Re: [PATCH] LVM VG size is not same as the device where PV is. (#480793)

[Hans de Goede <hdegoede redhat com>]

Joel Granados Moreno wrote:
> This takes care of the corner case in which the clamping for the vgsize
> does not give us enough space for the PV metadata.
> Also revert the previous attempt to fix this in autopart.py.  It broke
> everything else and is not a good idea.

Looks good to me.

Regards,

Hans

> ---
>  autopart.py     |    2 +-
>  partRequests.py |    2 ++
>  2 files changed, 3 insertions(+), 1 deletions(-)
>
> diff --git a/autopart.py b/autopart.py
> index 2938d9b..2c6f5e8 100644
> --- a/autopart.py
> +++ b/autopart.py
> @@ -534,7 +534,7 @@ def growLogicalVolumes(diskset, requests):
>  		    log("considering %s, start size = %s",req.logicalVolumeName, req.getStartSize())
>  		    
>  		# get remaining free space
> -		vgfree = lvm.getVGFreeSpace(vgreq, requests, diskset) - (2*vgreq.pesize)
> +		vgfree = lvm.getVGFreeSpace(vgreq, requests, diskset)
>  
>  #		print "vgfree = ", vgfree
>                  if DEBUG_LVM_GROW:
> diff --git a/partRequests.py b/partRequests.py
> index 26b77f0..f907525 100644
> --- a/partRequests.py
> +++ b/partRequests.py
> @@ -762,6 +762,8 @@ class VolumeGroupRequestSpec(RequestSpec):
>                  size = pvreq.getActualSize(partitions, diskset)
>                  #log("size for pv %s is %s" % (pvid, size))
>                  clamped = lvm.clampPVSize(size, self.pesize)
> +                if clamped == long(size):
> +                    clamped = clamped - (2*self.pesize)
>                  log("  got pv.size of %s, clamped to %s" % (size,clamped))
>                  #log("  clamped size is %s" % (size,))
>                  totalspace = totalspace + clamped