I’ve been migrating some very large VMs, using export/import, and have noticed this behavior, I now has been reported and maybe answered, but would like to know the logic/reason behind this.
The command to import (convert) from a qcow2 to a raw (LVM) device on FC. /usr/bin/nice -n 19 /usr/bin/ionice -c 3 /usr/bin/qemu-img convert -t none -T none -f qcow2 /rhev/data-center/333f0407-ace1-4cb3-8147-39931c4942c5/92f8390f-812e-4872-b2dc-44433f53146a/images/6c5986df-65d2-41cb-9901-1c9ade164b55/5e24f52c-c8eb-4f9d-8eeb-5e3f0982add1 -O raw /rhev/data-center/mnt/blockSD/1c4b3226-b530-4554-81e9-ead20c1512c5/images/6c5986df-65d2-41cb-9901-1c9ade164b55/5e24f52c-c8eb-4f9d-8eeb-5e3f0982add1 The best BW achieved with this on a 1Gbps link was from 15 to 30MB/s. I tried modifying ionice, priority, and nice, and got almost the same values, after spending must of the night up. I stop the process and tried to to this without the import. Just activated the LVM using the ID of the VM Disk image, and executing without the -t and -T option, that I understand it’s related to the cache of the source and destination. But could use almost entire Gb link 120MB/s. Is there any reason, for doing this this way ? Greetings, -- Jorge Luis Andrade Escobar [email protected] <mailto:[email protected]> [email protected] <mailto:[email protected]> Red Hat Certified Architect (RHCA) Level IV Red Hat Certified Instructor/Examiner (RHCX) # 100-000-967 gpg: 685C 68E8 9F55 771B ED7C 3CC7 5E65 8385 88D3 09A5 Phone: +502 2305-0800 ext. 4005 http://www.i-t-m.com <http://www.i-t-m.com/>
signature.asc
Description: Message signed with OpenPGP using GPGMail
_______________________________________________ Devel mailing list [email protected] http://lists.ovirt.org/mailman/listinfo/devel
