From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <4BC5D83D.7040600@redhat.com> Date: Wed, 14 Apr 2010 17:59:09 +0300 From: Avi Kivity MIME-Version: 1.0 References: <4BC58DF3.608@redhat.com> <4BC5D58F.2030809@cfl.rr.com> In-Reply-To: <4BC5D58F.2030809@cfl.rr.com> Content-Transfer-Encoding: 7bit Subject: Re: [linux-lvm] Problems resizing physical volume Reply-To: LVM general discussion and development List-Id: LVM general discussion and development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , List-Id: Content-Type: text/plain; charset="us-ascii"; format="flowed" To: Phillip Susi Cc: LVM general discussion and development On 04/14/2010 05:47 PM, Phillip Susi wrote: > On 4/14/2010 5:42 AM, Avi Kivity wrote: > >> (resending, first try didn't seem to make it to the archives) >> >> Due to dracut creating huge initrds, I now need a larger /boot. So I'd >> like to reduce my lvm physical volume on the same disk in order to make >> room. >> > This doesn't directly address your issue, but I'd point out that if > /boot were itself an lv instead of a normal partition, resizing it would > be easy ;) > I don't think my boot loader supports LVs? > Personally I just leave /boot on the root lv. > > I'm triple booting, so I have multiple root volumes. >> Looks like all extents are already allocated (that's the 'a', right?). >> But it claims that 10G are free! And sure, if I create a new volume, >> space is found at extent 4384: >> > The output is kind of silly in that it shows the free extents too > without any indication that they are free. > Yes. >>> [root@violet-regb ~]# pvmove /dev/sda2:4704-4765 /dev/sda2:4384-4445 >>> No extents available for allocation >>> >> At this point I gave up. Is there any way to accomplish this? >> > pvmove normally is used to move segments off the pv to other pvs in > order to free up that pv for removal. So it is failing since it can not > find space on other disks, since you limited its search to specific > extents on the same disk. You need to add --alloc anywhere to stop it > from excluding the current disk from its search. The need for this is > kind of silly given that you explicitly tell it where to move to, it > should probably be assumed, but isn't. > I'll try it out. Thanks. -- error compiling committee.c: too many arguments to function