Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Sign In with Google

Our Sites

covecube.com 
community.covecube.com 
blog.covecube.com 
wiki.covecube.com 
bitflock.com 
stablebit.com 

Poll

No poll attached to this discussion.

In order to better support our growing community we've set up a new more powerful forum.


The new forum is at: http://community.covecube.com


The new forum is running IP.Board and will be our primary forum from now on.


This forum is being retired, but will remain online indefinitely in order to preserve its contents. This forum is now read only.


Thank you,

Slow copying TO my DrivePool

edited February 2013 in DrivePool
Just installed 1.2.7226 and its in the process of duplicating/balancing - although I migrated over 4 of my WHS v1 drives with the files already duplicated so was expecting the duplicating process to go faster than if DP had to make dupe copies.

I'm trying to copy a bout 70 GB's of data - via a Gigabit network - from my Laptop to a Network Share - and the average speed is only in the 3.2-4.0 MB/s range.  TeraCopy is reporting it will take almost 5 hours to copy!  I'm seeing much faster speeds copying FROM the Network shares. 

I stumbled on an old thread noting this:  "go into the stablebit drivepool settings in the dashboard and enable all of the direct file i/o and fast i/o options and try again. If they were unticked you will see a large speed increase" - but a later post notes those settings are not accessible any longer?  Am I missing something in some setting somewhere - or is this b/c DP is actively duplicating/balancing - or is this the norm copy speed to the DrivePool?  I have several hundred GB's of files I need to move to my server - but at this rate, it's going to take days.

Help?

Comments

  • I read through the manual - didn't see anything that would explain why this would be so slow.  I've never encountered this slow copying with anything else on my network or when copy to my old WHS v1.

    If the slowdown is being caused by the flurry of activity on the server and/or the duplication and balancing in place (which I don't see an option to pause?) - is it better to create a folder OUTSIDE the pool - on the POOL drive - copy the files to that folder - then MOVE the files into the appropriate Pool folder(s)?

    hoping somebody can advise this evening...thanks much


  • Resident Guru
    It certainly shouldn't be that slow on a gigabit network (I'm testing mine right now, averaging 53MB/s copying to a duplicated folder and 78MB/s reading it back afterwards, with the server otherwise idle).

    Possibly your gigabit network might be running at a lower speed (e.g. a bottleneck somewhere); also, if the laptop has wifi check that it isn't still using that even when you've got it plugged in via a network cable.

    The settings mentioned in the old thread are obsolete; DrivePool does offer read striping and network IO boosting via the Dashboard but those should be on by default anyway (check under Dashboard, StableBit DrivePool, Disks).
  • Don't think it's my network.  Not having any other issues speedwise.  Wifi is diasbled on the laptop.  Yeah, I saw those settings and they are checked off.

    The duplicating and balancing is done - so tested copying a file to a few places;
     a) to a DrivePool Shared folder:  approx 30-35 MB/s, slightly faster copying back (35-42 MB/s)
     b) to a separate folder on the DrivePool Drive (not duplicated, outside the pool) - pretty much the same speeds
     c) to the second partition of the primary WHS drive (shared a folder there) - 58 MB/s - almost 65 MB/s copying back.

    So it seems the DrivePool virtual drive is slowing things down?  How do I troubleshoot this?
  • Resident Guru
    Ah, wait... if you were copying those files to the server WHILE the pool was already in the middle of a balancing run, that could explain it. There's a reason balancing is normally scheduled for 1am, it can be very disk-intensive. :)

    Re your testing after the balancing was finished, it almost looks like your issue might be more system disk versus non-system disk rather than drivepool drive versus non-drivepool drive? Can you clarify your "b" result, as you can't actually have a folder that is both "on the drivepool drive" and "outside the pool".
  • B) was me copying to the old "shares" folder from the movement of data from the old WHS v1 folders. What I meant to say was on the DrivePool drive but NOT in one of the duplicated folders.  So since the duplication run was done, makes sense that the speeds were the same as (a).

    So how do I go about testing?  I'm now moving some files from a USB 2.0 drive connected to the server to the DrivePool and it's maxing out at 31.5 MB/s (pretty consistent speed).  I only have the one eSATA port currently and it's in use for my Sans Digital 4 drive enclosure - so can't test copying from an eSATA drive until my add-in card gets here.
  • Resident Guru
    31.5 MB/s is about as good as it can physically get with a USB 2.0 interface.

    I have some travel to do, will next check the forum in a few hours (possibly several hours).
  • If you can point me towards any testing resources - that would be appreciated.  I'm headed to bed here (East Coast time zone).
  • edited February 2013 Resident Guru
    A simple test of DrivePool's impact on file IO:

    Let's say P is your pool drive and K thru M are your pooled drives.

    1. Create two large files of identical size (preferably 1+ GB) but different (preferably random) content on a non-pooled drive.
    2. Copy the first file into a non-duplicated folder on P. Time this.
    3. Find the particular pooled drive where DrivePool stored the file.
    4. Copy the second file into a non-pool folder on that same drive. Time this.

    Theoretically, the difference in copy time between the two files is DrivePool's impact on file IO for reading and writing. In practice, there can be discrepancies due to background processes, disk fragmentation/layout, etc, and you should repeat the test multiple times to build up a working trend.

    (I find TeraCopy handy for this sort of thing; if you keep it from closing after a copy, it can tell you how long the copy took and calculate the average speed for you).
  • OK, will try later.  Assuming the different data is to avoid cache issues?  I use TeraCopy also.  I'll report back.
Sign In or Register to comment.