Stephen Hahn wrote: > * Milan Jurik <Milan.Jurik at sun.com> [2007-09-13 06:49]: >> Enda O'Connor ( Sun Micro Systems Ireland) p????e v ??t 13. 09. 2007 v >> 10:49 +0100: >> >>> Did anyone consider staying with SVR4 and just making the tools work >>> properly. i.e a green field site and using SVR4 build a complete >>> set of tools from ground up. > > Are we talking about the format, the set of interfaces, or the > current implementation? Are we including or excluding the patch > format, interfaces, and implementation? patching to be included, to my mind patching currently is just an after though, no sustaining is being done really.( by this I mean adding functionality to support say SMF ) patching in say 2.6 worked, but s10 had essentially the exact same set of tools ( with the addition of pdo, which I have other thoughts on really that we won't go into here :-) ).
IBM use SVR4 but they invest in both the tools and the process to deliver by said tools, and that is the difference. Same for HP, I have used their sustaining tools and they work as expected. In pretty much every update we are expected to somehow cram a major "feature" into patches and expect our users to be happy. This is not the IBM way for instance, they have two updates ( aka Technology Levels )a year, one of which has hardware support the other pure sustaining, so you can choose to go the sustaining route once a year in the fall and not have to have the pain of being forced to take zfs, TX and so on when not wanted. I don't want to seem against this proposal, when in fact I agree totally that we need major surgery, but I just worry that in one or two releases we will be back where we started. > >> Based on all already published info I saw - no in reality. SVR4 is not >> broken itself but they think that some new packaging system will solve >> their ignoring of feature delivering magically somehow, ignoring the >> flexibility of SVR4. > > "SVR4", in its current implementation of both packaging and patching, > is an unsafe, imprecise, and inadequate system. I believe the > implementation and interfaces are not worth repairing; I have done > maintenance programming and porting in the past, and believe my > assessment of costs is rational. I have not seen a better system ( tools withstanding ) really. RPM's and all that. I agree that our current approach is in dire need of overhaul. But I have not seen what we are going with yet, so from my point of view it's hard to tell. I did ask some time back ( more than once ) but got no answer as to what we were planning. I work very much in patch and packaging space so have a lot of experience, I have worked with HP and AIX among others, also have a lot of experience with Oracle patching ( or lack thereof ) as I worked in that space for a number of years. It would be useful to get a look at the prototype though :-) Enda > > As I've said before, we're asking the Install CG to host our > exploration, in which we invite others to participate. I am not > asking that other efforts in this area be denied, nor am I saying that > the goals of this effort are fixed for the duration. (I am resisting > scope expansion in certain directions when larger issues of > feasibility for this initial phase, however.) > > - Stephen >
