Nils - I would certainly agree that diskless/swapfull clients are an important part of the OS makeup, but a project would be a better place to start. The tool enhancement of 'smdiskless' and the OS services necessary to provide diskless support are welcome.
being an old Sun 3/60 administrator (maybe redundant) I would be more than happy to help... rich Nils Goroll wrote: > Hi all, > > I would like to push some diskless topics within the OpenSolaris > community and I would like to contribute a working framework + tools > which I am using in production. > > I am posting this to a wider range of lists because somehow the > diskless topic does not fit into any of the exising communities - or > many of them. > > The following communities relate to diskless, in one way or another, > IMHO: > > HA Clusters > Installation and Packaging > NFS > OS/Net (ON) > Xen > ZFS > Zones > > I feel that there really should be a Diskless community, but I dont > know if there is enough interest in the topic to justify a > community. > > If diskless was to become a project, this would be my project request: > > > - A short (one or a few words) synopsis of the Project's purpose, > and an optional shortened or decorative name (see 4.1). > > Diskless / OS-Service > > - A list of Community Groups sponsoring the Project. > > Who would? > > - A list of Participants constituting the initial Project Team, > including a single individual who will serve as the preferred > point of contact for matters regarding the Project. > > me (nigoroll) > > - A one-paragraph description of the Project, for an audience of > Participants who may not be familiar with the area in which work > is proposed. This should contain a brief description of the > problem(s) the Project is expected to solve, and of the manner in > which it will do so. > > SHORT: > > The project aims to revive the almost forgotten diskless client > technology which has the power to centralise patch and change > management and thus enables another path to efficient and cost > effective management of larger installations by improving the > tools and concepts for diskless client deployment and manage- > ment. > > - A listing of related ongoing or proposed Projects, including > information about any dependencies on or by this Project and any > duplication of purpose or overlap with other ongoing work. This > listing should also include the name of the consolidation the > Project Team is targeting, if applicable. > > xen - uses diskless technology to boot > newboot - changed the netboot process and might need to be amended by this > project > > what else? > > - Optionally, additional information which may be of interest to > prospective Project Team members and/or consumers of the Project's > output. > > Solaris can run off the network without the need for any local > disks. The basic concept of "diskless clients" dates back to the > 1990s when Sun built workstations without any disks (SPARCstation > ELC). > > Nowadays, the main use for diskless boot technology is to jumpstart > systems from an install server, but it is also being used productively > in real life data centres. > > Diskless clients use NFS to access a split Solaris installation: A > dedicated (per diskless client) root filesystem (read-write) and a > shared /usr filesystem (read-only). Besides the advantage that this > concept ensures that many servers use the *same* operating environment > (not just copies of it), diskless client technology centralises patch > and change management and thus allows for efficient and cost effective > management of larger installations - at the expense of more complexity > on the central server (OS-Server). > > Unfortunately, the tools to maintain a diskless client environment > built into (Open)Solaris today, are outdated, erroneous, inefficient - in > short, unsuable. The project aims to improve this situation. > > To kick off the project, I (nigoroll) can contribute: > > - An overhauled OS-Server structure concept which is being used in > production and allows to run various releases / patch levels off a > single OS-Server (can be on an HA cluster). > > - The necessary information on how to serve highly available > OS-Services on Sun Cluster > > - A "proof of concept"-stage code base in Perl, which re-implements > diskless client patching and does away with many drawbacks of the > dcpatch ksh script. > > - Suggestions, concepts and implementations on how to make the PXE > boot process more secure. > > > > What I (nigoroll) CAN NOT provide is any guarantee as to how much time > I will be able to invest into this project. I will try my best, though. > > see a Sun SPARCstation ELC http://john.ccac.rwth-aachen.de:8000/alf/sparc_elc/ > > > This message posted from opensolaris.org > _______________________________________________ > install-discuss mailing list > install-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/install-discuss >
