Status: proposal
Simplified Linux Server (SLS) SIG
1. Overview
The Simplified Linux Server (SLS) SIG provides a common platform for delivering turnkey CentOS-based solutions that are managed via a web and/or REST-based interface. The focus is on taking the various application silos and providing a cohesive solution to simplify deployment and maintenance.
2. Big Picture
Database servers, web servers, file servers, proxy servers and many other applications use CentOS as a starting point. Typically, these solutions are deployed using in-house tools (e.g. Puppet), the command line, and plain old Linux expertise. This limits CentOS’ ability to be considered by many IT solution providers due to high knowledge barrier required to obtain a comprehensive solution. Consider this example: a system administrator needs to create a CentOS-based mail server on a VM or cloud instance. The administrator installs CentOS, along with:
- Postfix
- Cyrus IMAP (or Dovecot)
SpamAssassin
- ClamAV (from EPEL)
- OpenSSL for user certificates
They pull it all together with OpenLDAP for user and group management. Now all that needs to be done is configuring all those open source packages into a cohesive mail solution: /etc/postfix/main.cf, /etc/clamd.conf, /etc/imapd.conf, etc. When that's done, they breakout the OpenSSL documentation to figure out all the security certificate jargon.
The SLS SIG provides a platform for pulling all of these disparate applications together, and then wrapping up the solution in a web/REST-based interface. This is different from a management focused SIG. Unlike management interfaces (ie. webmin, et al), SLS focuses on results-based administration rather than alternate ways to modify configuration files. Rather than showing all the options, it shows the best practices results and takes care of configuring the system to those results-based configurations. The result is simplified, scalable, and unified configuration rather than granular customization. Additionally, this framework lowers support requirements and costs by making server administration accessible to GUI-oriented administrators and by creating similar deployment scenarios with fewer permutations.
3. Deliverables
The SIG members are involved in the development and maintenance of turnkey solutions based on CentOS.
Software Maintenance - This SIG will co-ordinate maintenance of common upstream packages in order to avoid duplication and provide a better experience for all CentOS users.
Quality Assurance - This SIG provides solutions with tight integration amongst the various applications in CentOS. This provides a unique opportunity to provide additional QA testing (e.g. Samba + OpenLDAP).
4. Interface with Other SIGs
- Core
- Cloud - SIG solutions already run in the cloud (Amazon, Linode) and co-operation would reduce duplication of effort
5. Mechanism To Deliver Software
The SLS SIG will begin with a single repository to house development, testing, and production code with the appropriate tagging and branching to ensure that developers and users are always getting the appropriate level of stable code.
6. Who is involved
Filippo Carletti from Nethesis (NethServer)
David Loper from ClearFoundation (ClearOS)
Peter Baldwin from ClearFoundation (ClearOS)
John Crisp from SME Server
Anyone from the community could join and contribute.
7. Software
http://www.clearfoundation.com/docs/developer/packaging/clearos_6.x/start
http://dev.nethserver.org/projects/nethserver/wiki/Packages#From-third-parties
8. Future potential collaborators
- Everyone involved with devops tasks
- Upstream developers, i.e. developers of software packaged by the SIG
9. Why this SIG is important
- To foster adoption of CentOS in diverse/new scenarios and situations, especially where a sysadmin is not available
- To let people without Linux experience to start using CentOS as a friendly and turnkey system
- To provide CentOS users a single reference site to download rpm packages commonly used in Small Businesses
10. Resources needed
We expect the following from the CentOS Project:
- Buildsystem access to build rpms
- QA resources to test the devel/testing grade code
- Release mechanism to push released code publicly
- Git repositories to maintain code (including commit access for SIG contributors)
- Wiki and documentation space for content related to the SIG
- Bugs and issue tracker project to track user issues and incident reports
- Private security email address for issues that need to be kept private
- CentOS Project involvement to liaison with CentOS Core SIG to sync and test with potential CentOS Linux releases
11. Current Status
A plurality of founding members of the SIG have packages that could be easily migrated to the CentOS repo.
12. Bootstrap process
The first step in this bootstrap is ensure that all expectations are aligned with the CentOS board. Once this SIG is established as a necessary resource the founding members will migrate their packaging and deployment resources over to CentOS infrastructure and start promoting development work as a participatory effort in that setting.
We will aim to document this process as it happens so that other wishing to join at a later date will have a clear pathway on how to participate. Part of this process will be a discussion with the CentOS board, and community-at-large, on the best way to deliver services on top of the CentOS platform.
13. Comparable focus in the marketplace
Already, integration for this type of relationship exists in the marketplace. A notable example is the Zentyal project within Ubuntu. See https://help.ubuntu.com/12.04/serverguide/zentyal.html
Another work in progress is the Fedora Server Working Group: http://fedoraproject.org/wiki/Server