Actions
Bug #4821
closedmonitor: actually setting an exclusive in _pick_random_mon would break things
% Done:
0%
Source:
Development
Tags:
Backport:
Regression:
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):
Description
if (other >= 0 && n >= other) n++;
Is that supposed to be an equality instead of a greater-than-or-equals check? I can't make sense of it as it stands.
Updated by Sage Weil almost 11 years ago
no i think it's right. if o is set, max is num_mon-1, and we shift o or greater one to the right to still get a uniform dist. right?
Updated by Greg Farnum almost 11 years ago
max is monmap->size(); there's no modification if other is specified. I guess it would work if we decremented max, and that would preserve a uniform distribution, yes.
This isn't a big hurry though as nobody uses that parameter and I believe it's correct when other is unset.
Updated by Sage Weil almost 11 years ago
- Status changed from New to Resolved
- Assignee changed from Joao Eduardo Luis to Sage Weil
oh i see, i was looking at the MonClient version (that this was probably modeled after).
Actions