21 replies to this topic

[mediadogg]

This tool is not strictly an add-in, but this seems the most appopraite place to discuss it.

The WHS Migration Assistant (WHS2WHS) is designed to help primarily with the planning of the data move from Windows Home Server to Windows Home Server 2011. There are two versions of WHS2WHS: the Basic version and the Pro version. The Pro version, in addition to the planning functions, has functions for partially automating the actual data movements and other functions for saving, restoring and printing multiple plans. NOTE: WHS2WHS Pro does not directly copy, erase nor move any data. It provides a planning template as well as a framework for invoking other tools, using the plan as a guide.

Method of Operation:

Think of the move operation as going from left to right. WHS2WHS presents a screen that shows the shared folders from the source WHS system on the left, and the Volumes (drives) associated with the WHS 2011 system on the right. This is due to the fact that WHS 2011 uses a different scheme for managing shared folders. Instead of locating folders on a large pool of data that Drive Extender maps onto multiple drives, WHS 2011 requires the administrator to allocate shared folders to logical disk volumes, which usually correspond directly to physical disk drives, but not necessarily. Technologies such as RAID can be used to create logical volumes with various reliability and performance attributes comprised of multiple physical drives. WHS 2011 works with the logical drives (or “volumes”).

Basic Version - Click Here (free)
Pro Version (on sale!) - Click Here

Here is a screen shot.



For more details, please take advantage of the following task-oriented articles:

Basic and Pro

• Getting Started: How To Setup a Plan
• Automated Planning Features
• Physical Drive Inventory

Pro Only

• Automated Copy Operations
• Copying Share -> Share
• Copying Local WHS V1 Disk -> Share
• Copying Share -> Local WHS 2011 Volume
• Copying Local WHS V1 Disk -> Local WHS 2011 Volume
• Copying Share -> Mapped Network Drive

I have also attached the User Guide: Version 1.04 Release 6/10/2011

Attached Files

•  WHS2WHSUserGuide.pdf   210.45KB   175 downloads

[mediadogg]

Just discoverd a small bug that was introduced late in the game, so I missed it. I will post updated code for both Basic and Pro later today.

[Terry (WGS)]

Thanks for the heads up - have added it to the add-in list (even if it is a tool!)

[mediadogg]

Windows Home Server Migration Assistant (Basic and Pro) V1.01 Available

• Fixed a bug in the "Folder Add" menu
• Documentation Improvements

Customers of the Pro version get free upgrades.

Today's Migration Tip: (I'm about 15% along on a 20TB trek).

I can't overemphasize the value of planning and keeping track. The opportunities to lose data abound in the confusion of migration. Case in point: I had a hard drive crash during the copy process. WHS 2011 wanted no parts of the drive after that. I finally mounted it in a USB dock and then ran a complete Chkdsk / Scan. The drive was all happy, but a couple of files, even after appearing to be OK in both RichCopy and Beyond Compare, were actually corrupt! VLC player would not play the files (Microsoft disk repair actually warns you of this possibility - that's why I checked). So, I recopied them, and now all files are verified. After that 2 day distraction, I was gratified to pick up my pictorial plan, with annotations, and was able to jump right back on board. WHS2WHS Basic is a geat way to keep track - and it's free!

[mediadogg]

Windows Home Server Migration Assistant (Basic and Pro) V1.02 Available

(Basic and Pro)

• - Check for already running added
• - Log / View added to Main Menu of Basic version
• - "Open" added to Target disk context menu
• - Source abd Target servers UserID is now restored across sessions
• - Tab order for controls has been improved to be more intuitive
• - Target disk volumes now ordered in drive letter sequence
• - Improved error messages and log entries
• - Double click on blank rows of disk volume brings up disk info edit
• - Discover folder size estimate improved (see note for limitations)
• - When a copy is checked "Complete" on the plan, the Souce server name is replaced by the Target server name in the folder path

(Pro Only)
- Improved discovery of BC and RichCopy on C: drive
Known Bugs / Limitations
- Discover folder sizes reports lower due to skipping certain Microsoft system folders and files (Looking into how to fix this)

Today's Tips:

• Discovering one of the biggest advantages of WHS 2011: Removing a drive in WHS V1 disables the entire share pool for the entire time, which could be 1-2 days for a 2TB drive. Removing a drive in WHS 2011 simply removes only the folders located on that specific drive. Server rolls on, no complaints. And if that drive is part of a RAID cluster, you are likely not even to lose access to those folders. Quite an advantage.
• Adding a pre-populated drive to WHS 2011: Easiest and best way is to first mount the drive on WHS 2011. Define the desired folders tobe on that drive. WHS 2011 does all the work. Bypass setting up the backup schedule for now. Next, remove the drive and take to to whatever other system and glom the data into those folders. Then take it back to WHS 2011. Later, you can set up a backup schedule. If the folders you want to set up are not one of the WHS pre-defind folders, you can even skip the initial step. On any system, set up a folder structure ServerFolders\MyFolder1, ServerFolders\MyFolder2, etc. Then take the drive over to WHS 2011 and use task "Add a Folder." The folder structure is optional, but I like to keep it that way to be consistent with the way WHS 2011 sets it up.

[S-F]

I think I must be missing something but, Why wouldn't you simply copy and past the contents using terracopy?

[mediadogg]

I think I must be missing something but, Why wouldn't you simply copy and past the contents using terracopy?

Thank you for your question.
Answer: Of course you can. I also use copy, xcopy, robocopy, beyond compare, drag and drop, etc.

Here is what you missed (first saentence of my post): "The WHS Migration Assistant (WHS2WHS) is designed to help primarily with the planning of the data move from Windows Home Server to Windows Home Server 2011"

In fact WHS2WHS Basic does no data copying at all! It is simply a neat way of planning it out in a pictorial way. You can drag and drop folders around from disk to disk. You can define disks and folders, edit them or WHS2WHS will go out and find them for you, and then you can manually update those as well. As you move the folders around, it keeps track of the free space on each disk, adding and subtracting as you go. you can mark folders "Complete" as you make progress copying the data, using any method, and you can print the planning template. The plan keeps notes about each disk that you type in, or you can just print the template and write on it. I'm sure you have seen those furniture layout type tools - well, its sort of like that, and its free.

For added convenience, WHS2WHS Pro, will use all the planning information to partially automate the actual data moving process. I have built-in some automation for Microsoft RichTools and Beyond Compare - only because they are what I use, nothing other than that. I also put in a way for you to call your favorite program, and pass the pre-formatted paths. It will also pass user credentials and disk information in case you need that. I find it very useful, even though I am reasonably expert at using all those copy methods manually, it helps me to be able to execute them within a framework. And it helps me keep track, so I don't goof up and lose my data.

That's it. Thanks for the opportunity to explain things, Why not give it a try, just for grins?

[mediadogg]

Working on some improvements.

Today's Tip: What do do if your WHS shares become "unknown"

This can happen if you remove the drive (or it otherwise becomes unavailable) where the shared folders are located.

But it can also happen if you inadvertantly change drive letters. This happened to me today, and it shook me up for a second.

A couple of solutions:
- reset the drive letter(s) back to what they were
- use the automatically supplied to rediscover the missing folder task: just be aware that if the folder is one of the pre-defined folders, you will not be able to do this. You will get a "Folder already exists" message. In this case, I just renamed the real folder (in my case it was "Recorded TV" renamed to "Recorded TV1"). Then I went ahead re-established the missing Recorded TV onto the correct drive. After that, I just cut/pasted the contents from "...TV1" ro "...TV" and then deleted the "...TV1" folder. This went fast becuase the cut/paste is just an on-disk folder move.

[mediadogg]

Windows Home Server Migration Assistant (Basic and Pro) V1.04 Available

(Basic and Pro)

• Added disk inventory. View Inventory plus Physical Drives Used
• Added Tooltips, tooltips On/Off buttons
• Added help info header in edit dialogs
• Added task-oriented Help Topics
• Added smart sorting to Disk columns (by folder, drives not included)
• Added discovery of DE/Shares
• Rewording of some menu items and messages for clarity
• Discovery of folders sizes is more accurate
• Additional log messages and error recovery
• View Plan Image now suppresses display of File dropdown
• Various bug fixes

(Pro Only)

• Disk nventory is saved and loaded with plan save / load

Please see the set of task-oriented Help articles in the lead post.

As usual, customers of the Pro version get a free upgrade.

[mediadogg]

Not getting any feedback on this, but I will continue to post my own. I am about 50 - 60% complete with my 20TB migration. I am so happy with Migration Assistant Pro. Helps me keep track through all the distractions of disk failures, motherboard hangs, MS upgrades that interrupt overnight copies, etc. I just refer to my plan, and keep notes as to where I am in the process. Fortunately, both Beyond Compare and RichTools will just take off from where they were, and MA remembers all your copy setups across sessions. So I just restart the copies and go (again) for coffee.

Today's Tips:

• If / when you get a message from WHS saying maybe you should check/repair a disk or continue ... You guessed it, better stop and do the check/repair. It is relatively fast (unless you select surface scan), and it will save you a ton of grief later.
• When messing with drives taken directly from WHS V1, and using the files found in DE\Shares, be careful not to mess with anything else. I found that Microsoft's very smart reparse points or whatever they are called, still know about the shares that are not on that disk, but located elsewhere. And they still sort of work, even when you move the drives. Example, you can goto drive:\shares and you will see all your WHS V1 shares. They mostly don't work, but they try. I would NOT recommend attempting to directly erase this folder or its contents. Instead, once you have copied (or moved) all the data from drive:\DE\Shares, go to disk manager, delete the volume, re-establish it, reformat and (ideally) run a WHS Repair and then you will have a nice clean new disk to use in the new configuration. (If anybody can shed more light on this, I would like to know.)

[mediadogg]

Continuing this public monologue ... (hopefully it is helping somebody) ...

Today I discovered that my RAID JBOD Spanned cluster went belly up over night. Luckily, given available backup data and the ability to regenerate some missing data that did not make my backup window, I will be OK - just sets me back a couple of days of intense copying. Thanks again to MA, I have ongoing migration snapshots that helped me figure out what I needed to check and re-copy.

Now I need to get a handle on why RAID seems to be useless for me:

- maybe I have a bad board: I hope not. It is a well-respected Asus-NVDIA implementation. I haven't read any overly bad press on it.
- maybe I have a bunch of bad drives. Highly possible. I buy cheap consumer-grade SATA drives from the usual vendors. I have a couple of the WD "EARS" variety that always seem to be causing issues, even though they always pass SMART tests. I think I will RMA the most likely culprit today. It was the lead disk in the cluster, and I had removed the WD Advanced Format jumper (pins 7-8) since I'm running under Windows 7 / Server 2008 R2.

But the latest news (for me) is this: There is such a thing as "RAID-compatble" drives!!! Silly me. And you guessed it, WD and some other vendors do not recommend their consumer grade cheapo drives for RAID. This has to do with the implementation of "deep cycle error recovery" or something like that. Basically, it causes the disk to appear to be dead to the world for up to two minutes while it attempts to reccover from nasties. That's OK for consumer grade, but for enterprise (RAID), you need to disable this feature, which then gives you "TLER" - Time Limited Error Recovery, which limits the delay to 7 seconds. Most RAID controllers give the drive substantially longer that that (maybe 15 seconds) before it gets dropped from the array. See a WD KB article here.

Available drives run the gamut from (more expensive) TLER-enabled RAID drives to el-cheapo consumer drives that can be configured for TLER (apparently WD does not advertise this, but you can get some kind of utility from them - see their web site). Now I understand why my first two attempts to set up RAID ararys failed almost immediately due to dropped el-crappo consumer grade non-TLER enabled drives.

I've got my migration on hold for a bit until I figure out whether I want to use RAID, and whether I will go the route of TLER-ing the cheap drives or buying a couple of enterprise grade drives.

Anybody with some good poop on this to share with me, please do so. Thanks in advance.

(Just found that even WD Caviar Black are called "Desktop" hard drives - therefore still not recommended for RAID according to that article. Hmmm ...)

[DieHard]

I'm watching this thread, so please continue to comment on your experience,

[mediadogg]

I'm watching this thread, so please continue to comment on your experience,

Great. Well I got a copy of that TLER utility and attempted to set it ON for the three drives I had in a HW JBOD span with my Asus-Nvidia setup. One of them, "EADS" style got set. The other two were "EARS" and it seems are fixed by WD to not allow the commands.

So I was considering swapping the EARS drives for a couple of EADS drives from my pool (and thereby needing a bunch more copy/verify time), but then I read that one has to be careful with that setting for drives that could fail during a rebuild. For example, if you get a failure in RAID 5 and you start a rebuild, then the TLER setting will make it much easier for a second (unrecoverable) drive failure to be reported while the build is in process - good bye whole array's worth of data. If TLER were not set, the drive could mess around for up to two minutes and possibly recover, thereby enabling the rebuild to continue.

So, that pretty much makes my decision: no RAID5 unless I use enterprise drives that are reliable enough to risk the TLER setting (and WD come that way by default.)

I am now about to retry a JBOD spanned array, using the TLER'd drive as the lead drive, followed by the two EARS drives. But first I am going to break the array, turn off RAID on those ports, and do a complete bad sector scan on all 3 drives, then restore them to RAID status and remake the array. Probably will take 1-2 days. But keeps a retired geek busy.

If that breaks again before I get a chance to complete the server and resume my Robocopy sync with the backup server (WHS V1), then I might go ahead a spring for a couple of Caviar Black drives. Cheaper than ENterprise drives, but possibly reobust enough to risk the TLER trick (that is if WD hasn't precludued it).

[mediadogg]

I found this review of an Adaptec controller, as I was exploring options to upgrade my RAID experience. The most interesting part of that review was not so much for the controller, but for the advice: (1) don't use cheap drives, and beware of cheap hot-swap SATA backplanes. Well, I'm using both! Read the review for yourself - form your own opinion.

[mediadogg]

I'm back. Feels like I've been in a bad dream. Nightmare. System hanging on boot, sporatic drive dropouts, corrupted file copies, broken RAID and spanned arrays.

I was blaming everybody and everything. What kind of dummy am I?

It came to a head when I made the heartbreaking discovery that both my copy programs were being faked out by identical file attributes masking underlying corruption (are we talking about politics?) after viewing one of my favorite videos - it was corrupt, while the original was fine - but both copy programs said the files were identical (what the ... ???). So I turned on the Big Gun: Beyond Compare's binary compare. Sure enough, that file, as well as some others (not all of them, thankfully) had massive differences in the middle of the (typically huge) file. What to do now?

Then it hit me: what component affects everything? What one item did I not upgrade when I went to my new chassis? You got it: THE POWER SUPPLY!!!!

My heroic but inadequate 450W power supply had been trying to keep up with the demands of 11 green drives, powering up whenever they pleased, RAID and spanned drive support trying to keep multiple drives in sync, 8GB of memory, 2 external USB drives, an HD PCIe graphics adapter - it's a wonder the poor thing didn't go up in smoke! I can easily see now how the boot hangs, the frozen mouse pointers, file transfer issues, failing drives and CHKDSKs, etc. were happening.

I did a sizing according to the ASUS MB recommendations, and my PSU was about half the recommended power for a system with 6 drives - mine has 11 (not counting the 8 drives in my external Sans Digital cabinet.)

Upshot: This morning I installed an 800W Thermaltake (heh heh!) monster PSU. My rig literally screamed in ecstacy during boot. Drive lights flashing like crazy. Smooth, fast file transfers. No more drops when multitasking and multiple shares being accessed (been running all morning.)

I'm now painstakingingly running a binary compare on my spanned arrays, and re-copying the files that fail the comparison. So far, so good. My mouse has not frozein once. And I have been playing videos and doing other work simultaneously to put on some stress. Disk lights just keep on flashin' (sounds like a song ...)

If this continues, and my confidence is rebuilt in myself and my system, I will likely go back and revisit RAID.

So, todays TIPS:

(1) Don't turn off TLER on non-enterprise drives. Anything that delays the drive's response (such as a weak PSU) will cause the drive to drop out of an array.
(2) If you are a DIY'er, make sure to properly size your PSU. With modern power management, it is not unusual for power surges to be sporatic and uneven. In the old days, everything powered on in the beginning and stayed on. If you didn't blow a fuse when you hit the power switch, you were good to go. Not any more!

[mediadogg]

In celebration of my getting somewhat back on track with my migration:
Windows Home Server Migration Assistant V1.05 Available

• Pro Version big price drop - On Sale!
• Removed Stop button from Copy loop (Pro)
• Added sort to include drive letter to context menu
• Fixed bug in "mark complete" context menu item

[emraith]

I found this review of an Adaptec controller, as I was exploring options to upgrade my RAID experience. The most interesting part of that review was not so much for the controller, but for the advice: (1) don't use cheap drives, and beware of cheap hot-swap SATA backplanes. Well, I'm using both! Read the review for yourself - form your own opinion.

Hmm, I just picked up an IcyDock 4-in-3 backplate for my planned migration to WHS 2011... Was around $120..so not sure what that user's review means as far as cheap, there were some around $50..but considering this is a review of a +$500 raid controller, I'm guessing he would include the IcyDock with the cheap price range, heh. Oh well, guess I'll find out eventually what happens. Not screwing with raid myself though.

[mediadogg]

Hmm, I just picked up an IcyDock 4-in-3 backplate for my planned migration to WHS 2011... Was around $120..so not sure what that user's review means as far as cheap, there were some around $50..but considering this is a review of a +$500 raid controller, I'm guessing he would include the IcyDock with the cheap price range, heh. Oh well, guess I'll find out eventually what happens. Not screwing with raid myself though.

Yeah, I hear ya. At the time of that post, I was ready to believe anything. But as my subsequent post reveals, it was faulty HW planning on my part. No problems since I replaced the power supply with an 800 watt unit. I will do more testing and will post results soon. I have since completed most of my initial migration. Now working on replication and backup methods. I was able to arrange for a bunch of drives to be left over in a Sans Digital tower. The RocketRaid 622 software appears to be very capable and easy to use. So I am building two Raid 5 clusters with a spare fail over drive, as we speak. So far, so good. The software is taking forever to initialize the drives. I suspect it is doing a surface scan to isolate bad clusters - a very good thing.

And I am using Startech drive bays (3 into 2) and don't have any issues so far.

Thanks for your perspective. I'm eager to hear about your progress.

[mediadogg]

Windows Home Server Migration Assistant V1.06 Available

• Fixed documentation error and menu misspelling
• Improved sorting of disk blocks based on drive letter

Customers of the Pro version need only to login to the Shopping site to download the free upgrade.

Status on my own migration:
I'm (crossed fingers) nearly there. I backed off for a few days to give myself a break and also play with my Grandson (so cute!).

Server is still running without major issues, although I do see failed backups now and then. Another thing I noticed is that my 3TB Hitachi USB drives will often not show up when the server is rebooted. Don;t know why. And the WHS Console doesn't complain. The drives don;t show up, the shares don't show up - it's like they never existed. But when I recycle the power on the USB drives, the plug and play finds them, and the shares pop up in the Console. It could be worse.

I realize that using USB drives is not the most ideal thing for shares, but these particular shares are convenience copies of some favorite DVDs. I figure that the fact that the drives totally power down when not being used is a hedge against not having a backup (other than the original discs that I could re-rip). Less spinning - less heat - maybe they will last a long time before failing. Who knows.

So far, the spanned volumes I created with Disk Management, the RAID mirror I set up with the NVidia motherboard support, and the RAID 5 volumes I set up with the Sans Digital / Hignhpoint RR622 controller ad software, have been operating as expected. No drives dropped or failures since I put in my new power supply.

Absent any unexpected critical failures, I anticipate another week or so of finalizing my replication and backup operations. I plan to have a three-tier setup: MainWHS will have shares on a combination of single-disk, mirrored, and spanned volumes. Selected folders will then be copied via automated scripts to a couple of larger RAID5 volumes in the Sans Digital Tower (4 x 2TB = 6TB RAID5 and 3x1TB = 2TB RAID5, plus 1 x 2TB hot spare supporting both arrays).

All major shares are then further incrementally backed up overnight to a WHS V1 system running on my beloved Acer EasyStore, which will also have selective folder replication.

Even with all that, I don't feel confident that I will never lose data. For example, I discovered that at some point, my data replication methods (RichCopy) dutifully copied a few corrupt files onto the backup, rendering both the orginal and backup useless. I need to modify the script to avoid copying a zero length file.

[mediadogg]

I just had an idea that I think is good, so I will share it.

Background:
Using RAID5, you are between a rock and a hard place. Most of us do not have official enterprise drives capable of running with TLER enabled. That makes RAID5 very risky in the event of a drive failure. It is well documented that the probability of consumer grade drive dropouts during the long rebuild time is very high - and as we all know, two drives lost in RAID5 = "toast."

Partial Solution:
Since my HighPoint RR622 and SW supports a pool of hot spare drives, I set up a 2TB drive to serve as a hot spare for the two RAID5 clusters I have defined, obviously counting on the improbability of both clusters needing the spare concurrently. So today I decided to buy one enterprise 2TB drive to use as the spare instead of the consumer grade spare. My rationale is that even though the arrays are built from subpar drives, if I get a failure, at least I have substantially reduced the chance of a second drive failure during rebuild - which I think really bangs on the drive being added, leading to the high failure rates noted during the process.

As a hot spare gets used to replace a failed drive, I will replace it with another enterprise drive, spreading the cost out over time, and gradually my arrays will migrate to all enterrise drives. (About 2.5 times the ccost / TB of consumer grade.)

What do you think? Good idea, or am I still in denial?