Re: SQL Server 2000 Performance - 10K vs. 15K
- From: "MeanOldDBA" <MeanOldDBA@xxxxxxxxxxxxxxxxxxxxxxxxx>
- Date: Sun, 18 Sep 2005 15:11:01 -0700
Have you looked at your disk queue lengths in performance monitor? I would
be looking at both the logical and physical disk reads per disk to identify
where/if the bottleneck is occurring at the hardware level. In addition,
monitor the cache hit ratio, buffer cache hit ratio, target and total server
memory counters, pages/sec, and processor to make sure you have no issues
After you moved to the new hardware, did you run UPDATE STATISTICS on the
databases? If not, you might want to consider that.
Post the counter averages above, and include max(counter) values for the
physicals. That should help us quickly identify if it's hardware.
Other than that, people have given you really good advice on using Profiler.
You need to find out if there is a hot spot causing you issues. Sometimes,
moving to a bigger system can change dynamics enough to cause particular
processes to run poorly. After they are identified and tuned, you can really
begin enjoying the full benefit of the upgrade.
Mean Old DBA
When life gives you a lemon, fire the DBA.
"Hank Arnold" wrote:
> I'm definitely not in the same league as John and Andrew, but it sure makes
> sense to investigate where the bottlenecks are before investing that kind of
> money. You know how the finance folks are!! ;-)
> That's one powerful system you have there. It *should* really rock.....
> Hank Arnold
> "randy" <rhenson@xxxxxxxxxxxx> wrote in message
> > Hi SQL Gurus, I've got a scenario that I need some help with.
> > I recently upgraded our company's erp system from the old foxpro
> > version, to the latest sql version. I'm seeing a couple of performance
> > issues. It doesn't matter which workstation config. I've got some with
> > 2.8GB processor and 2GB of RAM, still have the same problem. I can run
> > the client on the server, and still have to same issue.
> > It's like the reads are OK, but the writes seem slower than on the
> > foxpro version. Not a lot slower, but still slower -- and you know
> > how users can be!!
> > My server config is just as the software company recommends. Dell
> > PowerEdge 2850, Dual 3GB processors, 6GB RAM with sql tuned to use 5gb.
> > . RAID controller is PERC4e/DI, with 2X4 backplane. SQL Server 2000
> > Enterprise with AWE enabled. Log files and data files on separate raid
> > arrays on different channels. OS and Log array is RAID 1 on channel 0,
> > Data array is RAID10 on channel 1. RAID Policy is set to write-back.
> > My database is about 10GB.
> > I said all that to ask this question, all of my drives are 140GB 10K,
> > would the performance increase be significant enough to spend the money
> > if I upgraded the 4 hard drives on the RAID-10 Data drive to 140GB and
> > 15K?? I've got the budget to do it, but if it's not a big difference,
> > I won't bother spending the money.
> > I appreciate any guidance you will be able give me on this.
> > Randy
- Prev by Date: Re: raid 1 vs raid 5
- Next by Date: Re: SQL Server 2000 Performance - 10K vs. 15K
- Previous by thread: Re: SQL Server 2000 Performance - 10K vs. 15K
- Next by thread: Re: SQL Server 2000 Performance - 10K vs. 15K