Just wanted to post an update to my current Drobo + WHS status.
[list=1]
[]I am now running WHS 2011 (64-bit) on the same hardware I used to run WHS v1 on.
It runs much better than WHS v1 did, and does not suffer the throughput interruption issue that Drive Extender’s migration would cause.
Media streaming is smooth.
[]I now have 4 Drobos hooked up to my WHS 2011 machine, for a total of just under 22 TB of storage
[]I can see all my Drobo’s status info in the WHS 2011 Dashboard thanks to Ross Dargan’s nifty WHS Drobo Status add-in.
It sure beats having to RDP and run Drobo Dashboard for basic usage and health monitoring.
I’ve been working with him to iron out a few bugs and it’s looking quite good. If you have Drobo on WHS 2011, I encourage you to give it a try and give him some feedback.
[]I am using Drive Bender for the storage pooling and duplication functionality that DE used to provide.
I chose Drive Bender because it works for me and still leaves a standard NTFS filesystem on each member drive, so if things really go south I can still mount the drive on another machine and retrieve whatever was stored on it (no weeding through tombstones like in WHS v1).
[]I have two storage pools (and drive letters) set up in Drive Bender:[list]
[]WHS 2011 default pool
1x Drobo S Gen 1
1x Drobo S Gen 2
Dual-Disk redundancy provided by Drobo S units
Duplication of “important” folders (User folders, Purchases, etc) is handled by Drive Bender (so I have protection of the “important” files even if an entire Drobo fails)
Fault tolerance:[list]
[]Loss of single drive on either/both unit(s) - no data loss
[]Loss of two drives on either/both unit(s) - no data loss
[]Loss of three+ drives on a single unit - partial data loss (only duplicated folders preserved, rest would need to be restored from backup)
[]Loss of single Drobo chassis - partial data loss (only duplicated folders preserved, rest would need to be restored from backup)
[]Loss of three+ drives on a both units - complete data loss (all would need to be restored from backup)
[]Loss of both Drobo chassis - complete data loss (all would need to be restored from backup)
[/list]
[]Movie Storage
2x Drobo v2
Single disk redundancy provided by Drobo v2 units
No duplication via Drive Bender
Fault tolerance:[list]
[]Loss of single drive on either/both unit(s) - no data loss
[]Loss of two+ drives on single unit - partial data loss (whatever was on the failed unit is lost)
[]Loss of single Drobo chassis - partial data loss (whatever was on the failed unit is lost)
[]Loss of two+ drives on both units - complete data loss (all would need to be restored from backup)
[]Loss of both Drobo chassis - complete data loss (all would need to be restored from backup)
[/list]
[/list]
[]Storage backup is being handled manually, as WHS 2011’s native backup functionality still has a 2 TB limit (not sure why… annoying it is.)
[]Client PC backup is still being done automatically by WHS 2011, much the same as it was in WHS v1.
[/list]
Hey mate
- i looked into this - but i wanted to have it more like flexraid’s pooling wher eit woudl try and keep diretories together - rather thn trying to balance load across the drives - i thikn it would make recovery muh easier if i lost a whoole directory - rather than losting 25% of the files from every directory
i bet you are exited about windows 8 storage pools and possible ReFS 
Hehehe Indeed. I do like the idea of keeping folders together, honestly not sure if DB does it fully, partially or none-ly, haven’t really looked at it too closely.
Yeah… Windows 8 Storage Pools. Hopefully they do it right, at least it’ll save them a bit of face after the DE fiasco.
its none-ly
if you copy a foldder with 300 identicaly sized files, only a 3 disk DB pool - you’ll get 100 files on each disk. which i think is a major oversight and turns recovery intoa bit of a nightmare
i use my array for media storage, i’d rather loose one or two tv series so they can be replaced/re-ripped
losing a third of every series woudl just be ridiculous.
Good to know. My discs come in via autoloader, so I’d probably just drop them all in and let whatever’s still there get auto-skipped.
hehehe
have you tried working out the cost of your setup…
i just worked out how much my sixteen 3tb drives are actually worth now…
i had to have a little sit down!
No, because I have zero desire to depress myself!
I’m perfectly happy in my ignorant view of being a “technology halfway house” - they come in, stay a while, then go away, and somehow I’ve done some kind of good service for the world.
hahaa - know that feeling - its how i got over 850 feedback on ebay!
btw i was looking at this page after searching for drivebender (from a recent post about hardlinks etc), and noticed that a news update was released for DriveBender v2 coming soon:
http://blog.division-m.com/2013/09/07/drive-bender-v2/
might be handy for you
(also had a couple of quick questions about it too)…
-
did you (bhiga) choose drivebender mainly to use hardlinks and dedupe features (as apparently they only work on single filesystems)?
-
does your drivebender for point 4 above, essentially just create an ntfs container file, eg like a trucypt container, which just happens to reside on the drobos as a huge file, which then contains your actual drivebender filesystem?
(if so, how could you mount it using ntfs if one of the 2 ‘bended pairs’ of drobo happened to lose multiple drives beyond their redundancy mode,
eg xDR +1 extra failed drive)
this thread is 18 months old?
afaik he still uses drivebender with whs2011 
and the v2 features might be handy for both of you
-
Nope, the intent was simply to (relatively) easily pool multiple physical storage devices into a single logical volume with duplication across physical devices. Hardlink support is still beta in the current version of Drive Bender and I’m not using it.
-
No. There’s no container file. Drive Bender sits “on top” of the standard NTFS filesystem. I can take any of my Drive Bender member drives and mount them on another machine and access whatever files that are stored there normally. There’s a root folder structure above the files that’s a GUID.
Oversimplified, Drive Bender looks at all the member drives, combines all the contents of folders with the same GUIDs and presents that combined content as a logical volume.
There are no hardlinks or “tombstones” to confuse things like WHSv1’s Drive Extender. Either the file is there, or it isn’t. In the case of duplicated files, duplicate copies are tucked away in a GUID subfolder, but they’re there.
Drive Bender will still mount a pool that is missing members in read-only mode. So that’s recovery step #1.
If for some reason Drive Bender won’t mount the pool*, recovery step #2 is to just take the member drives and recombine their data together manually, or just grab what I need.
- I have had Drive Bender fail to mount a pool after a very early upgrade SNAFU. Their support sent me a utility to run, I sent the log dump back, and they sent me a repaired configuration that got me fixed. Total turnaround was 3 days from start of problem to resolution.
thanks for the information bhiga,
sorry if i missed something, but when you say that it sits on top of the standard ntfs filesystem, and still leaves a standard NTFS filesystem on each member drive so you can mount it…
… does this mean that it is drivebender itself, which actually does (or did) the initial formatting?
eg, because as far as i understood it, if a drobo formats a drive/pack, then i didnt think you could easily pick out a drive and try to mount it individually?
For drivebender I dont think it matters who / what does the formatting on the individual disk - as long as it is NTFS compliant, then you can always remove the disk and use it in another system?
on drobo - the volume is NTFS formatted, but the disks are most definitely not
Docchris is correct.
It doesn’t matter who formatted the member volume. I don’t think Drive Bender lets you format either member volumes or its virtual volume. The Drive Bender volume already appears formatted.
Drive Bender allows you to add an existing volume and its data to the pool.
You can also have it mount drive letters for each of the volumes individually for testing/debugging but I do not recommend doing this for Drobo, as it makes the capacity gauge go nuts. I had quite the scare when the capacity gauge (and Drobo Dashboard) said I only had 84 MB in use, even though I could still see and access all my data fine.
So more clearly on the “taking the pool apart” front, you can freely remove any storage device from the Drive Bender pool and mount that storage device as you would normally on a non-DB machine.
many thanks dochris and bhiga for the info,