Corrupt files with Storage server and MSA2010 SAN

I am getting the following error on my Windows Storage Server 2008's Server Event log:

The file system structure on the disk is corrupt and unusable. Please run the chkdsk utility on the volume Data.


I am using Windows 2008 > iSCSI Initialor > HP MSA2010

There are no errors listed on the MSA. I have ran a chkdsk in the past and it fixed a million security permission and file system problems. It seems to be happening again. I update all drivers and firmware on everything end to end.

Could networking problem cause this? I am showing "receive errors" on the MSA controller. Does that mean it is too much traffic? I have the MS server and the MSA connected to our normal network and to a SAN-traffic only network with it's own Gigabit Cisco switch.


*** Resolved ***
Apparently you can't have multiple computers accessing the same vdisk or it causes corruption.

Hi,

I know this is completley usefless to you question but as its a SAN topic, I'd like to ask my own.

We have a Clarion SAN at present fibre linked, and are about to upgrade.

The options are an NX4 EMC or an HP left had SAN moving to ISCSI from fibre. To be brutaly honest my SAN knowledge is not great so I have both companies arguing their product is better.

Can anyone simply advise which of these products is better, bear in mind that down the line virtualization will be happening so shared storage is a must. There is not much difference on their quotes.

Sorry if you feel I've hijacked your post if so I'll post somewhere else.

Many thanks

Here is my two cents on which SAN. We just purchased two Lefhand SANs (Now HP SAN IP, I think) with iSCSI and we do virtualization to over 30 VMs on HP blades.

Either solution will work, but I like Lefthand and iSCSI for the flexibility.

What are your requirements? What type of machine will you virtualize? Exchange? DCs? SQL? etc?

Requiremnts at the moment is simply for file access but for the SAN the following is a must

high availability in the box through the dual storage processors and controllers

Shared storage for Virtualization, and ease of managemnet.

I've been checking and have seen that Each Left Hand node has several single points of failure within them as they are in essence servers, so to get the high availability required, two left hand nodes actually equal one highly available SAN assuming they are using RAID within the San and RAIN across the two nodes.

I also have concerns with HP around the I/O through-put and power, and if I need to expand the storage, then it’s not as simple as adding a tray of disks, it’s another two nodes minimum at circa £40K

Can you confrim if this is the case

Cheers for the feedback

I believe with Lefthand all you have to do is buy another tray of disks to expand it. Also, Lefthand is supposed to be very flexible when adding removing space. In the past, I had to migrate the data, delete the volume expand/shrink it, then rebuild it and migrate the data back. With Lefthand you are supposed to be able to the changes on the fly without moving data around.

If you can wait until Wednesday I can let you know for sure because I have two days of Lefthand training. This Monday and Tuesday.

I have a lefthand San as well and have been using it for about 7 months. Adding space is easy you just buy more boxes and add them to the management groups its easy to scale. What skepticals is saying is true it's very flex-able and the management is very user friendly and pretty feature rich for a small to medium business SAN.

I'm using raid 1 network raid san so I have 2 boxes with synchronous communication. If one box fails you don't even know. I can do maintenance on one reboot it 0 interruption. Be aware when you setup a volume and want a highly available volume 30gb consumes 30 gb on both ends.

They just added raid 5 and raid 6 network raid as well so you can get more space performance with less boxes now. It takes 4 boxes to setup raid 5 but one box can be brought down your still up with no downtime and you get the total space/iops of 3 out of the 4.

I then have another left hand at a remote site providing disaster recovery. So lefthand can do sync, async, cloning, failover/high avilablity, snapshots, remote snapshots, application/vss snapshots, expand/shrink volumes on the fly, and performance monitoring. Not bad for a SMB product.

As far as performance the thing that always kills you is disks. The more disks the more iops. Since you can just add more disks to a management group the performance and storage scales every time you add a box. For us it had plenty of performance with just 1 box if that changes ill scale it up for more drives.

Hp does have documentation on iops with their product so as long as you use perfmon and monitor your environment for at least a week you will know how much iops you need. Until then you don't really know how much you do or do not need.