On 14.07.2012 15:09, Josef Bacik wrote:
Daniel Blueman reported a bug with fio+balance on a ramdisk setup.
Basically what happens is the balance relocates a tree block which will drop
the implicit refs for all of its children and adds a full backref. Once the
block is relocated we have to add
One point regarding the merge: wouldn't it be better to put the seq as a sort
criterion at the end, so the merge can happen in one run through the list
instead of this potentially quadratic time? I've seen some warnings from CPU
stuck 22s which recovered after the test.
On 16.07.2012 09:41, Arne
Daniel Blueman reported a bug with fio+balance on a ramdisk setup.
Basically what happens is the balance relocates a tree block which will drop
the implicit refs for all of its children and adds a full backref. Once the
block is relocated we have to add the implicit refs back, so when we cow the