Bug 1362198

+++ This bug was initially created as a clone of Bug #1349284 +++

+++ This bug was initially created as a clone of Bug #1299841 +++

Description of problem:
Today, When a file is heated which is of size greater than the size of hot tier's high water mark, the file is promoted. Gluster should intelligently determine that the size of the file is more than high watermark and should not get promoted. 

with the following scenario,

hot tier capacity: 100Gb
high water mark - 70% i.e., 70Gb
File is created of size 80Gb

Having such file in hot tier has multiple consequences.

1) When this file is in hot tier, we always exceed high watermark and get into complete demotion mode. As a result, we make all other on hot tier a possible candidate for demotion.
2) No other files will be promoted until this file moves to cold tier
3) Possibility of moving such a huge file back and forth from cold to hot tier which might affect performance

Version-Release number of selected component (if applicable):
glusterfs-3.7.5-15.el7rhgs.x86_64

How reproducible:
Always

Steps to Reproduce:
1. create a gluster volume
2. create a 80Gb file, say file1
3. Attach hot tier of size 100Gb and set high watermark at 70%
4. heat the file - file1 created in step 2

Actual results:
File gets promoted

Expected results:
File should not get promoted

Additional info:

--- Additional comment from Red Hat Bugzilla Rules Engine on 2016-01-19 06:38:14 EST ---

This bug is automatically being proposed for the current z-stream release of Red Hat Gluster Storage 3 by setting the release flag 'rhgs‑3.1.z' to '?'. 

If this bug should be proposed for a different release, please manually change the proposed release flag.

--- Additional comment from Vijay Bellur on 2016-06-23 03:18:47 EDT ---

REVIEW: http://review.gluster.org/14780 (cluster/tier: dont promote estimated block consumption > hi watermark) posted (#1) for review on master by Milind Changire (mchangir)

--- Additional comment from Vijay Bellur on 2016-06-23 03:59:25 EDT ---

REVIEW: http://review.gluster.org/14780 (cluster/tier: dont promote estimated block consumption > hi watermark) posted (#2) for review on master by Milind Changire (mchangir)

--- Additional comment from Vijay Bellur on 2016-06-29 06:31:50 EDT ---

REVIEW: http://review.gluster.org/14780 (cluster/tier: dont promote if estimated block consumption > hi watermark) posted (#3) for review on master by Milind Changire (mchangir)

--- Additional comment from Vijay Bellur on 2016-07-05 06:27:23 EDT ---

REVIEW: http://review.gluster.org/14780 (cluster/tier: dont promote if estimated block consumption > hi watermark) posted (#4) for review on master by Milind Changire (mchangir)

--- Additional comment from Vijay Bellur on 2016-07-05 12:40:26 EDT ---

REVIEW: http://review.gluster.org/14780 (cluster/tier: dont promote if estimated block consumption > hi watermark) posted (#5) for review on master by Milind Changire (mchangir)

--- Additional comment from Vijay Bellur on 2016-07-12 02:44:12 EDT ---

REVIEW: http://review.gluster.org/14780 (cluster/tier: dont promote if estimated block consumption > hi watermark) posted (#6) for review on master by Milind Changire (mchangir)

--- Additional comment from Vijay Bellur on 2016-07-12 09:06:15 EDT ---

REVIEW: http://review.gluster.org/14780 (cluster/tier: dont promote if estimated block consumption > hi watermark) posted (#7) for review on master by Milind Changire (mchangir)

--- Additional comment from Vijay Bellur on 2016-07-15 04:57:49 EDT ---

REVIEW: http://review.gluster.org/14780 (cluster/tier: dont promote if estimated block consumption > hi watermark) posted (#8) for review on master by Milind Changire (mchangir)

--- Additional comment from Mohammed Rafi KC on 2016-07-15 06:33:30 EDT ---

upstream master patch : http://review.gluster.org/#/c/14780/

--- Additional comment from Vijay Bellur on 2016-07-19 06:54:31 EDT ---

REVIEW: http://review.gluster.org/14780 (cluster/tier: dont promote if estimated block consumption > hi watermark) posted (#9) for review on master by Milind Changire (mchangir)

--- Additional comment from Vijay Bellur on 2016-07-20 02:52:21 EDT ---

REVIEW: http://review.gluster.org/14780 (cluster/tier: dont promote if estimated block consumption > hi watermark) posted (#10) for review on master by Milind Changire (mchangir)

--- Additional comment from Vijay Bellur on 2016-07-20 09:04:04 EDT ---

REVIEW: http://review.gluster.org/14780 (cluster/tier: dont promote if estimated block consumption > hi watermark) posted (#11) for review on master by Milind Changire (mchangir)

--- Additional comment from Vijay Bellur on 2016-07-21 03:30:03 EDT ---

REVIEW: http://review.gluster.org/14780 (cluster/tier: dont promote if estimated block consumption > hi watermark) posted (#12) for review on master by Milind Changire (mchangir)

--- Additional comment from Vijay Bellur on 2016-07-27 09:54:35 EDT ---

REVIEW: http://review.gluster.org/14780 (cluster/tier: dont promote if estimated block consumption > hi watermark) posted (#13) for review on master by Milind Changire (mchangir)

--- Additional comment from Vijay Bellur on 2016-07-29 10:14:40 EDT ---

COMMIT: http://review.gluster.org/14780 committed in master by Dan Lambright (dlambrig) 
------
commit 1f4e41e8c2f5f4af4564caba0a08996853f089f4
Author: Milind Changire <mchangir>
Date:   Thu Jun 23 12:37:51 2016 +0530

    cluster/tier: dont promote if estimated block consumption > hi watermark
    
    Add test to fail promotion if estimated block consumption grows
    beyond hi watermark.
    
    Skip file migrations until next cycle if tier_get_fs_stat() fails
    in tier_migrate_using_query_file()
    
    Change-Id: Ice04572fa739c09109c4433e65965197482a7beb
    BUG: 1349284
    Signed-off-by: Milind Changire <mchangir>
    Reviewed-on: http://review.gluster.org/14780
    Smoke: Gluster Build System <jenkins.org>
    CentOS-regression: Gluster Build System <jenkins.org>
    NetBSD-regression: NetBSD Build System <jenkins.org>
    Reviewed-by: mohammed rafi  kc <rkavunga>
    Reviewed-by: Dan Lambright <dlambrig>