• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Drives for a "RAID10" VM Datastore??

farscapesg1

2[H]4U
2FA
Joined
Aug 4, 2004
Messages
2,658
Currently I'm using 4 older Toshiba 160GB 7200 RPM laptop drives in a striped+mirror ZFS array and I'm getting results that I'm not exactly happy with with, plus I already find myself constrained on space at only about 300GB...

I have the option of getting enough 300 GB SAS 10K drives (and already planning on picking up a M1015 or BR10i card anyways) for free, but used. Or I can just bite the bullet and pick up 4 new SATA drives. Unfortunately, SSD is out of my budget, as are new SAS drives ;)

I'm sure I would get better IOPS out of the SAS drives... but would it be that much more than say WD Black drives or the Seagate Momentus XT drives? This is for a home VMWare setup, so no database-intensive VMs. Just running a WHS2011 VM for system backups/file server, a couple XP/Win7 VMs, etc.

The SAS drives I can get are Seagate Cheetah T10 ST3300555SS drives. Unfortunately, I can't seem to find the idle power value for these drives to compare them to running the WD Black drives....
 
I wasn't sure what info you needed, but I'll take a shot of being more detailed ;)

I'm currently running a Supermicro X8ST3-F with a Xeon 5640 and 24 GB of RAM. This system is running ESXi 4.1 as an all-in-one solution running OpenIndiana for ZFS.

The OI server has two pools, both currently connected to the 1068e SAS ports on the motherboard.
Code:
NAME SIZE Bonnie Date(y.m.d) File Seq-Wr-Chr %CPU Seq-Write %CPU Seq-Rewr %CPU Seq-Rd-Chr %CPU Seq-Read %CPU Rnd Seeks %CPU Files Seq-Create Rnd-Create 

RAIDZ_datastore 7.25T start 2011.08.18 12G 76 MB/s 97 130 MB/s 14 85 MB/s 12 85 MB/s 99 285 MB/s 14 659.0/s 1 16 +++++/s 22177/s  

VM_Store 298G start 2011.08.18 12G 75 MB/s 98 114 MB/s 11 68 MB/s 9 76 MB/s 88 157 MB/s 7 1111.5/s 1 16 +++++/s 30959/s

The first (VM_Store) is a striped/mirrored pool comprised of 4 Toshiba MK1656GSY 160GB drives. These are laptop drives running at 7.2K RPM and the datastore is provided by NFS.
The second (RAIDZ_Datastore) is configured from 4 WD20EADS drives and will be used for media storage and the datastore is provided by NFS.

Running CrystalDiskMark on the XP VM when stored on the VM_Store pool gives the following results... which seem really low to me;

Sequential Read 39.08 Write 59.09
512K Read 38.10 Write 58.18
4K Read 5.927 Write 3.784
4K QD32 Read 4.510 Write 2.846

The VM's also seem really choppy, compared to when I run them on the RAIDZ pool...

For now it will remain as an all-in-one server, but I'm considering building a seperate storage server to free up memory and CPU usage for other VM's. I have an IBM M1015 SAS card on the way to seperate the datastores onto different controllers.

At the moment, the ESXi server is currently running the OI VM, vCenter, 1 XP VM, 1 Win7 VM, Untangle, and a WHS2011 VM. When I get enough funds I plan on building a second ESXi host to replace the Dell 745 I'm using temporarily in my home lab.
 
Back
Top