Bug 764368 (GLUSTER-2636)

Summary: "profile info" lists 'nan' in latency when heavy i/o is running
Product: [Community] GlusterFS Reporter: M S Vishwanath Bhat <vbhat>
Component: glusterdAssignee: Pranith Kumar K <pkarampu>
Severity: medium Docs Contact:
Priority: medium    
Version: mainlineCC: gluster-bugs, mzywusko, pkarampu, rabhat
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---
Description Flags
profile info results taken for every 4000 secs none

Description M S Vishwanath Bhat 2011-03-31 01:47:50 EDT
Created attachment 467
Comment 1 M S Vishwanath Bhat 2011-03-31 04:45:27 EDT
Running profile info lists nan when dbench is running with 150 clients. 
 I was running dbench with 15 clients for 10 hours and was also running fileop with 35 directory depth. I was taking profile info after every 4000 seconds. After first 4000 seconds i started getting nan in profile results. I have attached profile results file for reference. I have also archived the logs.
Comment 2 M S Vishwanath Bhat 2011-04-05 03:01:56 EDT
This bug can be reproduced even without heavy i/o. 
I Created a replicate setup with two bricks. Created 100 files and read from it. Ran profile info and i consistently got nan in latency. I have archived the logs.
Comment 3 Vijay Bellur 2011-04-07 11:40:12 EDT
PATCH: http://patches.gluster.com/patch/6735 in master (mgmt/glusterd: Fix gluster-profile performance issues)
Comment 4 Vijay Bellur 2011-04-07 11:40:16 EDT
PATCH: http://patches.gluster.com/patch/6736 in master (debug/io-stats: Fix races and performance issues in profile)
Comment 5 Raghavendra Bhat 2011-04-13 03:22:52 EDT
Its fixed now. profile info does not say nan now.