know when the storage VM has
come up?
>>>>> >>
>>>>> >> In the Infrastructure tab, do you see system VMs in running
state ?
>>>>> >> Check my video for help :)
>>>>> >> https://vimeo.com/52150218
>>>>> >>
>>>>> >> PS: I ran the testing procedure on my mac air, and it was exceptionally
>>>>> >> fast. I was amazed.
>>>>> >>
>>>>> >>
>>>>> >>>
>>>>> >>> I checked under instances and storage and nothing changed
for about 10
>>>>> >>> minutes.
>>>>> >>>
>>>>> >>> Last three lines in console are:
>>>>> >>>
>>>>> >>> [sshexec]      [java] WARN  [network.element.VpcVirtualRouterElement]
>>>>> >>> (secstorage-1:) Network Ntwk[203|Storage|4] is not associated
with any
>>>>> >> VPC
>>>>> >>> [sshexec]      [java] INFO  [network.security.SecurityGroupListener]
>>>>> >>> (AgentManager-Handler-2:) Received a host startup notification
>>>>> >>> [sshexec]      [java] INFO
>>>>>  [cloud.consoleproxy.ConsoleProxyManagerImpl]
>>>>> >>> (AgentManager-Handler-2:) Successfully sent out command
to start HTTP
>>>>> >>> handling in console proxy agent
>>>>> >>>
>>>>> >>> I decided to proceed to the next stage, on the off chance
it had
>>>>> worked.
>>>>> >>>
>>>>> >>> I get stuck on this step:
>>>>> >>>
>>>>> >>> In step 2, select "tiny Linux"
>>>>> >>>
>>>>> >>>
>>>>> >>> There are no templates to select from!
>>>>> >>>
>>>>> >>>
>>>>> >>
>>>>> https://www.evernote.com/shard/s18/sh/1a151896-8cbb-41c8-93d2-d844e6461146/08efa5dddbd86cb96d7895692f9ec20d
>>>>> >>>
>>>>> >>
>>>>> https://www.evernote.com/shard/s18/sh/ee10c37e-dad1-454f-a43f-e64d9ef7d9e3/bce4465e03169a3883f825fbb6f43cd1
>>>>> >>>
>>>>> >>
>>>>> https://www.evernote.com/shard/s18/sh/58221024-608f-4a7f-9531-53c267532efc/cbdb8204c77471ae2fa194355af40430
>>>>> >>>
>>>>> >>> I am unable to proceed.
>>>>> >>>
>>>>> >>> P.S. Testing CloudStack on my MBA is exceptionally slow.
Is there any
>>>>> way
>>>>> >>> to do the whole thing on or something? I practically cannot
do
>>>>> anything
>>>>> >>> else on my computer while testing.
>>>>> >>>
>>>>> >>> On 28 October 2012 00:12, Chip Childers <chip.childers@sungard.com>
>>>>> >> wrote:
>>>>> >>>
>>>>> >>>> Yeah.  Funny enough, that IS the infra menu option.
>>>>> >>>>
>>>>> >>>> I've never seen this before.
>>>>> >>>>
>>>>> >>>> I assume you are logged in as admin / password?  Can
you do a fresh
>>>>> >>>> build (start with "ant clean-all build-all")?
>>>>> >>>>
>>>>> >>>> - chip
>>>>> >>>>
>>>>> >>>> Sent from my iPhone.
>>>>> >>>>
>>>>> >>>> On Oct 27, 2012, at 5:06 PM, Noah Slater <nslater@apache.org>
wrote:
>>>>> >>>>
>>>>> >>>>> I doubt it. Looks like part of the application code.
>>>>> >>>>>
>>>>> >>>>> My real problem is that I am unable to complete
this step, as there
>>>>> is
>>>>> >>>>> nothing labeled "Infrastructure" in the UI.
>>>>> >>>>>
>>>>> >>>>> On 27 October 2012 19:04, Chip Childers <chip.childers@sungard.com>
>>>>> >>>> wrote:
>>>>> >>>>>
>>>>> >>>>>> Comments below:
>>>>> >>>>>>
>>>>> >>>>>> - chip
>>>>> >>>>>>
>>>>> >>>>>> Sent from my iPhone.
>>>>> >>>>>>
>>>>> >>>>>> On Oct 27, 2012, at 1:28 PM, Noah Slater <nslater@apache.org>
>>>>> wrote:
>>>>> >>>>>>
>>>>> >>>>>>> Sig: OK
>>>>> >>>>>>> Hashes: OK
>>>>> >>>>>>> Source checks: OK
>>>>> >>>>>>> RAT: OK
>>>>> >>>>>>> Build: OK
>>>>> >>>>>>>
>>>>> >>>>>>> However, during this:
>>>>> >>>>>>>
>>>>> >>>>>>> mvn -P deps
>>>>> >>>>>>>
>>>>> >>>>>>> I noticed line like this:
>>>>> >>>>>>>
>>>>> >>>>>>> [INFO] Installing
>>>>> >>>>>>>>
>>>>> >> /private/tmp/cloudstack/apache-cloudstack-4.0.0-incubating-src/pom.xml
>>>>> >>>>>> to
>>>>> >>>>>>
>>>>> >>>>
>>>>> >>
>>>>> /Users/nslater/.m2/repository/org/apache/cloudstack/cloudstack/4.0.0-incubating-SNAPSHOT/cloudstack-4.0.0-incubating-SNAPSHOT.pom
>>>>> >>>>>>>> [INFO]
>>>>> >>>>>>>>
>>>>> >>>>>>>> [INFO]
>>>>> >>>>>>>>
>>>>> >>>>
>>>>> ------------------------------------------------------------------------
>>>>> >>>>>>>> [INFO] Building Apache CloudStack Utils
4.0.0-incubating-SNAPSHOT
>>>>> >>>>>>>> [INFO]
>>>>> >>>>>>>>
>>>>> >>>>
>>>>> ------------------------------------------------------------------------
>>>>> >>>>>>>
>>>>> >>>>>>>
>>>>> >>>>>>> Any idea why that says incubating-SNAPSHOT?
>>>>> >>>>>>
>>>>> >>>>>> That's the default build version. If we do a
binary release, it
>>>>> would
>>>>> >>>>>> be changed via the maven release tools.
>>>>> >>>>>>
>>>>> >>>>>>>
>>>>> >>>>>>> A bigger problem is that I am stuck with
the additional testing.
>>>>> >>>>>>>
>>>>> >>>>>>> Specifically, this step:
>>>>> >>>>>>>
>>>>> >>>>>>> Click "Infrastructure", click the "View
All" button in the "Zones"
>>>>> >>>>>> display
>>>>> >>>>>>>> box, click "add zone"
>>>>> >>>>>>>
>>>>> >>>>>>>
>>>>> >>>>>>> Here's what my console looks like:
>>>>> >>>>>>
>>>>> >>>>
>>>>> >>
>>>>> https://www.evernote.com/shard/s18/sh/4df76607-12ea-42d6-b73a-ff36a9435d7b/510f4a49695062e03aa03fa45d32ad82
>>>>> >>>>>>>
>>>>> >>>>>>> Note the oddly unlabelled cloud icon, which
when clicked,
>>>>> displays:
>>>>> >>>>>>
>>>>> >>>>
>>>>> >>
>>>>> https://www.evernote.com/shard/s18/sh/95042d8c-06fa-4c56-8f89-e9b4bc2b80a4/f1b06913bdb8a327cf1c554f88899380
>>>>> >>>>>>>
>>>>> >>>>>>> I would like to be able to complete this
testing, but would be
>>>>> >> prepared
>>>>> >>>>>> to
>>>>> >>>>>>> vote without it.
>>>>> >>>>>>>
>>>>> >>>>>>> Can someone help me?
>>>>> >>>>>>
>>>>> >>>>>> Browser specific issue?  Is it consistent?
>>>>> >>>>>>
>>>>> >>>>>>>
>>>>> >>>>>>> thanks,
>>>>> >>>>>>>
>>>>> >>>>>>> --
>>>>> >>>>>>> NS
>>>>> >>>>>
>>>>> >>>>>
>>>>> >>>>>
>>>>> >>>>> --
>>>>> >>>>> NS
>>>>> >>>>
>>>>> >>>
>>>>> >>>
>>>>> >>>
>>>>> >>> --
>>>>> >>> NS
>>>>> >>
>>>>> >>
>>>>> >
>>>>> >
>>>>> > --
>>>>> > NS
>>>>>
>>>>>
>>>>
>>>>
>>>> --
>>>> NS
>>>>
>>>
>>>
>>>
>>> --
>>> NS

Mime
View raw message