You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@cloudstack.apache.org by Prasanna Santhanam <ts...@apache.org> on 2013/05/17 18:04:47 UTC

CLOUDSTACK-2554: Yet another blocker?

This one's filed as Major but I tried the XCP 1.6 hypervisor today on
our automated test infrastructure and I'm unable to even bring up a
zone with the XCP hypervisor:

The error when AgentManager attempts to reconnect to an XCP host is as listed
in the bug. Happy to provide any more logs:

https://issues.apache.org/jira/browse/CLOUDSTACK-2554

To me this seems like another blocker.

-- 
Prasanna.,

------------------------
Powered by BigRock.com

Re: CLOUDSTACK-2554: Yet another blocker?

Posted by Prasanna Santhanam <ts...@apache.org>.
On Tue, May 21, 2013 at 12:23:10AM +0000, Chiradeep Vittal wrote:
> Also, why would the 4.2 system vm template block 4.1 ?
It doesn't.  I encountered the issue on master and the bug reporter
encountered it on 4.1. I fixed both. Except master had additional
fixes that are related to functionality targetted for 4.2. So both are
good now.

> >Prasanna,
> >
> >Just to confirm, the reviewboard submission for the issue that was
> >targeting the 4.1 branch was all inclusive of these fixes.  Correct?
Master fixes include fixes to some additional functionality added for
CPU and memory overcommits.

-- 
Prasanna.,

------------------------
Powered by BigRock.com

Re: CLOUDSTACK-2554: Yet another blocker?

Posted by Chiradeep Vittal <Ch...@citrix.com>.
Also, why would the 4.2 system vm template block 4.1 ?

On 5/20/13 12:40 PM, "Chip Childers" <ch...@sungard.com> wrote:

>On Sun, May 19, 2013 at 06:38:01PM +0530, Prasanna Santhanam wrote:
>> On Fri, May 17, 2013 at 06:28:38PM +0000, Koushik Das wrote:
>> > The problem is that there is no default ctor in
>> > XcpServerResource.java. Now the default ctor was changed to a
>> > non-default one by this commit. Adding a default ctor should fix the
>> > issue but need to check if there will be any other side effects due
>> > to the version field introduced.
>> > 
>> > commit 63630a412fddc92fdc68dc27e12d2a68dd09f1dd
>> > Author: Edison Su <su...@gmail.com>
>> > Date:   Wed May 8 11:40:37 2013 -0700
>> > 
>> >     CLOUDSTACK-1907: Debian Squeeze 6.0 (64-bit) is not experimental
>>any more
>> > 
>> > -Koushik
>> > 
>> 
>> Thanks - I understand the intentions of Edison's checkin here. Our
>>systemVMs
>> have upgraded to squeeze which XCP had only experimental support for.
>>So for
>> the older XCPs in CloudStack(s) systemVMs will remain running as
>>before. But
>> for the newer CloudStack(s) the new template will be usable. But agent
>> reconnects fail because of the empty constructor. So thanks for solving
>>half my
>> problem. How would one enforce this in the inheritance though? Any
>>ideas?
>> 
>> Meanwhile, I've gone ahead and made multiple fixes to get XCP to work,
>>each one
>> going deeper down the rabbithole. So it turned out to be more involved
>>than
>> what appeared on surface. My commits are in the branch
>> CLOUDSTACK-2554, rebased and applied on master now. I ran the full setup
>> and ensured XCP systemVMs came up with the latest Debian template. I'm
>> running a BVT across Xen, XCP and KVM to ensure it's not regressed
>> beyond what it was.
>> 
>> Here are the master commits:
>> 76e19d5f5090880d27b6334da6ade032273288f5 CLOUDSTACK-2554: Incorrect
>>compute of minmemory and cpu
>> fa086d0bb1d662c82cded083d3e9937349ee5a80 CLOUDSTACK-2554: Ensuring that
>>we honor hypervisor xen's memory constraints
>> ac48c38122ee057abfa416c0bf85231debc0cff3 Fixing multiple minor
>>annoyances
>> 66303b5ee52e1172e19b608e8c8b75a2b9e883d0 CLOUDSTACK-2554: CloudStack
>>fails to load XCP 1.6 hypervisors
>> 
>> What was particularly annoying was single line messages to large feature
>> commits (23e54bb0) with little to no comments within code but affecting
>>all
>> hypervisors and many files (~33 files)
>> 
>> Please include enough information in the commit message. At the very
>>least I'd
>> expect the commit message to be as long as the number of files you've
>>changed.
>
>Prasanna,
>
>Just to confirm, the reviewboard submission for the issue that was
>targeting the 4.1 branch was all inclusive of these fixes.  Correct?

Re: CLOUDSTACK-2554: Yet another blocker?

Posted by Chip Childers <ch...@sungard.com>.
On Sun, May 19, 2013 at 06:38:01PM +0530, Prasanna Santhanam wrote:
> On Fri, May 17, 2013 at 06:28:38PM +0000, Koushik Das wrote:
> > The problem is that there is no default ctor in
> > XcpServerResource.java. Now the default ctor was changed to a
> > non-default one by this commit. Adding a default ctor should fix the
> > issue but need to check if there will be any other side effects due
> > to the version field introduced.
> > 
> > commit 63630a412fddc92fdc68dc27e12d2a68dd09f1dd
> > Author: Edison Su <su...@gmail.com>
> > Date:   Wed May 8 11:40:37 2013 -0700
> > 
> >     CLOUDSTACK-1907: Debian Squeeze 6.0 (64-bit) is not experimental any more
> > 
> > -Koushik
> > 
> 
> Thanks - I understand the intentions of Edison's checkin here. Our systemVMs
> have upgraded to squeeze which XCP had only experimental support for. So for
> the older XCPs in CloudStack(s) systemVMs will remain running as before. But
> for the newer CloudStack(s) the new template will be usable. But agent
> reconnects fail because of the empty constructor. So thanks for solving half my
> problem. How would one enforce this in the inheritance though? Any ideas?
> 
> Meanwhile, I've gone ahead and made multiple fixes to get XCP to work, each one
> going deeper down the rabbithole. So it turned out to be more involved than
> what appeared on surface. My commits are in the branch
> CLOUDSTACK-2554, rebased and applied on master now. I ran the full setup
> and ensured XCP systemVMs came up with the latest Debian template. I'm
> running a BVT across Xen, XCP and KVM to ensure it's not regressed
> beyond what it was.  
> 
> Here are the master commits:
> 76e19d5f5090880d27b6334da6ade032273288f5 CLOUDSTACK-2554: Incorrect compute of minmemory and cpu
> fa086d0bb1d662c82cded083d3e9937349ee5a80 CLOUDSTACK-2554: Ensuring that we honor hypervisor xen's memory constraints
> ac48c38122ee057abfa416c0bf85231debc0cff3 Fixing multiple minor annoyances
> 66303b5ee52e1172e19b608e8c8b75a2b9e883d0 CLOUDSTACK-2554: CloudStack fails to load XCP 1.6 hypervisors
> 
> What was particularly annoying was single line messages to large feature
> commits (23e54bb0) with little to no comments within code but affecting all
> hypervisors and many files (~33 files) 
> 
> Please include enough information in the commit message. At the very least I'd
> expect the commit message to be as long as the number of files you've changed.

Prasanna,

Just to confirm, the reviewboard submission for the issue that was
targeting the 4.1 branch was all inclusive of these fixes.  Correct?

Re: CLOUDSTACK-2554: Yet another blocker?

Posted by Prasanna Santhanam <ts...@apache.org>.
On Fri, May 17, 2013 at 06:28:38PM +0000, Koushik Das wrote:
> The problem is that there is no default ctor in
> XcpServerResource.java. Now the default ctor was changed to a
> non-default one by this commit. Adding a default ctor should fix the
> issue but need to check if there will be any other side effects due
> to the version field introduced.
> 
> commit 63630a412fddc92fdc68dc27e12d2a68dd09f1dd
> Author: Edison Su <su...@gmail.com>
> Date:   Wed May 8 11:40:37 2013 -0700
> 
>     CLOUDSTACK-1907: Debian Squeeze 6.0 (64-bit) is not experimental any more
> 
> -Koushik
> 

Thanks - I understand the intentions of Edison's checkin here. Our systemVMs
have upgraded to squeeze which XCP had only experimental support for. So for
the older XCPs in CloudStack(s) systemVMs will remain running as before. But
for the newer CloudStack(s) the new template will be usable. But agent
reconnects fail because of the empty constructor. So thanks for solving half my
problem. How would one enforce this in the inheritance though? Any ideas?

Meanwhile, I've gone ahead and made multiple fixes to get XCP to work, each one
going deeper down the rabbithole. So it turned out to be more involved than
what appeared on surface. My commits are in the branch
CLOUDSTACK-2554, rebased and applied on master now. I ran the full setup
and ensured XCP systemVMs came up with the latest Debian template. I'm
running a BVT across Xen, XCP and KVM to ensure it's not regressed
beyond what it was.  

Here are the master commits:
76e19d5f5090880d27b6334da6ade032273288f5 CLOUDSTACK-2554: Incorrect compute of minmemory and cpu
fa086d0bb1d662c82cded083d3e9937349ee5a80 CLOUDSTACK-2554: Ensuring that we honor hypervisor xen's memory constraints
ac48c38122ee057abfa416c0bf85231debc0cff3 Fixing multiple minor annoyances
66303b5ee52e1172e19b608e8c8b75a2b9e883d0 CLOUDSTACK-2554: CloudStack fails to load XCP 1.6 hypervisors

What was particularly annoying was single line messages to large feature
commits (23e54bb0) with little to no comments within code but affecting all
hypervisors and many files (~33 files) 

Please include enough information in the commit message. At the very least I'd
expect the commit message to be as long as the number of files you've changed.

Thanks,

-- 
Prasanna.,

------------------------
Powered by BigRock.com

RE: CLOUDSTACK-2554: Yet another blocker?

Posted by Koushik Das <ko...@citrix.com>.
The problem is that there is no default ctor in XcpServerResource.java. Now the default ctor was changed to a non-default one by this commit. Adding a default ctor should fix the issue but need to check if there will be any other side effects due to the version field introduced.

commit 63630a412fddc92fdc68dc27e12d2a68dd09f1dd
Author: Edison Su <su...@gmail.com>
Date:   Wed May 8 11:40:37 2013 -0700

    CLOUDSTACK-1907: Debian Squeeze 6.0 (64-bit) is not experimental any more

-Koushik

> -----Original Message-----
> From: Prasanna Santhanam [mailto:tsp@apache.org]
> Sent: Friday, May 17, 2013 9:35 PM
> To: CloudStack Dev
> Subject: CLOUDSTACK-2554: Yet another blocker?
> 
> This one's filed as Major but I tried the XCP 1.6 hypervisor today on our
> automated test infrastructure and I'm unable to even bring up a zone with
> the XCP hypervisor:
> 
> The error when AgentManager attempts to reconnect to an XCP host is as
> listed in the bug. Happy to provide any more logs:
> 
> https://issues.apache.org/jira/browse/CLOUDSTACK-2554
> 
> To me this seems like another blocker.
> 
> --
> Prasanna.,
> 
> ------------------------
> Powered by BigRock.com