Обсуждение: Re: Is PostgreSQL ready for mission criticalapplications?

Kaare Rasmussen <kar@webline.dk> wrote:
:> But I am not imagining the random "I have rolled back the current
: transaction
:> and am going to terminate your database system connection and exit."
: messages.

: I'm wondering if you ever reported these problems to this list or the
: the hackers list? I've been reading both regularily and don't recall
: seeing this descussed before, but maybe I'm wrong.

: Generally I find the responsiveness from the development team way better
: than any commercial products. _All_ problem reports are treated with
: concern. So if you didn't report them before, please take the time to
: document your experience and send the problem report to the correct
: place.

No I haven't reported them. I have reported a minor bug that I could reproduce
to the bug tracking system. But all the other problems I had, were, as I said,
not reproducable. I tried to come up with a small test case for some of the
bugs and failed. Sure I can report them all, but the developers will tell me,
and rightly so, that they can't do anything with it, because they can't
reproduce it. I know that this is not very helpful, but I know no easy way out
here.

Jochen
--
Jochen Topf - jochen@remote.org - http://www.remote.org/jochen/

On Tue, Nov 23, 1999 at 11:48:58AM +0100, Jochen Topf wrote:
> Kaare Rasmussen <kar@webline.dk> wrote:
>
> No I haven't reported them. I have reported a minor bug that I could reproduce
> to the bug tracking system. But all the other problems I had, were, as I said,
> not reproducable. I tried to come up with a small test case for some of the
> bugs and failed. Sure I can report them all, but the developers will tell me,
> and rightly so, that they can't do anything with it, because they can't
> reproduce it. I know that this is not very helpful, but I know no easy way out
> here.
>

Jochen -
One thing I've noticed following the HACKERS list for a while, is that
the developers are very good at helping turn not-reproducible bugs into
reproducible ones, as well. By that I mean, they can often help narrow
the scope of variables to consider significant, based on the exact error
message, even if it's a rare bug. Sometimes, just knowing a particular
error can be triggered on anyone's system can lead a developer in the
right direction when back-tracing possible execution paths, and often
suggest some simple tests. As I said, I've seen it happen more than
once in the last year. I understand your reticence in reporting bugs
without a simple reproducible test case: it's worthless when contacting
your typical commercial vendor, where your talking to some code tester,
who probably isn't a developer.

Heck, I recently had a case where I submitted a bug, with simple
reproducible case to a commercial vendor. I even diagnosed exactly what
the cause was. When that company got ready to release a new point release,
I was contacted by their QA department, because they couldn't reproduce
the bug, and wanted me to pull their beta, upgrade my working mission
critical system, and see if the bug was still there! Apparently, they
couldn't be bothered to keep an installation of their currently released
product around to test on, but I could be bothered to potentially break
my system for them. Sheesh!

So, in short, report those bugs, even if elusive. At best, the bugs
will get stamped out faster, at worst, you'll point out pontentially
suspicious sections of code to the developers.

Ross

P.S. I noticed you mentioned the 'bug tracking system'. I know that a
web based bug tracker was tried out earlier this year, but was abandoned
in favor of the mailing lists.
--
Ross J. Reedstrom, Ph.D., <reedstrm@rice.edu>
NSBRI Research Scientist/Programmer
Computer and Information Technology Institute
Rice University, 6100 S. Main St.,  Houston, TX 77005

> Jochen -
> One thing I've noticed following the HACKERS list for a while, is that
> the developers are very good at helping turn not-reproducible bugs into
> reproducible ones, as well. By that I mean, they can often help narrow
> the scope of variables to consider significant, based on the exact error
> message, even if it's a rare bug. Sometimes, just knowing a particular
> error can be triggered on anyone's system can lead a developer in the
> right direction when back-tracing possible execution paths, and often
> suggest some simple tests. As I said, I've seen it happen more than
> once in the last year. I understand your reticence in reporting bugs
> without a simple reproducible test case: it's worthless when contacting
> your typical commercial vendor, where your talking to some code tester,
> who probably isn't a developer.

Another thing is that when we hear a few unrepeatable bug reports, we
usually get cracking at finding the cause.  Also, sometimes some one
will come along with a reproducable case, and we jump on that case very
quickly because we know others are seeing the problem too.

--
  Bruce Momjian                        |  http://www.op.net/~candle
  maillist@candle.pha.pa.us            |  (610) 853-3000
  +  If your life is a hard drive,     |  830 Blythe Avenue
  +  Christ can be your backup.        |  Drexel Hill, Pennsylvania 19026