Bug 762777 (GLUSTER-1045) - [3.0.5rc9]:locktests fails while I/O going on
Summary: [3.0.5rc9]:locktests fails while I/O going on
Keywords:
Status: CLOSED NOTABUG
Alias: GLUSTER-1045
Product: GlusterFS
Classification: Community
Component: write-behind
Version: 3.0.4
Hardware: All
OS: Linux
urgent
medium
Target Milestone: ---
Assignee: Raghavendra G
QA Contact:
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2010-07-03 09:36 UTC by Raghavendra Bhat
Modified: 2015-12-01 16:45 UTC (History)
3 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed:
Regression: ---
Mount Type: ---
Documentation: ---
CRM:
Verified Versions:


Attachments (Terms of Use)

Description Raghavendra Bhat 2010-07-03 09:36:11 UTC
locktests fails on a file if I/O is going on that file, when flush-behind option is on(which id the default). Turning flush-behind off works fine and locktests work fine.

Steps to reproduce:

Keep writing to a file in a loop (dd in this case), and parellely run locktests on that file ( 2 clients were used in this case). locktests fails saying some of the opearations are failed:



process number running test successfully :
50 process of 50 successfully ran test : WRITE ON A READ  LOCK
50 process of 50 successfully ran test : WRITE ON A WRITE LOCK
50 process of 50 successfully ran test : READ  ON A READ  LOCK
50 process of 50 successfully ran test : READ  ON A WRITE LOCK
50 process of 50 successfully ran test : SET A READ  LOCK ON A READ  LOCK
49 process of 50 successfully ran test : SET A WRITE LOCK ON A WRITE LOCK
50 process of 50 successfully ran test : SET A WRITE LOCK ON A READ  LOCK
0 process of 50 successfully ran test : SET A READ  LOCK ON A WRITE LOCK
50 process of 50 successfully ran test : READ LOCK THE WHOLE FILE BYTE BY BYTE
50 process of 50 successfully ran test : WRITE LOCK THE WHOLE FILE BYTE BY BYTE
Init


Here out of 50 processes 49 processes are successful at one point and none of the processes are successful at the other.


Note You need to log in before you can comment on or make changes to this bug.