[Info-ingres] Was row-level locking ever an "extra cost" option?

[Paul Mason]

I remember in my brief stint as an OpenROAD programmer (OR 3.0, win32s on Windows 3.1, a good day was only two reboots!) being told that even though the new system we were writing would run on OI 1.2 which had outer joins, I should write it as if it were going to run on 6.4 which didn't - because it was a new feature and therefore probably buggy. ISTR writing a very long snarky comment in the code about it.

From: info-ingres-bounces at lists.planetingres.org <info-ingres-bounces at lists.planetingres.org> On Behalf Of Roy Hann
Sent: 10 July 2019 10:28
To: info-ingres at lists.planetingres.org
Subject: Re: [Info-ingres] Was row-level locking ever an "extra cost" option?

Chris Clark wrote:

> On Friday, July 5, 2019 at 10:04:04 AM UTC-7, Karl and Betty Schendel wrote:
>> > On Jul 5, 2019, at 6:43 AM, Roy Hann <specially at processed.almost.meat<mailto:specially at processed.almost.meat>> wrote:
>> >
>> > My memory fades. Was row-level locking ever an "extra cost" option?
>> >
>> > I don't think it was but funny things happened in the past, like
>> > creating tables (briefly) being an extra cost option...!
>>
>>
>> I'm not 100% certain, but I don't believe it was ever license-controlled.
>
> I can confirm Marty is correct, it was added as a new feature to the base
> product. You did need to use larger (than 2Kb) page sizes to make use
> of it so some minor DBA work was needed if upgrading.

Thank you Chris, Karl, Marty.

Row-level locking was released in (I think) Ingres II 2.0, circa
circa 1997, which was also the release that supported so-called large
pages.

The reason I was asking about this was curiosity about why the system I
am looking at just now has loads of homebrewed optimistic locking code.

Your combined answers make me wonder if they were just unwilling to
pioneer the use of two new features (in 1997). Code-your-own
optimistic locking may have seemed like a safer idea at the time. (I am
not sure they were right then, and I am certain it's a problem now.)

I think they could have done other things to avoid the contention they
were expecting (if not experiencing), but here we are.

Roy

_______________________________________________
Info-ingres mailing list
Info-ingres at lists.planetingres.org<mailto:Info-ingres at lists.planetingres.org>
https://lists.planetingres.org/mailman/listinfo/info-ingres<https://lists.planetingres.org/mailman/listinfo/info-ingres>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.planetingres.org/pipermail/info-ingres/attachments/20190710/7b9beb99/attachment.html>