Yaniv Kaul
2015-Nov-05 11:10 UTC
[Libguestfs] RFE: 'fstrim' minimum block size when sparsifying qcow2 should be 64K
As the minimum block size to discard in qcow2 is 64K, any point in fstrimming smaller contiguous blocks? TIA, Y.
Richard W.M. Jones
2015-Nov-05 13:08 UTC
Re: [Libguestfs] RFE: 'fstrim' minimum block size when sparsifying qcow2 should be 64K
On Thu, Nov 05, 2015 at 01:10:54PM +0200, Yaniv Kaul wrote:> As the minimum block size to discard in qcow2 is 64K, any point in > fstrimming smaller contiguous blocks?For virt-sparsify we just use the default that the util-linux 'fstrim -m' option uses, which (according to the fstrim manual) is 0 so it will try to discard any contiguous extent. Unfortunately it's hard to tell what the best minimum free extent would be. The disk image might not be qcow2 -- if it was raw then any 4K or even 1K block could be worth freeing. Is it a lot slower if we use a smaller minimum free extent? Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones Read my programming and virtualization blog: http://rwmj.wordpress.com virt-df lists disk usage of guests without needing to install any software inside the virtual machine. Supports Linux and Windows. http://people.redhat.com/~rjones/virt-df/
Possibly Parallel Threads
- Any risk in sparsifying a base image (that has a snapshot on top of it)
- Re: Fwd: [Bug 1277705] virt-sparsify --in-place should not sparsify a snapshot
- Re: failure to virt-sysprep (FC27?)
- Re: Fwd: [Bug 1277705] virt-sparsify --in-place should not sparsify a snapshot
- Re: failure to virt-sysprep (FC27?)