Bug 1591624

Summary:	[RFE]: Provide an option to create block hosting volume as arbiter volumes
Product:	[Red Hat Storage] Red Hat Gluster Storage	Reporter:	krishnaram Karthick <kramdoss>
Component:	heketi	Assignee:	Raghavendra Talur <rtalur>
Status:	CLOSED ERRATA	QA Contact:	Nitin Goyal <nigoyal>
Severity:	high	Docs Contact:
Priority:	unspecified
Version:	cns-3.9	CC:	akhakhar, akrishna, bkunal, hchiramm, jarrpa, jmulligan, kramdoss, madam, pprakash, rcyriac, rhs-bugs, rtalur, sankarshan, storage-qa-internal, suprasad
Target Milestone:	---	Keywords:	FutureFeature
Target Release:	CNS 3.10
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:		Doc Type:	Enhancement
Doc Text:	Earlier, all block hosting volumes were created only with group gluster-block option set on the Gluster volume. With this update, this feature enables an administrator to define options while creating block hosting volumes by setting block_hosting_volume_options in heketi.json OR by setting environment variable HEKETI_BLOCK_HOSTING_VOLUME_OPTIONS. However, Red Hat recommends not removing the group gluster-block option, but additional options can be added to this option. For example, to make block hosting volumes be of arbiter type, set the option to "group gluster-block,user.heketi.arbiter true".	Story Points:	---
Clone Of:		Environment:
Last Closed:	2018-09-12 09:23:45 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:	1568860

Description krishnaram Karthick 2018-06-15 07:03:26 UTC

Description of problem:
block hosting volumes are created automatically during the deployment of CNS. Until CNS 3.9, the block hosting volumes were of type replica 3. With arbiter volume type being introduced in CNS 3.10, an option or a provision should be made for the block hosting volume to be a arbiter volume by default. This has to be handled during the initial deployment. 

Also, when the block hosting volume is out of space, new block hosting volume being created automatically should also be of type arbiter instead of replica 3.

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


How reproducible:


Steps to Reproduce:
1.
2.
3.

Actual results:


Expected results:


Additional info:

Comment 3 Raghavendra Talur 2018-07-10 14:31:00 UTC

patchposted at https://github.com/heketi/heketi/pull/1253

Comment 4 Raghavendra Talur 2018-07-12 09:16:43 UTC

How it works,

    When a block hosting volume is created, either by auto creation when
blockvolume is requested or by heketi cli volume create --block true option,
heketi applies "group gluster-block" gluster option.

    However, if one wants to have a arbiter volume as a block hosting volume,
there was no way specify it.
    
    This feature allows users to specify the options used for new block hosting
volume creation 
a. Config file option under glusterfs section in heketi.json: block_hosting_volume_options
b. environment variable: HEKETI_BLOCK_HOSTING_VOLUME_OPTIONS

    Users can replace the default option list by setting these options. Hence,
to create block hosting volumes that are also arbiter volumes, set the value to
"group gluster-block,user.heketi.arbiter true".  It is NOT recommended to
remove group gluster-block option from block hosting volumes. During volume
creation, skip-block-hosting-volume-options flag can be set to true to not set
options provided in conf or env. However, GlusterVolumeOptions that are
provided in the command will be applied.

   Note that no storageclass changes are required. However, editing of
heketi.json config and/or env of pod is required.

Comment 5 Raghavendra Talur 2018-07-12 09:25:17 UTC

Jose, is there a way to provide env to heketi deployment config as part of the installation?

As you can see from comment 4, we have a new option HEKETI_BLOCK_HOSTING_VOLUME_OPTIONS. This tweaks how block hosting volume is created. If one were to set this option after the logging and metrics pods are deployed, it would be too late.

Comment 6 Jose A. Rivera 2018-07-12 11:51:57 UTC

In general, that can be done. It's obviously too late to provide that got OCP 3.10 GA, so it would have to go into 3.10.z which would be at the earliest three weeks after GA.

Comment 7 Humble Chirammal 2018-07-13 07:10:21 UTC

Fixed in version : rhgs-volmanager-rhel7:3.3.1-20

Comment 9 Nitin Goyal 2018-07-30 06:22:13 UTC

Hi,

I have verified this bug ob below Rpm and container Image. It is working fine. Hence marking this as verified.

Rpm: ->
heketi-7.0.0-5.el7rhgs.x86_64

Container Image: ->
rhgs-volmanager-rhel7:3.3.1-21

Comment 10 Nitin Goyal 2018-07-30 06:35:49 UTC

I had mention wrong container Image by mistake the correct one is below.
rhgs-volmanager-rhel7:3.3.1-22

Comment 11 Anjana KD 2018-09-07 08:10:06 UTC

have updated the doc_text field. Kindly review.

Comment 12 John Mulligan 2018-09-07 17:39:19 UTC

Doc Text looks OK

Comment 14 errata-xmlrpc 2018-09-12 09:23:45 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://access.redhat.com/errata/RHEA-2018:2686