Dynamic Disks on Hardware RAID5

[gmail2]

We've currently got 4 disks in a RAID5 configuration which has 3 partitions (C:, D: and E. E: drive is 150 GB and is where user data is stored. However, it's no longer big enough for what we need to store there. We're about to create a second array with 2x300GB disks in a RAID1 configuration. This will give us an additional 300GB worth of data, but the array will appear to windows as a separate "disk". Ideally, what I'd like to do is combine the current 150GB with the new 300GB worth of storage, that way we don't have to decide which shares should go on which disks. But the only way I can see that we can do this is by using dynamic disks ... so, I have some questions !!

1. Are dynamic disks recommended for server situations? When I googled it, I mainly got pages relating to XP

2. Does it cause "confusion" for the hardware RAID controller?

3. If I setup a Spanned volume on the dynamic disks, how is the data written to the disks? Will it exhaust all the space on Disk0 and then move onto Disk1?

4. If one of the disks in one of my hardware RAIDs fails, and is replaced, will it have any impact on the spanned volume (which is escentially a software RAID)?

Would really love to hear from somebody who has tried this out. Or alternative, if somebody has another solution on how I can group the 150GB and 300GB worth of storage together, I'd really appreciate it.

Thanks in advance

Irish Poetry - Karen O'Connor
Get your Irish Poetry Published
Garten und Landschaftsbau

 

[technome]

1. Are dynamic disks recommended for server situations? No!!!!! your adding a point of failure if your OS has file corruption.

Windows Dynamic disks adds no confusion to the hardware raid, hardware raid knows nothing about Dynamic disks structure or files associated with Dynamic disks.

"3. If I setup a Spanned volume on the dynamic disks, how is the data written to the disks?
Yes it will it exhaust all the space on Disk0 and then move onto Disk1
If you span your adding even more points of raid failure, the original corruption possibility of Dynamic disks, then the danger from disk failures in the span, basically you lose the ability to lose even 1 disk; a span acts like raid 0 as to failure rates.

A non perfect remedy ....
Most raid adapters will allow you to add drives to an existing array from within the raid card setup, but the space added does not automatically become part of the existing array volume. Once the drives are added to the array, you can use Diskpart.exe (with limitations) or a third party partitioning software such as Acronis disk Manager, to extend the original into the space created by the addition drive. The problem in your situation, only 150 meg of each of the 300 Meg drives can be added to the original raid array; all drives in a raid 5 are treated as the same size as the smallest capacity member.

What ever you do, DO NOT SPAN.




........................................
Chernobyl disaster..a must see pictorial

http://www.kiddofspeed.com/default.htm

 

[gmail2]

Yea, since posting this I'd read up on this some more and pretty much decided that it may not be such a good idea. From what I can see, it's escentially a cheap solution for those who can't afford a hardware RAID setup.

I had originally suggested to my manager that we expand the current array, but he didn't want to to this because last time he tried it, it took nearly a week for the array to re-stripe. There would be more than 150GB of wastage if we did go for that option as our current disks are 72GB (the E: partition is 150GB) so we've had 228GB wasted !!

I guess there's no other remedy really, maybe I'll just have to re-think how we distribute space to users instead and use that 150GB for something else like WSUS

Thanks for the reply

Irish Poetry - Karen O'Connor
Get your Irish Poetry Published
Garten und Landschaftsbau

 

[kmcferrin]

Who says not to use Dynamic disks on a server? Most people connected to SANs (and especially running VMWare and the like) are using Dynamic disks with no issues.

Also, I'm not convinced that spanning across arrays does anything to add to the potential failure rate. Let's say that he did expand a dynamic disk from a RAID5 array to include a RAID1 array as well. So now Windows sees two "physical" disks (actually arrays) that are presented as a single dynamic disk. If one of the drives in either array dies, then Windows still sees the "physical" disk as present, and no data loss has occurred. If he somehow managed to lose two disks in the same array then he would certainly be hosed. But that's little different than what would happen with any RAID5 or RAID1 array that loses two disks.