0

I have a Windows 2008 R2 AD server operating on my network, it recently ran out of space and I upgraded to 2TB SATA drives instead of the 300GB SAS drives it came with, I configured my 4 drives in RAID-10. I restored operation through the Windows Server Backup feature and everything seemed normal.

Now my workstations are taking a very long time to log on and once on they are taking a very long time to interact with client-side software that communicates data from the server. Sometimes things are taking minutes between communication requests.

Why are things running so very slowly? Thank you for your assistance.

user1695505
  • 101

1 Answers1

0

This could be for any number of reasons, without further information I will be just guessing, firstly I would ask what else is on the server besides AD - because it shouldn't need that much space and its not good practice to put a file server (home directories) or roaming profiles on your AD server, second how does the new partition table compare to the old table, third, is the raid array rebuilding since you restored from a different raid backup. Most importantly, are you monitoring the disk I/O, there are lots of free tools to monitor disk IOPS you should benchmark the array to see what speed you are getting, you could also reboot the server and log into the raid controller to view your virtual disks for any possible errors, check what the write policy is and the block size - all of these things also effect performance - if the server only came with 1 disk then the RAID controller wasn't doing much before.

Can you give the model of the server, the raid card, how the server is configured besides just AD controller and have a look at task manager and see what the disk and network are doing. but to some up - the disk is busy doing something probably RAID 10 was enough to cause the bottle neck, it's much slower than RAID 0

Sum1sAdmin
  • 1,934
  • 1
  • 12
  • 20
  • It was previously using 4 300GB SAS drives, the partition table is similar but with an expanded primary partition to 1TB size. I needed this size because this AD server runs server-side software that stores data locally on the AD server's primary partition. I failed to verify the previous configuration of the RAID array in the default configuration of this PowerEdge T630 - but based on 4x300GB drives and only having 450GB available in the partition table I imagined it was RAID-10 previously also. – user1695505 Apr 07 '16 at 21:24
  • What was the RPM of the old disk vs the new SATA- it will 7.5k ,10k or 15k – Sum1sAdmin Apr 07 '16 at 21:29
  • 4 300GB SAS disk in RAID 10 gives 600GB useable storage. – Sum1sAdmin Apr 07 '16 at 21:33
  • Previous RPM was 10k, current disk RPM is 7.5k. Perhaps I am misremembering the previous configuration - could there only have been 300GB available with these 4 disks? What configuration would that represent? I am attempting to receive redundancy and performance - I need fault tolerance for 1 disk but I wonder if this server was operating for performance and not fault tolerance previously? – user1695505 Apr 07 '16 at 21:37
  • Ok - you have massively downgraded your performance, SATA is for pure storage such as backups, SAS is enterprise grade performance and you have also reduced the RPM, my advice - power off the server put the 4 SAS drives back in, boot up and run the PERC setup - destroy the array, create a new virtual disk, add the 4 physical disks, choose RAID 5 - a disk can fail and it should mail you to let you know - Dell have a 4 hour turn around on disks - also you are backing up right? so a disk can fail - no big deal - get another disk before another one fails. – Sum1sAdmin Apr 07 '16 at 21:43
  • Thanks for the info Rob-d, I will implement these changes and touch base here again tomorrow. – user1695505 Apr 07 '16 at 22:00
  • 1
    I wouldn't be so quick to assume that the slower disks are the cause of your problems. Please check the windows resource monitor, specifically the disk I/O and queue length metrics during a period of time when the symptoms are present. If the disks are bottle-necking your system then you would expect to see a queue length > 0. – blacklight Apr 07 '16 at 22:19
  • Here is an image of my disk activity during a backup [link](http://pasteboard.co/2X6cF5S.png) So if I switch back to the SAS drives I think I can recover my much needed performance, but I still need storage. Since I only have 4 ports I'd like to do two RAID arrays, both redundant but with acceptable performance and use the 2TB drives for data. I was going to use RAID-1 for each of these arrays. How does that sound? – user1695505 Apr 09 '16 at 02:09