Guys, because of this issue we couldn't certify MOS 6.0 with Cinder/LVM at least on Ubuntu because of this issue. Could we increase severity of it to HIGH/CRITICAL ? This problem occurs on scale because cinder dd'ing logical volumes by zeroes and on high load it basically could do that in reasonable time and and delete volume. So this bug wouldn't appear on 1 by 1 volumes creation/attach/delete and so on but in production use cases it will lead to the described issue and customer will get a lot of hanged tasks and a lot of trash in volumes db, wasting space and so on. We have to merge that ASAP.
Guys, because of this issue we couldn't certify MOS 6.0 with Cinder/LVM at least on Ubuntu because of this issue. Could we increase severity of it to HIGH/CRITICAL ? This problem occurs on scale because cinder dd'ing logical volumes by zeroes and on high load it basically could do that in reasonable time and and delete volume. So this bug wouldn't appear on 1 by 1 volumes creation/ attach/ delete and so on but in production use cases it will lead to the described issue and customer will get a lot of hanged tasks and a lot of trash in volumes db, wasting space and so on. We have to merge that ASAP.