I had (or have) the same concerns. An O'Reilly book i was looking at says
that the Perl DBI for postgres allows only table level locking.
And doesn't go into, at all, the concept of snapshots for record
stability.
So does anyone know anything about this? Are record / row locks comitted
or table locks? And what about the perl interface that postgres comes
with?
thanks for any info
.jtp
On Thu, 13 Jul 2000, ernie cline wrote:
> Tom Lane wrote:
> >
> > SELECTs certainly don't block each other (except the SELECT FOR UPDATE
> > variety).
>
>
> > Need more detail ...
>
> What else do you need to know? Postgres 7.0 compiled with default
> options, running on a VA Linux Fullon box, kernel 2.2.13, running VA
> Linux/Redhat 6.1. Box is a single processor pentium 3 500mhz, 128megs
> of ram, apache 1.3.9. Using Perl DBI interface to access postgres. A
> very small html page is built (about 4.5k), and it takes about 2.6
> seconds to execute and build the html. Running 2 of the pages at once,
> it takes about 5 seconds for them to complete. Not total, but 5 for
> each one. This is using just unix 'time' command, and moving my mouse
> to another rxvt window and hitting enter after first hitting it in
> another window ;).
>
> Is using Perl DBI slow? Would writing the cgi in C speed things up?
> (Please no holy wars for perl vs. C performance! <G>)
> Any help would be appreciated ...
>
> -ernie