I just read this post by Mike Dietrich. Thanks Mike!
There were a couple of very welcome bits of news from that, and the MOS note (Doc ID 742060.1) it points to.
First, the extended support fee for 11.2.0.4 has been waived for another 18 months, taking it to the end of 2018 (start of 2019). This is really big for us as we have some projects which are gradually dying as we move to Oracle Cloud Apps and we don’t want to spend time upgrading them. This is going to save some money! I suspect a lot of people will be really happy about this!
Second, there are some “proposed” dates for the release of Oracle database 12.2.0.1 on-prem. As the doc says, nothing is cast in stone so be warned. This is welcome news for me because there’s is only so many ways you can say, “I don’t know”, when people ask you about it. 🙂
Probably more interesting than the 12.2.0.1 date is the proposed 12.2.0.2 date, which is later this year. In my testing so far, 12.2.0.1 has been a lot more robust than 12.1.0.1 was, but many would still probably wait for 12.2.0.2 before doing something major with it (see below). If that really does get released this year I think it would be great for uptake!
What does this mean if you are thinking, “Do I upgrade to 12.1 now, or wait for 12.2?” Just my opinion, but I would say the following, some of which may seem a little contradictory…
- I am not one of those people that believes you should never use the first Oracle release. You should test it. If it works, great. If it doesn’t, wait for it to get fixed with patches or a patchset, then use it. Your experience will depend on the feature set you rely on.
- The difference between releases is now *massive*! In real terms 12.2 is more like 13.1. I don’t think it would be a major exaggeration to say the patchsets are now what we used to know as releases and releases are now what we used to know as new versions. I hope Oracle scrap the whole release 1, release 2 naming approach. It’s confusing as its meaning has been lost, especially as new functionality is being delivered in PSUs and patchsets all the time…
- I don’t think it is terrible to wait for the first patchset of a release, but as mentioned previously, a patchset contains a lot of new functionality now. It’s not just bug fixes! Each PSU or patchset could introduce something new that breaks your stuff, so you can’t assume a PSU or patchset is safe. You have to test!
- My plans are to continue with the 12.1 upgrades that are already scheduled. At this point 12.1 is a known quantity and 12.2 is not! If there are projects with timelines that we could consider 12.2 for, I will, but only on the basis we test the sh*t out of it. 🙂
Overall, some great news!
Cheers
Tim…
Hi Tim,
I know all arguments against and for installing first Oracle releases. I made worse experiences., too. But my expectation is easy and clear: A software which costs nearly 40.000 per core plus 5.000 support per core/year should tested enough, robust enough and hardened to use it in production and not in any labs! Best regards Perer
Peter: I agree it should, but I would say two things.
1) Do you trust every bit of new software? This is not my mobile phone I’m talking about. 🙂
2) The product could be 100% perfect, but maybe a change or new feature affects how your software responds to it. You are not just testing the product. You are testing your software against that product. That’s why application regression tests are so important. I’m never going to assume a patch is safe if I have the option to prove it. 🙂
Cheers
Tim…