Bug 1208367

Summary:	Data Tiering:Data moving to the tier(inertia) where data already exists instead of moving to hot tier first by default
Product:	[Community] GlusterFS	Reporter:	Nag Pavan Chilakam <nchilaka>
Component:	tiering	Assignee:	Joseph Elwin Fernandes <josferna>
Status:	CLOSED CURRENTRELEASE	QA Contact:	bugs <bugs>
Severity:	urgent	Docs Contact:
Priority:	urgent
Version:	mainline	CC:	bugs, sankarshan
Target Milestone:	---
Target Release:	---
Hardware:	Unspecified
OS:	Linux
Whiteboard:
Fixed In Version:	glusterfs-3.7dev-0.869.gitf5e4c94.el6	Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:
Clones:	1229244 (view as bug list)		Environment:
Last Closed:	2016-02-14 04:10:53 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:
Embargoed:
Bug Depends On:
Bug Blocks:	1186580, 1229244, 1260923

Description Nag Pavan Chilakam 2015-04-02 06:02:39 UTC

Description of problem:
=======================
adding new data to a tiered volume doesnt always get added to hot tier first by default.
Suppose if it is an existing distribute volume with already data available on it and after we convert the volume to a tiered volume by attaching tier, and then add any data, the data still gets written to cold instead of hot tier, hence making the hot tier useless.
In short, new data moves to the hot tier if the volume is a complete new tier volume, else it moves to cold tier if it was already an existing regular volume with data in it and got converted to tier vol.
Note: I haven't set any ctr options
Also, there is no volume quota policies set.
I had tested with bare minimum zero sized files,hence storage space should not be an issue

Version-Release number of selected component (if applicable):
============================================================
Upstream nightly build
glusterfs 3.7dev built on Mar 31 2015 01:05:54
Repository revision: git://git.gluster.com/glusterfs.git
Copyright (c) 2006-2011 Gluster Inc. <http://www.gluster.com>
GlusterFS comes with ABSOLUTELY NO WARRANTY.
You may redistribute copies of GlusterFS under the terms of the GNU General Public License.



How reproducible:
================
easily

Steps to Reproduce:
==================
1.have a regular distribute volume and add data to it
2.Now attach a tier to convert this volume to tier volume
3. Now again add data to the tiered volume.
It can be observed that the new data is not getting written into new hot tier but still getting added to the cold tier.


Expected results:
================
Any new data should get added to the hot tier

Comment 1 Joseph Elwin Fernandes 2015-04-14 11:48:41 UTC

I couldn't reproduce this bug with the latest upstream.

Distribute:
~~~~~~~~~~~~~~~
Step 1: Created a distribute volume
step 2: mounted using fuse and created 300 files.
step 3: all the files are in the volume bricks
step 4: attached a distribute tier
step 5: created new 300 files
step 6: all new files are on the hot-tier bricks


Distribute-Replica:
~~~~~~~~~~~~~~~~~~
Step 1: Created a distribute-replica 2 volume
step 2: mounted using fuse and created 300 files.
step 3: all the files are in the volume bricks
step 4: attached a distribute-replica 2 tier
step 5: created new 300 files
step 6: all new files are on the hot-tier bricks

Comment 2 Niels de Vos 2015-05-15 13:07:41 UTC

This change should not be in "ON_QA", the patch posted for this bug is only available in the master branch and not in a release yet. Moving back to MODIFIED until there is an beta release for the next GlusterFS version.