Commit the applied FixPak??
in [DB2]
|
From: The Boss on 30 Mar 2007 17:11 Darin McBride wrote: > Anurag wrote: > >> I have response from IBM (had opened a PMR) on these 2 questions. > > Can you send me the PMR number? Was it against DB2 or AIX? These > responses don't make a lot of sense to me - so I'd like to pursue it. > >> Lemme share this with you all: >>>> **** Can I COMMIT the upgrades without any disruption to >>>> applications / users using the existing DB2? >> NO, unless it's a Client Installation. >> >>>> **** Do I require downtime to COMMIT? >> YES. You need db2stop and ipclean before COMMITing. >> >> So, after all the accolades from management, I now have to request >> downtime for COMMIT (maybe I do that over Easter weekend and nobody >> makes a fuss :) > > Of course, it never hurts to be safe rather than sorry, but, still - > it'd be nice to be able to just commit. > > (Nothing in the fixpack readme talks about shutting down the server > prior to committing, which I presume would be there if DB2 developers > knew of any reason why you would need to do so.) I think you are right. From the AIX Commands Reference (installp command): <quote> When a fileset update is applied to the system, the update is installed. The current version of that software, at the time of the installation, is saved in a special save directory on the disk so that later you can return to that version if desired. Once a new version of a software product has been applied to the system, that version becomes the currently active version of the software. <..> When updates are committed with the -c flag, the user is making a commitment to that version of the software product, and the saved files from all previous versions of the software product are removed from the system, thereby making it impossible to return to a previous version of the software product. Software can be committed at the time of installation by using the -ac flags. Note that committing already applied updates does not change the currently active version of a software product. It merely removes saved files for previous versions of the software product. </quote> -- Jeroen
From: Ian on 30 Mar 2007 23:32 Anurag wrote: > I have response from IBM (had opened a PMR) on these 2 questions. > Lemme share this with you all: >>> **** Can I COMMIT the upgrades without any disruption to applications / users using the existing DB2? > NO, unless it's a Client Installation. > >>> **** Do I require downtime to COMMIT? > YES. You need db2stop and ipclean before COMMITing. > > So, after all the accolades from management, I now have to request > downtime for COMMIT (maybe I do that over Easter weekend and nobody > makes a fuss :) > > I concur with Darin. This is wrong. I have committed applied fixes with the database running more times than I can count. And I have never had an issue.
From: Gert van der Kooij on 31 Mar 2007 03:46 In article <4DkPh.114122$Ko5.70779(a)newsfe08.phx>, ianbjor(a)mobileaudio.com says... > Anurag wrote: > > I have response from IBM (had opened a PMR) on these 2 questions. > > Lemme share this with you all: > >>> **** Can I COMMIT the upgrades without any disruption to applications / users using the existing DB2? > > NO, unless it's a Client Installation. > > > >>> **** Do I require downtime to COMMIT? > > YES. You need db2stop and ipclean before COMMITing. > > > > So, after all the accolades from management, I now have to request > > downtime for COMMIT (maybe I do that over Easter weekend and nobody > > makes a fuss :) > > > > > > I concur with Darin. This is wrong. > > I have committed applied fixes with the database running more times than > I can count. And I have never had an issue. > > Then you concur with Anurag, Darin said no downtime was needed.
From: The Boss on 31 Mar 2007 11:02 Gert van der Kooij wrote: > In article <4DkPh.114122$Ko5.70779(a)newsfe08.phx>, > ianbjor(a)mobileaudio.com says... >> Anurag wrote: >>> I have response from IBM (had opened a PMR) on these 2 questions. >>> Lemme share this with you all: >>>>> **** Can I COMMIT the upgrades without any disruption to >>>>> applications / users using the existing DB2? >>> NO, unless it's a Client Installation. >>> >>>>> **** Do I require downtime to COMMIT? >>> YES. You need db2stop and ipclean before COMMITing. >>> >>> So, after all the accolades from management, I now have to request >>> downtime for COMMIT (maybe I do that over Easter weekend and nobody >>> makes a fuss :) >>> >>> >> >> I concur with Darin. This is wrong. >> >> I have committed applied fixes with the database running more times >> than I can count. And I have never had an issue. >> >> > > Then you concur with Anurag, Darin said no downtime was needed. And that's what Ian is agreeing with. I guess you mix up the meaning of 'to concur' with Dutch 'concurreren'. http://www.m-w.com/dictionary/concur -- Jeroen
From: Gert van der Kooij on 31 Mar 2007 12:40
In article <460e7808$0$69886$e4fe514c(a)news.xs4all.nl>, usenet(a)No.Spam.Please.invalid says... > Gert van der Kooij wrote: > > In article <4DkPh.114122$Ko5.70779(a)newsfe08.phx>, > > ianbjor(a)mobileaudio.com says... > >> Anurag wrote: > >>> I have response from IBM (had opened a PMR) on these 2 questions. > >>> Lemme share this with you all: > >>>>> **** Can I COMMIT the upgrades without any disruption to > >>>>> applications / users using the existing DB2? > >>> NO, unless it's a Client Installation. > >>> > >>>>> **** Do I require downtime to COMMIT? > >>> YES. You need db2stop and ipclean before COMMITing. > >>> > >>> So, after all the accolades from management, I now have to request > >>> downtime for COMMIT (maybe I do that over Easter weekend and nobody > >>> makes a fuss :) > >>> > >>> > >> > >> I concur with Darin. This is wrong. > >> > >> I have committed applied fixes with the database running more times > >> than I can count. And I have never had an issue. > >> > >> > > > > Then you concur with Anurag, Darin said no downtime was needed. > > And that's what Ian is agreeing with. > I guess you mix up the meaning of 'to concur' with Dutch 'concurreren'. > http://www.m-w.com/dictionary/concur > > I guess so :) Sorry, I withdraw my comment. |