On 31.05.2014 20:45, Benoît Canet wrote:
The Friday 30 May 2014 à 19:53:12 (+0200), Max Reitz wrote :
On 30.05.2014 13:18, Benoît Canet wrote:
These are the last bits required to make quorum usable in production.
v5: rebase on latest Stefan's block branch [Kevin]
v4:
update patchset to stefan's block branch
drop Max reviewed by because the series changes
Benoît Canet (3):
quorum: Add the rewrite-corrupted parameter to quorum.
block: Add drive-mirror-replace command
qemu-iotests: Add 096 new test for drive-mirror-replace.
Independently of this series, while looking at patch 1 again
(although I had reviewed it before already), I noticed that
quorum_get_winner() does not select the definite winner, but only a
winner. So if you have a quorum instance with four children and a
threshold of two, it may happen that there are basically two winners
which both fulfill the threshold condition. In this case,
quorum_get_winner() will just return the first winner. However, I
think it should then return that there is no winner (i.e., NULL).
On the other hand, the user should be aware that it may be a bad
idea to choose a threshold which does not exceed half of the
children count; thus, I'm just asking what you think about this. :-)
I think Quorum n/(2 *n) is a bad idea. (n + 1) / (2 *n) is a least required.
I wonder where we could put some documentation about this for the user.
Well, quorum_valid_threshold() exists and could check this… Maybe it
could at least emit a warning?
Max
Best regards
Benoît
Max
