Thanks very much!
I've decided to go straight to 8.1 though. There are just too many
performance improvements at this point that I might regret not having
and I don't want to do a dump reload again. I am about to compile it
now. If it isn't a panic grade failure in the latest 8.1 code then
I'd just assume take the stock release source code. I don't care at
all if this kills one connection at the ultra-low frequency with
which it occurs but what I can't have is the whole server rebooting
itself in the middle of processing hundreds of transactions. Once
that happens all of the web clients hang onto their bad connections
and then eventually die. Considering that I'm moving to 8.1 and am
not too familiar with applying patches am I crazy for just going with
the stock 8.1 code?
On Jan 20, 2006, at 10:36 PM, Tom Lane wrote:
> Rick Gigger <rick@alpinenetworking.com> writes:
>> I don't know if 2K could have passed since the last checkpoint.
>> ...
>> now that I think about it I was getting about 400 pages requests /
>> minute and each of those would have have been doing at least 2
>> transactions so yes, I guess that is very likely.
>
> Good, 'cause if you didn't have a couple thousand transactions between
> checkpoints then we need another theory ;-)
>
>>> You realize of course that that's pretty old ...
>
>> Yes. I will be upgrading immediately.
>
> You'll want to include this patch:
> http://archives.postgresql.org/pgsql-committers/2006-01/msg00289.php
> (or see adjacent messages if you plan to move to something newer than
> 7.3.*). We probably will not put out another set of releases until
> next month, unless something really big comes along. This one doesn't
> qualify as really big IMHO, because it's not a PANIC-grade failure in
> the later branches. But having been burnt once, I'm sure you'll want
> a patched copy ...
>
> regards, tom lane
>