[libvirt] [PATCH 3/3]: Implement logical stopPool and deletePool more fully

[Daniel Veillard]

On Tue, Sep 23, 2008 at 09:59:35AM +0100, Daniel P. Berrange wrote:
> On Tue, Sep 23, 2008 at 09:30:46AM +0200, Chris Lalancette wrote:
> > Daniel P. Berrange wrote:
> > >> 3)  In deletePool, remove the -f from the vgremove command.  Besides the fact
> > >> that we probably don't usually want to force things, the -f option doesn't exist
> > >> prior to F-9, so this would fail on F-8, RHEL-5, etc.
> > > 
> > > Hmm, are you sure it won't prompt for a y/n confirmation ? IIRC, that was 
> > > the reason I put in the -f. Perhaps simply ensuring stdin is /dev/null
> > > takes care of that problem.
> > 
> > This was already committed, but just to follow-up:
> > 
> > First, I was wrong about RHEL-5.  Apparently we've rebased the LVM tools in
> > RHEL-5, so it does have the -f flag for vgremove.  That being said, it's not
> > actually necessary, based on my testing.  Here's what I did:
> > 
> > 1)  pvcreate /dev/sdb
> > 2)  vgcreate MyVG2 /dev/sdb
> > 3)  lvcreate -L1G 0n guest1 /dev/MyVG2
> > 
> > At this point, I have a valid volume at /dev/MyVG2/guest1
> > 
> > 4)  lvremove -f /dev/MyVG2/guest1 (this corresponds to the code in
> > virStorageBackendLogicalDeleteVol)
> > 5)  vgremove /dev/MyVG2 (this corresponds to the new code in
> > virStorageBackendLogicalDeletePool)
> > 6)  pvremove /dev/sdb
> > 
> > All commands completed without prompts, and remove the logical volumes.  In step
> > 4), if I do a vgremove, *then* I get a prompt saying:
> > 
> > "Do you really want to remove volume group "MyVG2" containing 1 logical volumes?
> > [y/n]: n"
> > 
> > But that's only because you are trying to de-activate a volume group that has
> > active logical volumes.
> 
> Ah well that's the edge case then. There is no requirement in the libvirt
> virStoragePoolDelete method that says all volumes must be removed before
> calling it. So its certainly possible we'll ht this scenario. We either 
> have to explicitly remove all volumes ourselves in the Delete method, or
> we can trap & raise an error indicating that the app must remove them all.
> I tend towards the latter.

  Okay, so at this point i will rather revert that part of the patch
adding back the -f flag, it feels safer for the upcoming release.

Daniel

-- 
Daniel Veillard      | libxml Gnome XML XSLT toolkit  http://xmlsoft.org/
daniel at veillard.com  | Rpmfind RPM search engine http://rpmfind.net/
http://veillard.com/ | virtualization library  http://libvirt.org/