Re: IDS10 stability VS IDS11 features (aka "for IBM")
Posted in 2008
Topics: High Availability & Replication, Installation, Setup & Upgrades, Server Administration, Logging & Checkpoints, Versions, Editions & End-of-Life
Davorin Kremenjas wrote: > Right, > > probably a very lame try to imitate Gumby, but I'll give it a try :) > > I'd like to hear you opinions on this topic, here is mine. > > IDS11.50 is around for a good deal time now, _loads_ of new features. > I can see a lot of them making my DBA life easier so I'm looking to > upgrade reasonably soon. > > On the other hand, running IDS 10.FC8(ish) very stable in production > with all HDR/ER goodies I'm a happy man. Also, I'm doing regular > upgrades every time a new patch comes out (looking to roll out > IDS10.FC9W2 in the next few weeks on all prod instances). But, every > time I read the release notes/bug fix lists they scare the s**t out of > me. Too many crashes for my taste for (at least seemingly, R&D please > don't get offended now) trivial reasons (update stats or whatever). > > This makes me conclude the next: IDS10 support is ending in less than > 2 years (30 Sep 2010 according to IBM website) wouldn't it be nice to > "freeze" IDS10 code and work ONLY on bug fixes? Whoever wants to trade- > off features for stability, great, here's 11.50, just upgrade. For the > rest of us, we'd like to sleep well :) And in two years time, I'm sure > v11.50 will be just as stable as v10 is now, we'll have vNext too, so > the whole story can be repeated. > > Does this make any sense to the rest of you? Or am I just being too > naive? (Now I know at least the Clown will answer :o)) > > Cheers > > Davorin I think you didn't touch some important points in your exposure: - When you "fix" a bug you may introduce another - There are new features and new FEATURES. I mean small and big features. IDS 10 has seen small features. Ideally they would not cause critical bugs. To give you an example, we've seen view folding but not non-blocking checkpoints... - IBM has a very good practice: When a bug is found in one family it is checked against all currently supported families. Some bugs you see in the IDS 10 fix list were probably found in IDS 9.x, 11.10 or 11.5. This can happen because all families share a lot of code. Also, if a "small" feature can be introduced in previous families (because some code is shared and the changes are not big) this benefits lots of customers (who cannot make "big" upgrades - they may require 3rd party application certification) Besides this I would not agree with the idea that "IDS 10 is stable because it's old" and "IDS 11.5 is unstable because it's new". Again with an example: 10.00.xC7 became an "historical release", and not for the good reasons. IDS 10.00.xC8 seems pretty good. I haven't seen big issues with any 11.50.x release (obviously I know the majority of the customers hasn't upgrade yet). Regards. -- Fernando Nunes Portugal http://informix-technology.blogspot.com My email works... but I don't check it frequently...
On Dec 10, 11:02 pm, Fernando Nunes <domusonl...@gmail.com> wrote: > I think you didn't touch some important points in your exposure: > - When you "fix" a bug you may introduce another > - There are new features and new FEATURES. I mean small and big features. IDS > 10 has seen small features. Ideally they would not cause critical bugs. To give > you an example, we've seen view folding but not non-blocking checkpoints... > - IBM has a very good practice: When a bug is found in one family it is checked > against all currently supported families. Some bugs you see in the IDS 10 fix > list were probably found in IDS 9.x, 11.10 or 11.5. This can happen because all > families share a lot of code. Also, if a "small" feature can be introduced in > previous families (because some code is shared and the changes are not big) > this benefits lots of customers (who cannot make "big" upgrades - they may > require 3rd party application certification) > > Besides this I would not agree with the idea that "IDS 10 is stable because > it's old" and "IDS 11.5 is unstable because it's new". Again with an example: > 10.00.xC7 became an "historical release", and not for the good reasons. IDS > 10.00.xC8 seems pretty good. I haven't seen big issues with any 11.50.x release > (obviously I know the majority of the customers hasn't upgrade yet). > > Regards. > > -- > Fernando Nunes > Portugal > > http://informix-technology.blogspot.com > My email works... but I don't check it frequently... Hi Fernando, thanks for your input, it makes a lot more sense now. Davorin