WHS2011 folders & disk placement - noob questions!

[HPLovecraft]

Hi Everyone,

I've done plenty of reading but am still a little unsure - hope you don't mind a couple of noob questions in a longwinded post...

 

BACKGROUND: I have an HP Microserver N40L as a home server, which until recently had 2 x 2TB disks set up as RAID1 (controlled by the HP's onboard Raid gadget). I was running low on space, so I posted a query over on WeGotServed forum asking how to expand my RAID array. I had tried out RAID1 to guarantee continuity in the event of a drive failure. I also back up of course.

 

A couple of forum members over there were very helpful, and both recommended DrivePool as an alternative. I'd never heard of it, but did some research and realised it could offer me the benefits of RAID1, plus more flexibility and control. A couple of months later, Santa Claus has brought me some new HDDs, and I'm ready to go. I've already dismantled my RAID1 mirror, and have installed my new HDDs in the Microserver.

 

In there I have 4 disks devoted to storage:

• Seagate 1 - 2TB
• Seagate 2 - 2TB
• Hitachi 1 - 2TB
• Hitachi 2 - 2TB

(in fact I also have a couple of extra partitions from the Hitachi drives, but let's ignore those for now...). The system is on a separate 5th HDD, tucked away in the optical drive bay. Next step is to download & configure DrivePool.

 

QUESTION 1: By default, WHS2011 has created a range of server folders (My Documents, My Pictures, My Videos, etc). I planned to keep this structure - the family understand them & where they can find their stuff. Once they are sitting in a big pool, can I still simply replicate what I used to do with RAID... mount each main folder as a network drive on the client computers?

 

QUESTION 2: I gather I can force certain folders onto certain HDDs using the file placement rules. I plan to have My Docs and My Pics on one 2TB drive, with My Music & My Videos on a different 2TB drive. I plan to use DrivePool to duplicate everything - so I will have a max of 4TB of storage from my 8TB of disks, which is plenty for me. But can I also force which HDD the folders are duplicated onto?

 

Here's why I'm asking... the 2 Seagate drives are nothing special, they're not really intended for server use... they're just what I could afford at the time I first built the Microserver. The two Hitachi drives from Santa are superior units - intended for 24/7 server use, with a far better track record than the Seagates.

 

When setting up duplication, I'd like to 'pair' each Hitachi with a Seagate - so for example My Docs & My Pics are on a Hitachi, duplicated on a Seagate. Each pair would, by definition, be balanced within itself)

 

I don't like the idea of my files living on a Seagate, and being duplicated on another Seagate... I thought it might be wiser to mix them up.

 

Is this possible in DrivePool, or am I trying to dictate something that it's not deigned to do? Am I mad? Or paranoid? Or both..? Told you they were noob questions, but any guidance would be much appreciated!

[HPLovecraft]

...ahem, forgive me for replying to my own post...

 

I've been experimenting with the DrivePool trial software. It obviously requires an entirely different mindset to what I'm used to. It's clever stuff.

 

I haven't found a way of dictating which drive a duplicate of a file is placed on. But it just occurred to me, I could force DrivePool's behaviour by creating more than one pool..?

 

If I want data on a Seagate to always be duplicated on a Hitachi, I could pool only those two drives together? The software will have no choice where to put the duplicates.

 

Then I make a second pool with Seagate#2 and Hitachi#2.

 

I don't mind having more than one pool... my objective is drive redundancy, rather than a single massive pool. Does this sound sensible..? My brain hurts!

[Christopher (Drashna)]

QUESTION 1: By default, WHS2011 has created a range of server folders (My Documents, My Pictures, My Videos, etc). I planned to keep this structure - the family understand them & where they can find their stuff. Once they are sitting in a big pool, can I still simply replicate what I used to do with RAID... mount each main folder as a network drive on the client computers?

You can absolutely do that. Just "seed" the pool with the existing contents and run the "WSS Troubleshooter" and use the "Rebuild DrivePool Shares" option. It should then function exactly like it was before. Otherwise, you'll need to juggle data around.

http://wiki.covecube.com/StableBit_DrivePool_Q4142489

http://wiki.covecube.com/StableBit_DrivePool_Utilities

 

 

And then you can absolutely map the shares.

 

QUESTION 2: I gather I can force certain folders onto certain HDDs using the file placement rules. I plan to have My Docs and My Pics on one 2TB drive, with My Music & My Videos on a different 2TB drive. I plan to use DrivePool to duplicate everything - so I will have a max of 4TB of storage from my 8TB of disks, which is plenty for me. But can I also force which HDD the folders are duplicated onto?

 

Here's why I'm asking... the 2 Seagate drives are nothing special, they're not really intended for server use... they're just what I could afford at the time I first built the Microserver. The two Hitachi drives from Santa are superior units - intended for 24/7 server use, with a far better track record than the Seagates.

 

When setting up duplication, I'd like to 'pair' each Hitachi with a Seagate - so for example My Docs & My Pics are on a Hitachi, duplicated on a Seagate. Each pair would, by definition, be balanced within itself)

 

I don't like the idea of my files living on a Seagate, and being duplicated on another Seagate... I thought it might be wiser to mix them up.

 

Is this possible in DrivePool, or am I trying to dictate something that it's not deigned to do? Am I mad? Or paranoid? Or both..? Told you they were noob questions, but any guidance would be much appreciated!

 

Yes, you can use the "File Placement Rules" to limit what files or folders end up on which disk. However, you need to use version 2.X of DrivePool to have access to this feature.

http://stablebit.com/Support/DrivePool/2.X/Manual?Section=File%20Placement

 

As for forcing which folders are duplicated where, the file placement rules would be the only way to do that.

Also, StableBit DrivePool has no concept of "master" and "duplicate" file. Both files are the "master" file and may be used (or both used in conjunction).

 

As for the drives, you should be fine. However, if they're the "1TB/platter" drives, you may want to consider replacing them anyways. I've had 5 ST3000DM001 start to go bad on me in the last six months). 

 

However, if you install StableBit Scanner, you should be much better off. DrivePool will pull information from Scanner, and can cause DrivePool to evacuate a problem drive. (you may want to enable the SMART options for the StableBit Scanner balancer as well).

 

For the most part, there shouldn't be any reason to worry. And if you have StableBit Scanner installed, it should take care of everything for you.

 

...ahem, forgive me for replying to my own post...

 

I've been experimenting with the DrivePool trial software. It obviously requires an entirely different mindset to what I'm used to. It's clever stuff.

 

I haven't found a way of dictating which drive a duplicate of a file is placed on. But it just occurred to me, I could force DrivePool's behaviour by creating more than one pool..?

 

If I want data on a Seagate to always be duplicated on a Hitachi, I could pool only those two drives together? The software will have no choice where to put the duplicates.

 

Then I make a second pool with Seagate#2 and Hitachi#2.

 

I don't mind having more than one pool... my objective is drive redundancy, rather than a single massive pool. Does this sound sensible..? My brain hurts!

How dare you!!!!

Just teasing.  

There is no problem with doing so.

 

 

As for the file placement, again, make sure you have 2.X installed. Then open the balancing settings. There should be a tab.

 

No, no there isn't a way to ensure where "duplicates" end up. As I mentioned above, ... it should clarify why. However, you could limit the contents of a folder to the two drives (one seagate and one hitachi). 

This is a better option than two pools. And much more simple.

[HPLovecraft]

Many thanks for replying Christopher, that make things much clearer. It sounds like 'seeding' the pool is much easier / faster than actually copying all the files across - that's a bonus. Can I ask one more stupid question?

 

I keep all my files in WHS2011's default shared server folders. When I follow the procedure you linked to move existing files into the pool at http://wiki.covecube.com/StableBit_DrivePool_Q4142489 - is it the WHS2011 shared folders I should be moving? I assume I should do this within the WHS Dashboard, instead of Windows Explorer, so WHS knows where they have gone to..?

 

Actually, I can't guarantee that this is the last stupid question that I'll ask! 

[Christopher (Drashna)]

You are very welcome!

 

And yeah, "seeding" the pool is essentially just moving the contents to a different location on the same disks. This means it takes minutes (if that) instead of hours (roughly 4 hours per 1TB to move under ideal conditions, IIRC). You can seed a large pool in less than half an hour (i've done so several times).

 

 

As for moving the folders, you'd do so on the file system and ignore the dashboard.

After doing so, run the "WSS Troubleshooter" and select the "Restore DrivePool Shares" option.

This will "correct" the shared folders and use the new location on the pool.

http://wiki.covecube.com/StableBit_DrivePool_Utilities

[HPLovecraft]

Fantastic - the advice is much appreciated. Taken together with some other info from around this forum, I reckon I'm ready to plunge into the 'Pool... Will report back when it's hopefully all set up & running!

[Christopher (Drashna)]

You are very welcome!

And I'm glad to hear it!

 

If you have any other questions, don't hesitate to ask.

[HPLovecraft]

Yes, you can use the "File Placement Rules" to limit what files or folders end up on which disk. However, you need to use version 2.X of DrivePool to have access to this feature.

http://stablebit.com/Support/DrivePool/2.X/Manual?Section=File%20Placement

 

As for forcing which folders are duplicated where, the file placement rules would be the only way to do that.

Also, StableBit DrivePool has no concept of "master" and "duplicate" file. Both files are the "master" file and may be used (or both used in conjunction).

 

Hi Christopher - sorry if I'm making hard work of this, but I can't quite work out how to do the above trick... I've created a pool, but haven't seeded it with my existing content yet. I've just placed a test folder that contains a text file to check I understand the placement & duplication functions correctly.

 

Right now I have 4 drives in the pool. I'd like to force the test folder to always be on Drive 1, and always be duplicated on drive 3. Under 'file placement rules' I've highlighted the folder and ticked the checkbox next to 'Drive 1'. I also checked 'Never allow files to be placed on any other disks'. I don't see anywhere to check that dictates where the folder will be duplicated when duplication is turned on.

 

Then I enabled 'folder duplication' for the test folder. Once again, I couldn't see anything to specify where the duplicate should be stored.

 

By 'showing hidden files & folders' and looking at the drives, I can see that my test folder is currently being stored in duplicate on Drives 2 & 4. It's not present at all on the desired Drive 1. This is pretty much the opposite of what I'm trying to achieve - I suspect I'm getting something fundamentally wrong... Please can you suggest what it is?

 

 

[The reason I'm being so particular about what is stored on which drive, and on which drive it's duplicated, is to help with my backup regime. If I can restrict certain folders to certain drives, I can be confident I've backed all my files up by backing up only two of my four underlying drives each time. This will save me time & space. WHS2011's backup limitation is also a factor]

 

Many thanks for any advice!

[Christopher (Drashna)]

To clarify, the UI at the bottom is what the CURRENT file placement is, not what the projected layout.

 

Also, what are the balancing settings, on the "Settings" tab. This will influence how quickly the files will be migrated off of the disk.

And what balancers do you have enabled (on the balancing tab)?

[HPLovecraft]

Hi Christopher, I've been blundering around in there and have better success tonight. I also found this blog post, which doesn't quite look like my version, but which put me on the rails: http://blog.covecube.com/2014/04/stablebit-drivepool-2-1-0-503-beta-per-folder-balancing/

 

Here are screengrabs from my (more successful) experiments tonight...

 

My test folder, now permitted to appear on TWO drives (duh! if I wanted it duplicated I should have allowed it to live on two drives in the first place...). File duplication is turned ON.

 

 

My balancing settings. Currently set to non-automatic, but I planned to switch this to automatic once my files are all in place, and the folders have their rules all applied as I want them:

 

The balancing plug-ins that I've allowed...

 

...and their settings:

 

And finally (and I think this was maybe the eureka moment), I noticed this little pull-up arrow that allowed my to force an instant re-balance (I'd totally not noticed it before):

 

Once I instigated the rebalance, my test folder jumped to the correct drives. Interestingly the test folder remained on the other two drives, but it was emptied - the test file inside it had gone.

 

Does this combo of settings look appropriate for what I'm trying to achieve? Basically I'm trying to place my top-level folders on two drives, and consistently duplicate them on a different two drives - a sort of RAID1 result. But I want to prevent Drivepool from making executive decisions and splitting the contents of my top level folders all over the place. The only exception would be allowing Scanner to evacuate a drive that's about to fail.

 

I know this sounds really anal, and maybe I'm not allowing Drivepool as much free will as it's designed to have. But I need to know I can pull a HDD out of my server (let's imagine it's the one with 'My Pictures' on it), put it in a caddy and find that all My Pictures really are on it... none have been scattered onto a different drive. I'm conscious I may not be articulating this very well!

 

By the way, I appreciate the care & time you devote to this forum, answering questions, and helping out the likes of me!

[HPLovecraft]

Just a quick update on my post above... this weekend I committed the great data move into my 4-disk 8TB pool, using the settings shown in the screenshots above. All went great.

 

Christopher's file seeding method, and folder-resharing fix with the WSSTroubleshooter that he gave above (post #3) worked like a charm.

 

Thanks to the placement rules, I have been able to force my key folders onto two drives each (one where I put them, one where I have allowed them to be duplicated), whilst keeping their contents all together in the folders where they have always lived (rather than balanced around the 4 drives in a way that wouldn't have suited me personally).

 

The longest part of the process was the duplication, as you'd expect. After duplication, the dashboard told me that balancing was not optimal and invited me to run a balance. I had some nervousness about this (my placement rules are so strict that by definition the pairs of drives are always balanced), but I allowed DrivePool to go ahead... the progress bar chugged along, and at the end no files had been moved (which is what I wanted), and DrivePool seemed satisfied that all was well.

 

Effectively I still have a RAID1 style setup, but engineered through DrivePool, which gives me better control, visibility, and options when it comes to swapping out drives. I've paired different brands of drives so that a 'strong' one is duped to a poentially 'weak' one. A much friendlier version of the old RAID1 that I used to have through WHS2011.

 

There are some other benefits too, that I've yet to take advantage of: I'll be installing Scanner to keep an eye on my hardware; and I also have the option to fine-tune my placement rules to duplicate my REALLY crucial files (such as my KeePass database) on all 4 drives.

 

For any noobs like me who are similarly nervous about committing I'd recommend: looking in the blog (where there's loads of handy info expanding on what's in the manual); playing with some test folders & files first to understand what DrivePool is doing, and why, when set up a certain way. As I mentioned in a previous post, I needed to change my mindset a bit to 'get' DrivePool. It's a clever, clever piece of kit.

 

Next task - overhauling my backup regime!