Angela Byrne - Solaris Sustaining wrote: > So another process breakage hits us. > - redelivering an existing package to RE > > RE seem unable to always refuse such submissions > - and I do not know the details why :-( > - How do I find out who is doing it so we can stop it. > > So should we attempt to automate a fix or let the customers > try and do it manually. ? > > If we are unsure about the correctness of doing it automatically, > then I think we should not advise to do it manually either. > > We need to execute the fix and prove it, then see if we can > (or should) automate it. > > I would also say that telling customers to go get SUNWicda > from the u3 install CD is not appropriate. > > Making it available elsewhere is also problematical. > greeeat... > > If we go with manual means, I would want the prepatch script message > output to be improved over sending a Sun Alert out. (or both with the > script referring to the Sun Alert) > > > Angela > > > > Enda O'Connor ( Sun Micro Systems Ireland) wrote On 10/17/06 16:23,: > >>Angela Byrne - Solaris Sustaining wrote: >> >> >>>Angela Byrne - Solaris Sustaining wrote On 10/17/06 14:54,: >>> >>> >>> >>>>So I assume this script was put in place for a reason.. >>>>As such, we need to follow this idea to genesis add SUNWicda to the miniroot >>>>- if that is possible... >>> >>> >>>So Peter Harvey challenged this saying that the package SUNWicda >>>would already be present but that the content would not contain the >>>required SBD changes. >> >> >>But the one in U3 is a different VERSION >>in U2 we have >>VERSION=11.10,REV=2005.11.12.14.23 >>in U3 we have >>VERSION=11.10,REV=2006.08.16.17.07 >> >>So we need to pkgrm the old one and pkgadd the new one. >>Looks like FCS is differnet from U2 which is different from U3 VERSION, >>suspect u1 is also a different version. >> >>I think we will need to do this manually. >>That is a doc that explains >>1 now to remove the old SUNWicda >>2 install the new SUNWicda from U3 >> >>Doing all this at this late stage in a KU prePatchis too dangerous. >> >> >>A seperate patch coudl be created, but the KU would not depend on that >>patch ( unless it delivered non miniroot patch packages as well, and >>packages that are in the minimal metacluster as well ) >>That is still a lot of work/testing at this late stage >> >>So perhaps a info doc/sun alert to detail how to do this manually. >> >>Enda >> >> >> >> >>>As such, could we just add a regular sparse patch pkg for SUNWicda into the >>>KU to contain the update 3 changes. >>> >>>For normal patch installation, SUNWicda is not present so this would not >>>affect the patch install. >>> >>>But when we patchadd -C, the package SUNWicda would be present and would >>>automatically get the required update. >>> >>>Is this reasonable or is there some other strangeness in the miniroot >>>to stop this. >>> >>>Angela >>> >> >> > Obviously I prefer automated solutions, perhaps a seperate patch that is NOT required by the KU that delivers a prePatch that 1 removes SUNWicda if FCS/U1/U2 2 if 1 was done then pkgadd the new one from U3
I'd be very slow to include this in the KU as we are rejuvenating the KU and would not be able to fix any problem that arose from this new logic. Then the KU would have to say "please install patch xxxxxx before applying this KU ", only if we called patchadd -C obviously. Enda
