Bug 1331756

Summary: OSD's which are in "Down" State are also getting considered while reweight-by-utilization
Product: Red Hat Ceph Storage Reporter: Tanay Ganguly <tganguly>
Component: RADOSAssignee: Sage Weil <sweil>
Status: CLOSED NOTABUG QA Contact: ceph-qe-bugs <ceph-qe-bugs>
Severity: urgent Docs Contact:
Priority: unspecified    
Version: 1.3.2CC: ceph-eng-bugs, dzafman, hnallurv, kchai, kurs, nlevine
Target Milestone: rc   
Target Release: 2.0   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2016-05-02 19:24:24 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Attachments:
Description Flags
Steps none

Description Tanay Ganguly 2016-04-29 12:48:50 UTC 
Created attachment 1152279 [details]
Steps

Description of problem:
OSD's which are in "Down" State are also getting considered while reweight-by-utilization.

Version-Release number of selected component (if applicable):
ceph version 0.94.5-6redhat1trusty

How reproducible:
Always

Steps to Reproduce:
1. Mark 1 OSD down.
2. Immediately start the reweight-by-utilization, with the value which satisfies the "Down" OSD condition.

e.g. Down OSD 
 0 0.89999  0.80005   926G  782G  143G 84.50 1.10

 sudo ceph osd test-reweight-by-utilization 108 .05 4

Actual results:
"Down" OSD is also considered while reweighting

Expected results:
"Down" OSD should not be considered for reweight.

Additional info:
PFA, Detailed Steps
Comment 2 Samuel Just 2016-04-29 14:49:00 UTC 
So it should factor them into the average, but not reweight them?
Comment 3 Sage Weil 2016-04-29 20:29:51 UTC 
It feels like we should consider all non-out OSDs, regardless of whether they are up or down, since up/down is normally transient.
Comment 4 Neil Levine 2016-05-02 19:24:24 UTC 
Agreed.