Thanks for the update! It makes sense. I've logged this as ticket 3528, so somebody could have a look at that subzone query behaviour a bit better.
On Tue, Mar 08, 2016 at 09:40:57PM +0100, Thomas Mieslinger wrote: > Hi, > > maybe you remember my message regarding suboptimal ddns update performance. > A colleague helped me to identify the list_subzone_query as the performance > killer. > > The original where clause is > > disabled=0 and (name='%s' OR name like '%s') and domain_id='%d' > > with > > %.<record-to-update> > > as second argument. > > When like is used with a search argument that starts with a wildcard the > index can not be used. > > Because I know that my records to update have no subzones I short circuit > query with > > SELECT content,ttl,prio,type,domain_id,disabled,name,auth FROM records WHERE > disabled=0 and (name='%s' OR 'x'='%s') and domain_id='%d' > > Cheers > > Thomas > _______________________________________________ > Pdns-users mailing list > [email protected] > http://mailman.powerdns.com/mailman/listinfo/pdns-users _______________________________________________ Pdns-users mailing list [email protected] http://mailman.powerdns.com/mailman/listinfo/pdns-users
