March 5, 2009 at 2:01 pm
I have five SQL 2000 boxes that are we are going to consolidate into one:
1 @ dual cpu, dual core Xeon @ 3.2 ghz 4 gig RAM avg cpu 24%
(this one is very heavy on cursor creation/destruction)
1 @ dual cpu, dual core Xeon @ 2.7 ghz 4 gig RAM avg cpu 7% spiking to 24% at times
1 @ dual cpu Pentium III 750 mhz 2 gig ram avg cpu 5%
1 @ dual cpu P III 705 mhz 2 gig ram avg cpu 3%
disk utilization is very low on all. Memory is all used up on all of them. We're not experiencing any sorts of slow exec times. On one of the boxes, about 150 queries run longer than one second, which is actually less than 1.6 seconds. There's a few queries that run for a couple minutes (overnight jobs), but from our testing, they will be sub 45 seconds on the new box.
We're upgrading and consolidating onto one SQL 2005 64 bit, AMD Opties model 2356 2.3 ghz quad core (8 total), 16 gig RAM. 900 gigs of RAID 5 disk.
I think the new box is strong enough. Any other thoughts?
March 5, 2009 at 2:08 pm
exactly how many disks do you have available?
what version\edition of windows do you plan to use?
-----------------------------------------------------------------------------------------------------------
"Ya can't make an omelette without breaking just a few eggs" 😉
March 5, 2009 at 2:12 pm
To me it looks like in terms of memory and processor you should be just fine.
I would be more concerned with the disks that you are running on and the amount of IO that you had before to now.
You say you are going to 900 GB of Raid 5 disks. Depending on the size of the disks you may be only on 5-10 drives. That is not necessarily a problem unless previously you were on 5-10 drives per server (meaning you decreased your IO capacity by up to 75%). That is what seems to be the one area you may have potential issues. But that is completely dependent upon your IO patterns.
March 5, 2009 at 2:18 pm
Windows Server 2003 Enterprise x64.
Number of disks is unknown. I would guess somewhere around five or more.
March 5, 2009 at 3:37 pm
Another thing to note Raid 5 can be poor with on write intensive systems. CPU, Memory, SQL edition I think your good. I think you need to play our your disk subsystem more; or you might have issues.
Another note to think about bulking so much space together makes it more difficult to scan, chkdsk, etc.
Thanks.
Mohit.
Mohit K. Gupta, MCITP: Database Administrator (2005), My Blog, Twitter: @SQLCAN[/url].
Microsoft FTE - SQL Server PFE
* Some time its the search that counts, not the finding...
* I didn't think so, but if I was wrong, I was wrong. I'd rather do something, and make a mistake than be frightened and be doing nothing. :smooooth:[/font]
March 5, 2009 at 3:59 pm
as already mentioned, you really do need to take a good look at the disk subsystem
minimum you could get away with would be
RAID1 for OS (2 disks)
RAID1 for logs (2 disks)
RAID5 for sql data (3 disks min)
-----------------------------------------------------------------------------------------------------------
"Ya can't make an omelette without breaking just a few eggs" 😉
March 6, 2009 at 7:12 am
I was told it has 8 drives.
Disk time on all servers is pretty low. Only one of them has any sort of activity which around 10% average.
Disk queue length is 0 on all.
Lazy writes are 0 on all.
Latch wait on most are 0, one of them is 12-24% average.
March 6, 2009 at 7:45 am
where i work our VP doesn't trust AMD due to their financial problems
I would buy a Proliant DL380 G5 dual quad core xeon and 32GB of RAM. if you use the 4GB chips then it's only around $1400 for the RAM and cheaper than getting 16GB with 8GB chips. in 18-24 months when the price falls you can get 64GB for the same price.
the server holds 8 disks internally and you can buy a jbod. i'd get an MSA 70 and use all 146GB disks and make sure you break everything up. separate all the db files, indexes on separate disks, logs, tempdb, etc
March 8, 2009 at 4:16 am
Sailor (3/6/2009)
I was told it has 8 drives.Disk time on all servers is pretty low. Only one of them has any sort of activity which around 10% average.
Disk queue length is 0 on all.
Lazy writes are 0 on all.
Latch wait on most are 0, one of them is 12-24% average.
i'm assuming these disk are 146GB each in size. If you dont go for extra storage and rely purely on the 8 internal disks then go for something like
1 x 146 (OS)
2 x 146 RAID1 (LOGS)
4 x 146 RAID10 (Data)
1 spare
or
2 x 146 RAID1 (OS)
2 x 146 RAID1 (LOGS)
3 x 146 RAID5 (Data)
1 spare
or
4 x 146 RAID10 (OS and Logs)
4 x 146 RAID10 (Data)
These are absolute minimums and not ideal but with just 8 disks!!!
BTW, where do you intend backing up the databases too?
-----------------------------------------------------------------------------------------------------------
"Ya can't make an omelette without breaking just a few eggs" 😉
Viewing 9 posts - 1 through 8 (of 8 total)
You must be logged in to reply to this topic. Login to reply