Bug 1393423

Summary: XFS RAID alignment parameters are not honoured as per recommendation
Product: Red Hat Gluster Storage Reporter: SATHEESARAN <sasundar>
Component: gdeployAssignee: Sachidananda Urs <surs>
Status: CLOSED NOTABUG QA Contact: SATHEESARAN <sasundar>
Severity: high Docs Contact:
Priority: unspecified    
Version: rhgs-3.2CC: rhs-bugs, sasundar, smohan, storage-qa-internal
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2016-11-14 06:26:25 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:

Description SATHEESARAN 2016-11-09 14:24:37 UTC
Description of problem:
-----------------------
XFS RAID alignment recommendation as per RHGS Admin guide states usage of 'su' and 'sw' as per the stripe-size and number of data disks in RAID respectively.

But gdeploy uses the hard-coded values for 'su' ( stripe-unit ) & 'sw' ( stripe-width )

These values should be computed based on the following available in the config file:
[disktype]
raid6
[diskcount]
 4
[stripesize]
 256

Version-Release number of selected component (if applicable):
-------------------------------------------------------------
gdeploy-2.0.1-3.el7rhgs ( this is even applicable for older gdeploy too )

How reproducible:
-----------------
Always

Steps to Reproduce:
-------------------
1. Use gdeploy conf file to create bricks

Actual results:
---------------
gdeploy uses the hard-coded value for su=128k & sw=10

Expected results:
-----------------
for RAID-6 or RAID-10, 'su' - should be the stripe-size & 'sw' should be the number of data disks.

Additional info:
----------------
Snip of the hard-coded values :
<snip>
 gdeployfeatures/lv/lv.py:        section_dict['opts'] = "-f -K -i size=512 -d sw=10,su=128k -n size=8192"
</snip>

Comment 1 Sachidananda Urs 2016-11-09 14:53:51 UTC
(In reply to SATHEESARAN from comment #0)
> Description of problem:
> -----------------------
> XFS RAID alignment recommendation as per RHGS Admin guide states usage of
> 'su' and 'sw' as per the stripe-size and number of data disks in RAID
> respectively.
> 
> But gdeploy uses the hard-coded values for 'su' ( stripe-unit ) & 'sw' (
> stripe-width )
> 
> These values should be computed based on the following available in the
> config file:
> [disktype]
> raid6
> [diskcount]
>  4
> [stripesize]
>  256
> 
> Version-Release number of selected component (if applicable):
> -------------------------------------------------------------
> gdeploy-2.0.1-3.el7rhgs ( this is even applicable for older gdeploy too )
> 
> How reproducible:
> -----------------
> Always
> 
> Steps to Reproduce:
> -------------------
> 1. Use gdeploy conf file to create bricks
> 
> Actual results:
> ---------------
> gdeploy uses the hard-coded value for su=128k & sw=10
> 
> Expected results:
> -----------------
> for RAID-6 or RAID-10, 'su' - should be the stripe-size & 'sw' should be the
> number of data disks.
> 
> Additional info:
> ----------------
> Snip of the hard-coded values :
> <snip>
>  gdeployfeatures/lv/lv.py:        section_dict['opts'] = "-f -K -i size=512
> -d sw=10,su=128k -n size=8192"
> </snip>

sas, I think you are looking at wrong version of gdeploy. This has been fixed
as part of commit: https://github.com/gluster/gdeploy/commit/94eb2f757d50

This issue was reported by Dustin Black in sme-storage, and I have fixed it.

https://github.com/gluster-deploy/gdeploy/blob/2.0.1/gdeployfeatures/lv/lv.py

Can you please recheck the code you are referring?

Comment 2 SATHEESARAN 2016-11-14 06:26:25 UTC
(In reply to Sachidananda Urs from comment #1)
> 
> sas, I think you are looking at wrong version of gdeploy. This has been fixed
> as part of commit: https://github.com/gluster/gdeploy/commit/94eb2f757d50
> 
> This issue was reported by Dustin Black in sme-storage, and I have fixed it.
> 
> https://github.com/gluster-deploy/gdeploy/blob/2.0.1/gdeployfeatures/lv/lv.py
> 
> Can you please recheck the code you are referring?

Hi Sac,

You are correct. I was referring to the older gdeploy code base earlier ( gdeploy-2.0-16 ). But now I checked with the latest gdeploy - gdeploy-2.0.1-3.el7rhgs. It sets the value correctly. Also I misunderstood the values reported by xfs_info,

Thanks for clarifying. I am closing this bug as this is not an issue with latest gdeploy