Bug 1375296

Summary:	RFE: Support for installing SCL using NFS mount of /opt/rh
Product:	Red Hat Software Collections	Reporter:	Lee Whitty <lwhitty>
Component:	scl-utils	Assignee:	Joe Orton <jorton>
Status:	CLOSED CURRENTRELEASE	QA Contact:	BaseOS QE - Apps <qe-baseos-apps>
Severity:	unspecified	Docs Contact:
Priority:	medium
Version:	unspecified	CC:	hhorak, jorton, kanderso, lkuprova, lwhitty
Target Milestone:	---	Keywords:	FutureFeature
Target Release:	3.4
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2019-12-17 14:58:11 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 Lee Whitty 2016-09-12 15:55:18 UTC

Description of problem:

Customer would like to install SCL on an NFS server, and allow developers to mount this directory in order to use SCL. They were able to accomplish this using this blog article: http://developers.redhat.com/blog/2015/06/29/maintain-software-collections-easily-on-thousands-of-machines-using-scl-register/. They found that the NFS approach does work, however, they would have to maintain their own enable files to escape the hardcoded dependencies on /opt/rh in the stock ones.


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

NA


How reproducible:

- Install SCL on server
- Export /opt/rh directory
- Mount /opt/RH read only on developer system


Steps to Reproduce:
1.
2.
3.

Actual results:

Above steps work, but each developer system needs to be registered. Also, although it works, it isn't officially supported.


Expected results:

Would like to have developers mount /opt/rh from an NFS server (not necessarily a RHEL server exporting /opt/rh), and have developers able to use the SCL tools.


Additional info:

This is an RFE request, but didn't see option to categorize this as an RFE.

Comment 1 Honza Horak 2016-09-13 06:38:53 UTC

(In reply to Lee Whitty from comment #0)
> Customer would like to install SCL on an NFS server, and allow developers to
> mount this directory in order to use SCL. They were able to accomplish this
> using this blog article:
> http://developers.redhat.com/blog/2015/06/29/maintain-software-collections-
> easily-on-thousands-of-machines-using-scl-register/. They found that the NFS
> approach does work, however, they would have to maintain their own enable
> files to escape the hardcoded dependencies on /opt/rh in the stock ones.

Not sure I understand this. Is the problem the dependency itself, which is not handled transparently? Or is there another reason why clients need to maintain own enable files?

> Expected results:
> 
> Would like to have developers mount /opt/rh from an NFS server (not
> necessarily a RHEL server exporting /opt/rh), and have developers able to
> use the SCL tools.

I don't understand here either -- what else could it be other than RHEL server?
And is the problem here the need to run `scl register ...` on all clients?

Comment 3 Lee Whitty 2016-09-14 12:14:22 UTC

2016-09-14 - Reach out to customer and asked him to respond directly in order to clarify.

Thanks.

-Lee

Comment 6 Joe Orton 2019-12-17 14:58:11 UTC

An NFS-mounted /opt/rh will work where enabled in the collection, but requires running `scl register` on each client machine which mounts /opt/rh.  No additional feature work is planned here.