Public bug reported: In this thread:
http://lists.openstack.org/pipermail/openstack/2015-August/013614.html [qote] Running cmd (subprocess): sudo nova-rootwrap /etc/nova/rootwrap.conf qemu-img convert -O raw /var/lib/nova/instances/_base/999f7fff2521e4a7243c9e1d21599fd64a19b42e /dev/openstack-controller01-ky01-vg/5f831046-435c-4636-8b71-a662327b608c_disk Well, I understand what this step is doing - it's copying raw image to lvm. How can we speed it up? I don't wanna have instance from 100GB image booted for 4 hours [/quote] Regardless of how much data is in the source file, qemu-img has to write out the full virtual size on the target block dev to ensure any pre- existing data is overwritten. THis may involve writing 100's of GB of zeros for large images. qemu-img does has an optimization to let it use 'discard' as a fast alternative to writing zeros, but this requires that qemu-img has disk caching disabled , eg -t none. We should investigate to see if this improves performance for converting large images to block devices, and if so, enable this flag. ** Affects: nova Importance: Undecided Status: New ** Tags: libvirt ** Tags added: libvirt -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1484992 Title: Converting image from qcow2 to raw is slow for images with large virtual size Status in OpenStack Compute (nova): New Bug description: In this thread: http://lists.openstack.org/pipermail/openstack/2015-August/013614.html [qote] Running cmd (subprocess): sudo nova-rootwrap /etc/nova/rootwrap.conf qemu-img convert -O raw /var/lib/nova/instances/_base/999f7fff2521e4a7243c9e1d21599fd64a19b42e /dev/openstack-controller01-ky01-vg/5f831046-435c-4636-8b71-a662327b608c_disk Well, I understand what this step is doing - it's copying raw image to lvm. How can we speed it up? I don't wanna have instance from 100GB image booted for 4 hours [/quote] Regardless of how much data is in the source file, qemu-img has to write out the full virtual size on the target block dev to ensure any pre-existing data is overwritten. THis may involve writing 100's of GB of zeros for large images. qemu-img does has an optimization to let it use 'discard' as a fast alternative to writing zeros, but this requires that qemu-img has disk caching disabled , eg -t none. We should investigate to see if this improves performance for converting large images to block devices, and if so, enable this flag. To manage notifications about this bug go to: https://bugs.launchpad.net/nova/+bug/1484992/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : [email protected] Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
