Re: Horrible Bad Performance on Gargantuan Hardware

I reposted this in the SQL Server forum and there has been a response

http://www.microsoft.com/technet/community/newsgroups/dgbrowser/en-us/default.mspx?dg=microsoft.public.sqlserver.server&mid=28e98552-a0a8-46d5-b3b3-fa496c17f0fc&sloc=en-us

"Paul Pedersen" wrote:

I don't know enough to laugh, but I was watching it to see what the
responses would be.


"SSGTech Tim" <SSGTechTim@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:D6C8A638-3CA5-48F3-A037-315F21C39DD9@xxxxxxxxxxxxxxxx
No balls in the group huh, not one of you even thought to laugh at this
one.

"SSGTech Tim" wrote:

Hi,

I was asked to look at server running Windows 2003 Standard (Version
5.2(Build 3790.srv03_sp1_rtm.050324-1447: Service Pack 1), Quad Xeon
processors, running 4Gig of RAM with a 33Gig OS drive and a 1.5 Terrabyte
Data array partitioned into a 1 terrabyte partition and another .5
terraybte
partition.

The question I was asked was, ?Why do simple queries run horribly and get
exponentially worse as they become more complex and why does a system
running
about a fifth the capability run better?

The server is an MSSQL server and currently the ERP software provider has
been rewriting code based on customer accusations in relation to system
performance.

I was asked to take a look because the customer is willing to do anything
at
this point to get their problem resolved.

A simple truncation of three table and repopulation of 9,000 records into
each table is still running at an accumulated time now of 6 hours, 41
minutes.

Particulars:

After six hours running this query, the SQL server is now at 1.7 Gig of
RAM
consumed.
Page Faults are at 473,749 and rising.
Paged pool at 105k
Processors which started with only one spiked with a reported CPU usage
of
13% are now up to 32% and they are all starting to make like they are in
use.
Avg. Disk Bytes/Transfer for the D: drive (1.2 terrabyte) average at
24000
bytes
Avg. Disk Bytes/Transfer for the E: drive (log file data) average at
24000
bytes
% Disk time for D: 248
% Disk time for E: 1.409
BTW; the only thing going on aside from an hourly disk dump is this
query.

I'm guessing that they controller/drive configuration are inappropriate
for
the type of data read/write metrics that this system will be
encountering. I
also think that there maybe quite a bit wrong with the hardware.

I've thrown it back to them asking for the management utility to
interrogate
the RAID array and to have the hardware manufacturer validate the RAM.

Anyone have any ideas?



.

Relevant Pages

  • Re: SQL Resets
    ... Server Support or from another SQL Server experts provider. ... Disk Read Queue Length ...
    (microsoft.public.sqlserver.connect)
  • RE: Very slow response from server
    ... I suggest you backup the disk at first. ... I know the problematic PC is a file server & printer server, ... Very slow response from server ... That happens is network printing becomes very slow, ...
    (microsoft.public.windows.server.general)
  • Re: SQL Server I/O performance issue
    ... Geoff, thanks for that. ... Microsoft SQL Server MVP ... Both % Disk Time and Avg Disk Queue Length are showing very poor ...
    (microsoft.public.sqlserver.clustering)
  • Re: sunmanagers Digest, Vol 22, Issue 12
    ... series server. ... B B B B B i want to do install one more solaris Over there. ... Solaris cannot see all drives on Areca RAID controller ... popped it (a single disk at this point) into machine1. ...
    (SunManagers)
  • hai...find me a solution in M5000 server
    ... series server. ... B B B B B i want to do install one more solaris Over there. ... Solaris cannot see all drives on Areca RAID controller ... popped it (a single disk at this point) into machine1. ...
    (SunManagers)