[virt-tools-list] [virt-manager PATCH] diskbackend: make sure path existed before checking vol info

[chenhanxiao at cn.fujitsu.com]

> -----Original Message-----
> From: Giuseppe Scrivano [mailto:gscrivan at redhat.com]
> Sent: Friday, June 13, 2014 5:24 PM
> To: Chen, Hanxiao/陈 晗霄
> Cc: virt-tools-list at redhat.com
> Subject: Re: [virt-tools-list] [virt-manager PATCH] diskbackend: make sure path
> existed before checking vol info
> 
> Chen Hanxiao <chenhanxiao at cn.fujitsu.com> writes:
> 
> > https://bugzilla.redhat.com/show_bug.cgi?id=1108922
> >
> > Signed-off-by: Chen Hanxiao <chenhanxiao at cn.fujitsu.com>
> > ---
> >  virtinst/diskbackend.py | 3 ++-
> >  1 file changed, 2 insertions(+), 1 deletion(-)
> >
> > diff --git a/virtinst/diskbackend.py b/virtinst/diskbackend.py
> > index 5f72d00..5725e1f 100644
> > --- a/virtinst/diskbackend.py
> > +++ b/virtinst/diskbackend.py
> > @@ -65,7 +65,8 @@ def check_if_path_managed(conn, path):
> >      def lookup_vol_by_path():
> >          try:
> >              vol = conn.storageVolLookupByPath(path)
> > -            vol.info()
> > +            if os.path.exists(vol.path()):
> > +                vol.info()
> 
> I think that "check_if_path_managed" is used also in the case of a
> remote connection so os.path.exists is not correct as it checks for a
> local file.
> 
> 
> >              return vol, None
> >          except libvirt.libvirtError, e:
> >              if (hasattr(libvirt, "VIR_ERR_NO_STORAGE_VOL")
> 
> I have tried several times by using a similar command line as in the bug
> report but I wasn't able to reproduce the problem here (both
> virt-install-1.0.1-3.fc20.noarch and upstream version), what arguments
> did you feed to virt-install to get this failure?
> 

a) mkdir -p /home/rjones/d/libguestfs/builder/website/

b) ./virt-install  --connect qemu:///system  --virt-type kvm  --name demo  --memory 500   --cdrom /home/ISO/Fedora-20-x86_64-DVD.iso  --os-variant fedora20 --disk /home/rjones/d/libguestfs/builder/website/fedora-20,size=6

c) run and run again.

> Thanks,
> Giuseppe