osdir.com

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Problem with CLOUDSTACK-10240 (Cannot migrate local volume to shared storage)


Correct. I do think the problem here is only in the release notes.

Just to confirm, you found the problem while testing 4.12 (from master),
right?

On Tue, Jul 17, 2018 at 4:22 PM, Tutkowski, Mike <Mike.Tutkowski@xxxxxxxxxx>
wrote:

> Cool, if it’s just in master, then that makes it easier.
>
> Also, it means we did not have a process issue by introducing enhancement
> code in between release candidates.
>
> It would mean, however, that our documentation is a bit incorrect if, in
> fact, it states that that feature exists in 4.11.1.
>
> > On Jul 17, 2018, at 1:20 PM, Rafael Weingärtner <
> rafaelweingartner@xxxxxxxxx> wrote:
> >
> > Ok, thanks. I had the impression that we said it was backported to 4.11.
> >
> > I will get master and work on it then.
> >
> > On Tue, Jul 17, 2018 at 4:12 PM, Tutkowski, Mike <
> Mike.Tutkowski@xxxxxxxxxx>
> > wrote:
> >
> >> I only noticed it in master. The example code I was comparing it against
> >> was from 4.11.0. I never checked against 4.11.1.
> >>
> >>> On Jul 17, 2018, at 1:02 PM, Rafael Weingärtner <
> >> rafaelweingartner@xxxxxxxxx> wrote:
> >>>
> >>> Hey Mike, I got the branch 4.11 to start fixing the problem we
> discussed,
> >>> but I do not think my commit was backported to 4.11. I mean, I am at
> >>> "VirtualMachineManagerImpl" and the code is not here. I also checked
> the
> >>> commit (
> >>> https://github.com/apache/cloudstack/commit/
> >> f2efbcececb3cfb06a51e5d3a2e77417c19c667f)
> >>> that introduced those changes to master, and according to Github, it is
> >>> only in the master branch, and not in 4.11.
> >>>
> >>> I checked the "VirtualMachineManagerImpl" class at the Apache
> CloudStack
> >>> remote repository in the 4.11 branch, and as you can see, the code
> there
> >> is
> >>> the “old”   one.
> >>> https://github.com/apache/cloudstack/blob/4.11/engine/
> >> orchestration/src/com/cloud/vm/VirtualMachineManagerImpl.java
> >>>
> >>> I got a little confused now. Did you detect the problem in 4.11 or in
> >>> master?
> >>>
> >>>
> >>> On Tue, Jul 17, 2018 at 12:27 AM, Tutkowski, Mike <
> >> Mike.Tutkowski@xxxxxxxxxx
> >>>> wrote:
> >>>
> >>>> Another comment here: The part that is broken is if you try to let
> >>>> CloudStack pick the primary storage on the destination side. That code
> >> no
> >>>> longer exists in 4.11.1.
> >>>>
> >>>> On 7/16/18, 9:24 PM, "Tutkowski, Mike" <Mike.Tutkowski@xxxxxxxxxx>
> >> wrote:
> >>>>
> >>>>   To follow up on this a bit: Yes, you should be able to migrate a VM
> >>>> and its storage from one cluster to another today using non-managed
> >>>> (traditional) primary storage with XenServer (both the source and
> >>>> destination primary storages would be cluster scoped). However, that
> is
> >> one
> >>>> of the features that was broken in 4.11.1 that we are discussing in
> this
> >>>> thread.
> >>>>
> >>>>   On 7/16/18, 9:20 PM, "Tutkowski, Mike" <Mike.Tutkowski@xxxxxxxxxx>
> >>>> wrote:
> >>>>
> >>>>       For a bit of info on what managed storage is, please take a look
> >>>> at this document:
> >>>>
> >>>>       https://www.dropbox.com/s/wwz2bjpra9ykk5w/SolidFire%
> >>>> 20in%20CloudStack.docx?dl=0
> >>>>
> >>>>       The short answer is that you can have zone-wide managed storage
> >>>> (for XenServer, VMware, and KVM). However, there is no current
> zone-wide
> >>>> non-managed storage for XenServer.
> >>>>
> >>>>       On 7/16/18, 6:20 PM, "Yiping Zhang" <yzhang@xxxxxxxxxxx> wrote:
> >>>>
> >>>>           I assume by "managed storage", you guys mean primary
> >> storages,
> >>>> either zone -wide or cluster-wide.
> >>>>
> >>>>           For Xen hypervisor, ACS does not support "zone-wide" primary
> >>>> storage yet. Still, I can live migrate a VM with data disks between
> >>>> clusters with storage migration from web GUI, today.  So, your
> statement
> >>>> below does not reflect current behavior of the code.
> >>>>
> >>>>
> >>>>                      - If I want to migrate a VM across clusters, but
> >> if
> >>>> at least one of its
> >>>>                      volumes is placed in a cluster-wide managed
> >>>> storage, the migration is not
> >>>>                      allowed. Is that it?
> >>>>
> >>>>               [Mike] Correct
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>
> >>>
> >>> --
> >>> Rafael Weingärtner
> >>
> >
> >
> >
> > --
> > Rafael Weingärtner
>



-- 
Rafael Weingärtner