Using rhnChecksum and rhnChecksumType directly instead of rhnChecksumView seems
to fix the performance problem as well. 'explain analyze' output is included
to show the difference.
I saw the ill effects of using the same table twice within a subquery, but
honestly, it doesn't make a whole lot
The left join on rhnChecksumView is horrendously slow, exactly the same thing
that was happening with doing config file diffs. An export of a single channel
with 3000 packages takes hours compared to 10 minutes with this patch. I can
post the 'explain analyze' output if you'd like.
I'll brush
On Fri, Feb 03, 2012 at 03:55:53PM +, Parsons, Aron wrote:
> Can someone verify that this is a sound change? rhn-satellite-exporter is
> extremely slow on PostgreSQL when getting the file checksums. This speeds it
> up dramatically.
>
> It's the same issue we ran into with PostgreSQL doing
Can someone verify that this is a sound change? rhn-satellite-exporter is
extremely slow on PostgreSQL when getting the file checksums. This speeds it
up dramatically.
It's the same issue we ran into with PostgreSQL doing a left join on
rhnChecksumView in 'query_client_get_file'. PostgreSQL
On Sat, Nov 12, 2011 at 09:09:18PM +0100, Jan Pazdziora wrote:
> > Please find attached a proposed patch for this.
> > I've added database.control boolean directly in rhn.conf file.
> > I've modified the spacewalk-service script to manage this parameter. In all
> > cases, oracle is controlled, for