Well, we tested extensively on a staging server, and everything seemed to
work well. The production server was something else... truly a messy
process.
Some of the larger tables started rebuilding (as expected), but stopped
after about a million records. System was REALLY slow by then, comaning to a
halt. Not your average 50.000 record master merge.. The population couldn't
be stopped, couldn't remove change tracking. Terminal Server couldn't
connect, EM manager was unable to make contact.
Decided on a reboot, which resulted that this large table to go into
recovery mode. Other then massive CPU and disk IO, no visible result.
Fortunately I found an excellent piece of advice from Hilary on setting the
registry to avoid the rebuild, I dumped the gathered files, and was finally
able to remove the FTI from the table.
After this mess I decided to rebuild the other two indexes, but the net
result is doubling in time it takes to process all the queries. Load on the
system doesn't matter, even at 3 AM when there is nother activity it just
takes and takes. As an aside, resource_usage is set to 4
Nothing else changed, although in the meanwhile we have rebuild the major
indexes, and reorganized the tempdb which gets used a lot. Nog discernable
differences in FTI, unfortunatly. Since the big 6M+ index hasn't been
rebuild, I'd even say we doing less then before. Nothing else changed - so I
am at a loss. I realize we're using FTS quit extensive, every half hour we
execute 3000 queries. Since that takes 20 minutes, it is getting out of hand
Ant thoughts?
René
[quoted text, click to view] "John Kane" <jt-kane@comcast.net> wrote in message
news:eI0E$4GUDHA.2196@TK2MSFTNGP11.phx.gbl...
> René,
> Hopefully, this is not what is to be expected after upgrading to a new
> service pack, SP3 or any other one!
> Have you eliminated any non-FTS SQL Server related issues, such as
updating
> the table statistics or rebuilding the primary index? Are the query plans
> the same as they were with SP2 + slammer protection? If not, then what is
> the difference?
>
> Regards,
> John
>
>
>
> "René" <rene.de.vries/atsign/kexdotnl> wrote in message
> news:OeVyANGUDHA.556@TK2MSFTNGP12.phx.gbl...
> > Gents,
> >
> > We just made transition to SP3 on our production server. After a lot of
> > hassle rebuilding the FT indexes, we now running a few days with this
> > set-up.
> >
> > However, what we notice is that the FT query performance seems to have
> > slowed down. A process that took 10-12 minutes before, now takes about
20
> > minutes. Is this a what is to be expected? Nothing else changed but the
> > upgrade from SP2 + slammer protection to SP3.
> >
> > Something we missed?
> >
> > René
> >
> >
>
>