Bug 1164338

Summary:	enable_vol.sh doesn't ignore comments in /etc/fstab during check
Product:	[Red Hat Storage] Red Hat Gluster Storage	Reporter:	Martin Bukatovic <mbukatov>
Component:	rhs-hadoop-install	Assignee:	Jeff Vance <jvance>
Status:	CLOSED ERRATA	QA Contact:	Martin Kudlej <mkudlej>
Severity:	low	Docs Contact:
Priority:	unspecified
Version:	rhgs-3.0	CC:	bchilds, eboyd, matt, mkudlej, nlevinki, rcyriac
Target Milestone:	---	Keywords:	ZStream
Target Release:	RHGS 3.0.4
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:	2.39-1	Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2015-03-31 10:18:34 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:

Description Martin Bukatovic 2014-11-14 17:21:35 UTC

Description of problem
======================

Script enable_vol.sh doesn't ignore commented lines in /etc/fstab when it tries
to verify it. This may lead to unnecessary failure during brownfield usecase.

Version-Release number of selected component (if applicable)
============================================================

rhs-hadoop-install-2_29-1.el6rhs.noarch

Steps to Reproduce
==================

Follow RHS Installation Guide with brownfield cluster:

1. When checking gluster volume mount options, comment out the original line,
   while adding new with required mount options, eg:

~~~
#localhost:/BrownFieldVolume /mnt/brownfield glusterfs defaults,_netdev 0 0
localhost:/BrownFieldVolume /mnt/brownfield glusterfs entry-timeout=0,attribute-timeout=0,use-readdirp=no,_netdev 0 0
~~~

2. Run enable_vol.sh

Actual results
==============

The script fails because it doesn't ignore comments in /etc/fstab file:

~~~
ERROR: issues with 1 or more nodes spanned by BrownFieldVolume
A suggestion is to re-run the setup_cluster.sh script to ensure that
all nodes in the cluster are set up correctly for Hadoop workloads.
See the /var/log/rhs-hadoop-install.log log file for additional info.
~~~

The error in the log is:

~~~
--- 10.16.66.77: live BrownFieldVolume mount options check...
--- 10.16.66.77: /etc/fstab BrownFieldVolume mount options check...
ERROR on 10.16.66.77: BrownFieldVolume appears more than once in /etc/fstab
--- 10.16.64.42: live BrownFieldVolume mount options check...
--- 10.16.64.42: /etc/fstab BrownFieldVolume mount options check...
ERROR on 10.16.64.42: BrownFieldVolume appears more than once in /etc/fstab
--- 10.16.67.178: live BrownFieldVolume mount options check...
--- 10.16.67.178: /etc/fstab BrownFieldVolume mount options check...
ERROR on 10.16.67.178: BrownFieldVolume appears more than once in /etc/fstab
--- 10.16.67.194: live BrownFieldVolume mount options check...
--- 10.16.67.194: /etc/fstab BrownFieldVolume mount options check...
ERROR on 10.16.67.194: BrownFieldVolume appears more than once in /etc/fstab
~~~

Expected results
================

The script ignores comments in /etc/fstab and so it doesn't fail.

Additional info
===============

This is low impact bug, because admin will easily figure how to fix this issue.

On the other hand, this may happen to anyone enabling hadoop on existing
gluster volume (brownfield use case). Since admin needs to recheck/change mount
options on RHS Server nodes, he may like to save the previous configuration
in the comment.

Additional question: does it mean that having multiple mount points on the RHS
Server nodes would break something?

Comment 2 Jeff Vance 2014-11-14 20:47:34 UTC

"does it mean that having multiple mount points on the RHS Server nodes would break something?"
The glusterfs-fuse mount fstab entry should still only appear once, even though there are multiple brick mounts for a volume, right?

Comment 3 Martin Bukatovic 2014-11-18 18:37:59 UTC

(In reply to Jeff Vance from comment #2)
> The glusterfs-fuse mount fstab entry should still only appear once, even
> though there are multiple brick mounts for a volume, right?

Yep. But my question was a bit different: is there any reason why mounting the
hadoop enabled gluster volume more than once is a bad idea? Note that this
question is not directly related to the issue being reported/fixed here.

Comment 4 Jeff Vance 2014-11-20 00:54:00 UTC

It's been mentioned by our team that the user may want different vol mounts to support different mount options. So, we could see:
   /mnt/glusterfs1/HadoolVol and /mnt/glusterfs2/HadoopVol

bin.check_volmnt.sh will be changed to search for the volume mount rather than just the volume name. It is still expected that there will be only 1 mount per active volume, even if the volname appears in /etc/fstab more than once.

Comment 5 Jeff Vance 2014-11-20 01:00:32 UTC

Fixed in 2.39

Comment 8 Martin Kudlej 2015-03-12 10:11:06 UTC

Tested with rhs-hadoop-install-2_44-1.el6rhs.noarch and it works. -->VERIFIED

Comment 10 errata-xmlrpc 2015-03-31 10:18:34 UTC

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://rhn.redhat.com/errata/RHEA-2015-0761.html