Author Topic: Unable to populate my flexRAID with ROBOCOPY, teracopy  (Read 3125 times)

Offline toddlmorgan

  • Newbie
  • *
  • Posts: 3
  • Karma: +0/-0
    • View Profile
Unable to populate my flexRAID with ROBOCOPY, teracopy
« on: September 14, 2012, 09:50:32 am »
Let me start out with this. I REALLY want to like flexRAID ... but my user experience has been very poor. I have a purchased copy and I’m running a real-time RAID.

My primary problem is that I can't get all of my content onto flexRAID before it falls over. I've had a MCE since 2005 and I'm a hoarder so I have a LOT of content.

I should point out that on a positive note … I do NOT believe that I have LOST any content due to corruption or the like :- ) So that’s a big tick in flexRAID’s box.

I have an existing unRAID box which is now 100% full (about 18 TB) that I am trying to copy content off. I have mostly HiDef TV content all nicely arranged and named.

I have a reasonable good setup (Norco 4224, Xeon, 8GB RAM, Win 7 Pro, 12 x 3 TB drives - 1 is parity) rest are partially populated) - this was to be my UBER MCE box....
I’m using 2.0u9 with CruiseControl Realtime RAID.


I have attempted to copy off the content using many different techniques:
- Teracopy
- Robocopy
- Winmerge

I specifically used those as windows copying is unreliable - especially for large copying across the network (and as it takes soo long to copy files overnight etc I want to be reasonably certain it will actually work).

They all seem to fail with a variety of the following errors:
- OutOfMemoryErrors (I can usually got a 200-300 GB of content copied across) before this happens. Files are mostly large (1-10 GB) but there are the associated thumbnails etc

eg
[2012-09-14 22:46:53,924] WARN : Exception while dispatching incoming RPC call
com.google.gwt.user.server.rpc.UnexpectedException: Service method 'public abstract com.tchegbe.flexraid.client.bean.ActionStatus com.tchegbe.flexraid.client.rpc.RPCService.getFlags()' threw an unexpected exception: java.lang.OutOfMemoryError: Thread creation error: 8
   at java.lang.Void.<unknown>(Unknown Source)
   at java.lang.Void.<unknown>(Unknown Source)
   at java.lang.Void.<unknown>(Unknown Source)
   at java.lang.Void.<unknown>(Unknown Source)
Caused by: java.lang.OutOfMemoryError: Thread creation error: 8


From reading through the forums I gather this has something to do with the number of files and the fact that this is a 32 bit process which is constrained by the fact that you've, at least partially, written this as a Java app that you've compiled into a native app using excelsior jet (who are still looking at 64 bit compiler 1-2 years after they said they would get started).



PendingDelete errors

NOTE: I was OUT to dinner whilst this was happening so I didn't request any Delete. This was a robocopy process that was triggering this activity.

[2012-09-14 19:24:40,232] ERROR: Delete live file failed for: PendingDeletes(fileName\TV\Comedy\Some Funny Show 1x01.mp4,realPath=C:\FlexRAID-Managed-Pool\class1_0\{4369a460-f6d5-43f5-8c46-3d81f5ac35e4}\\TV\Comedy\Some Funny Show 1x01.mp4,link=C:\FlexRAID-Managed-Pool\class1_0\{4369a460-f6d5-43f5-8c46-3d81f5ac35e4}\_flxr_\l\\TV\Comedy\Some Funny Show 1x01.mp4,started=true,completed=false)
[2012-09-14 19:24:40,232] ERROR: [deleteLiveFile]->[deleteLiveFile by handle]->[ReadByHandle] ErrorCode=998 Message=handle=00005CAC requested=32768 read=-1 errorCode=998
java.lang.IllegalArgumentException: [deleteLiveFile]->[deleteLiveFile by handle]->[ReadByHandle] ErrorCode=998 Message=handle=00005CAC requested=32768 read=-1 errorCode=998
   at java.lang.Void.<unknown>(Unknown Source)
   at java.lang.Void.<unknown>(Unknown Source)
   at java.lang.Void.<unknown>(Unknown Source)
   at java.lang.Void.<unknown>(Unknown Source)
[2012-09-14 19:24:40,236] WARN : Disabling all live operations! All future operations will get an access denied error...


I'm guessing that this is something to do with the lack of support for programs that don't preallocate files ahead of time ... that or it's the internal reshuffling that's going on trying to balance the files across the other drives


I also get "too many links" when copying

[2012-9-4 3:22:46.726072][488]      [GetOperationResult] Unknown exception...
[2012-9-4 3:22:46.741121][488]      [onTruncate] GET_LIVE_INFO error = 31
[2012-9-4 3:22:46.741122][488]      [ERROR]:[OnSetEndOfFile] [onTruncate: GET_LIVE_INFO error] ErrorCode=31 Message=Too many links
         \TRIP VIEWING\ENCODED\Something else to watch.mp4 ->
         
         
So how do I get stuff on there … well … I verify … I reconcile …I backup the DBs… and I nuke the array and recreate it again. The painful part is that it takes somewhere between overnight and a whole day to (re)compute the parity. I'm a very stubborn person ... and I'm committed to this now as I'm out of storage on my other server.

I’ve now come to the conclusion that the only reliable way that I can actually get the content onto the drives is to put it on them BEFORE the array is constructed. I would have liked to know that beforehand as the folders are simply too large to get onto 1 drive without spanning etc … one of the reasons I like flexRAID and unRAID.

This is extremely disconcerting ... as others have said ... I have no issue if things need to be slowed down (if flexRAID can't keep up) ... but I do object to letting the system think it can handle things ... copying and then falling over and costing me literally wasted days of parity computation. I have happily copied many TBs of data off and on the unRAID server and onto other systems in one go without missing a beat ... it wasn't the fastest thing out there 50-60 MB/s but it just worked.

I have been reading the forums so I have all the required log files ready to go and 7zipped up.  I’m not willing to air all my privates in public so if Brahim can tell me how to get the files to him I will get them there.


On the usability front I would also suggest the following needs some further thought:
-   don’t just silently lock (a log file is silent as far as a UI is concerned) the array when the ONLY option is to fix it via Verify and Reconcile. There is only ONE realistic option. Just do it. It’s an internal matter. I shouldn’t need to be bothered by it. You should inform me that it’s happening eg flash a message on the Web UI or change the colour of the browser page or something visible. At the very least I’d prefer it if you provided an option to “automatically” heal the array. In other RAID systems I see this listed as “auto rebuild”. I can see perhaps that some people may wish to keep files available (ie “read-only” mode) but I want my array up and protected ASAP … not waiting days for me to notice that I can’t copy stuff onto it.
-   The GWT user interface frequently results in buttons that I can’t select “eg save or apply” at the bottom of pages due to the size of the operating system display. Or the restore from database list doesn’t actually scroll properly so I can’t restore the latest DB when I’m forced to do a reconciliation.

I know that there is a new release due today/tomorrow. One more roll of the dice perhaps ...


Offline Brahim

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 8,547
  • Karma: +204/-16
    • View Profile
Re: Unable to populate my flexRAID with ROBOCOPY, teracopy
« Reply #1 on: September 14, 2012, 10:27:11 am »
There is another thread that discuss the issue at hand here.

Use Windows Explorer to do the copying. Those programs, for whatever reasons, are creating too much fragmentation within the filesystem.
They fail to pre-allocate before starting to copy.

Secondly, you might need to re-create the RAID after changing the minimum allocation size property.
http://forum.flexraid.com/index.php/topic,1099.msg8937.html#msg8937

Offline DrBlaze

  • Sr. Member
  • ****
  • Posts: 281
  • Karma: +14/-0
    • View Profile
Re: Unable to populate my flexRAID with ROBOCOPY, teracopy
« Reply #2 on: September 17, 2012, 10:29:26 am »
" the restore from database list doesn’t actually scroll properly so I can’t restore the latest DB when I’m forced to do a reconciliation. "

On my system if I hit the refresh button I get scrollbars and can then select from the whole list.  Hope this helps.

Offline Brahim

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 8,547
  • Karma: +204/-16
    • View Profile
Re: Unable to populate my flexRAID with ROBOCOPY, teracopy
« Reply #3 on: September 17, 2012, 01:19:18 pm »
You are actually better off restoring an older backup than the last backup. So, do that.
As far as the UI issue, post a screenshot.

Offline creiglee

  • Newbie
  • *
  • Posts: 1
  • Karma: +0/-0
    • View Profile
Re: Unable to populate my flexRAID with ROBOCOPY, teracopy
« Reply #4 on: August 31, 2016, 12:36:18 am »


Have been using TC for two years now, on Win7 x64. Never had any "real" troubles with it. It's faster, yet I don't have exact information.

There's a bug in there somewhere, sporadically shutting down the process before it even starts, fixable by restarting TC a few (insert random number) times, till it works. It never quits in the middle, only when a Paste-command is made.

For the harder times, I've enabled the Scroll-Lock feature, to toggle between Windows' service, and TC.