Moving to Ceph is very slow when bs=1. Instead, use a larger block size in
combination with the (currently) PVE-specific osize option to specify the
desired output size.
Suggested-by: Dietmar Maurer <diet...@proxmox.com>
Signed-off-by: Fabian Ebner <f.eb...@proxmox.com>
---
Thanks to Dietmar for pointing me in the right direction.
We actually have the exact feature I wished for (see patch #11 in pve-qemu).
I missed it the first time around, because it is not described in the man page,
only the keyword is stated in the short synopsis.
Changes from v1:
* use osize option to control the output file size
PVE/QemuServer.pm | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index f401baf..a498444 100644
--- a/PVE/QemuServer.pm
+++ b/PVE/QemuServer.pm
@@ -6991,7 +6991,11 @@ sub clone_disk {
# that is given by the OVMF_VARS.fd
my $src_path = PVE::Storage::path($storecfg, $drive->{file});
my $dst_path = PVE::Storage::path($storecfg, $newvolid);
- run_command(['qemu-img', 'dd', '-n', '-O', $dst_format, "bs=1",
"count=$size",
+
+ # better for Ceph if block size is not too small, see bug #3324
+ my $bs = 1024*1024;
+
+ run_command(['qemu-img', 'dd', '-n', '-O', $dst_format,
"bs=$bs", "osize=$size",
"if=$src_path", "of=$dst_path"]);
} else {
qemu_img_convert($drive->{file}, $newvolid, $size, $snapname,
$sparseinit);
--
2.20.1
_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
