Find the answer to your Linux question:
Page 2 of 2 FirstFirst 1 2
Results 11 to 16 of 16
Your best bet is to isolate them. But my own preference would be to fix the network question and then look to the USB. .02...
Enjoy an ad free experience by logging in. Not a member yet? Register.
  1. #11
    Linux Guru
    Join Date
    Nov 2007
    Posts
    1,755

    Your best bet is to isolate them. But my own preference would be to fix the network question and then look to the USB.

    .02

  2. #12
    Just Joined! TheAlmightyOS's Avatar
    Join Date
    Feb 2009
    Location
    Earth?
    Posts
    42
    Quote Originally Posted by HROAdmin26 View Post
    Your best bet is to isolate them. But my own preference would be to fix the network question and then look to the USB.

    .02
    Good point. You wouldn't happen to have any advice on how to get started on that problem would you?

  3. #13
    Linux Guru
    Join Date
    Nov 2007
    Posts
    1,755
    Start at end of the "chain" (HDD > Cable > Controller > NIC > Network...) and move up until you find the bottleneck.

    Here's a good starting point.

    Your drives all have different max UDMA sync speeds (133, 100, and *none*) - I've seen issues with DMA settings with mismatched drives cause many a nasty thruput issue.

    Confirm the settings and confirm the HDD > controller communication with hdparm. Also isolate/test with specific HDD's which ties any possible issue to a certain HDD or a certain IDE controller.

  4. #14
    Just Joined! TheAlmightyOS's Avatar
    Join Date
    Feb 2009
    Location
    Earth?
    Posts
    42
    Well whatcha know. Dress me up and call me sue, it IS a controler issue!!
    Code:
    [root@EPSILON thealmightyos]# hdparm -Tt /dev/hda
    
    /dev/hda:
     Timing cached reads:   1240 MB in  2.00 seconds = 619.61 MB/sec
     Timing buffered disk reads:  204 MB in  3.01 seconds =  67.79 MB/sec
    [root@EPSILON thealmightyos]# hdparm -Tt /dev/hdb
    
    /dev/hdb:
     Timing cached reads:   1218 MB in  2.00 seconds = 608.33 MB/sec
     Timing buffered disk reads:  168 MB in  3.00 seconds =  55.91 MB/sec
    [root@EPSILON thealmightyos]# hdparm -Tt /dev/hde
    
    /dev/hde:
     Timing cached reads:   1238 MB in  2.00 seconds = 618.27 MB/sec
     Timing buffered disk reads:   60 MB in  3.07 seconds =  19.51 MB/sec
    [root@EPSILON thealmightyos]# hdparm -Tt /dev/hdf
    
    /dev/hdf:
     Timing cached reads:   1246 MB in  2.00 seconds = 622.60 MB/sec
     Timing buffered disk reads:   68 MB in  3.05 seconds =  22.33 MB/sec
    [root@EPSILON thealmightyos]#
    I feel like a dick now. Thanks for pushing me onto the right path. hde and hdf (the HDDS on the mobo) are vastly slower when it comes to buffered reads. Though the drives on the controller card are faster, they look to be much slower then your example.

    So now that we know, how do we go about fixing it?

  5. #15
    Linux Guru
    Join Date
    Nov 2007
    Posts
    1,755
    I would do a few more tests before changing anything...

    A) How are your partitions set up? Are you using any LVM or raid volumes? What you're looking for is to copy data from a volume that resides on ONE drive. So copy some data from hda and hdb to a network share - is thruput "normal?"

    Then try from hde and hdf and see if the results change.

    B) I can tell from your dmesg log that the 20GB drive (hde) is probably just not going to play nice on the same IDE cable as the 40GB (hdf.) The slow sync speed of hde will bring the whole bus down. I would remove that drive from the machine, or else put it on its own IDE channel/cable. That will allow the 40GB to get better performance.

    C) What are the current DMA/Readahead/etc. settings (as mentioned before):

    Code:
    hdparm /dev/hdX

  6. #16
    Just Joined! TheAlmightyOS's Avatar
    Join Date
    Feb 2009
    Location
    Earth?
    Posts
    42
    Well, hde and hdf are my root and home drives respectively. The raid card reserves hda, hdb, hdc and hdd for some reason. hde only has two partitions (root and swap). hdf is a single partition. So are hda and hdb. And no, thruput is not normal. No mater what drive I use. Even sudo copied a few items from the root drive. Same results

    if I remove my root drive then I will have to reinstall.

    Code:
    /dev/hda:
     multcount     = 16 (on)
     IO_support    =  0 (default 16-bit)
     unmaskirq     =  0 (off)
     using_dma     =  1 (on)
     keepsettings  =  0 (off)
     readonly      =  0 (off)
     readahead     = 256 (on)
     geometry      = 19457/255/63, sectors = 312581808, start = 0
    
    /dev/hdb:
     multcount     = 16 (on)
     IO_support    =  0 (default 16-bit)
     unmaskirq     =  0 (off)
     using_dma     =  1 (on)
     keepsettings  =  0 (off)
     readonly      =  0 (off)
     readahead     = 256 (on)
     geometry      = 65535/16/63, sectors = 240121728, start = 0
    
    /dev/hde:
     multcount     = 16 (on)
     IO_support    =  1 (32-bit)
     unmaskirq     =  1 (on)
     using_dma     =  1 (on)
     keepsettings  =  0 (off)
     readonly      =  0 (off)
     readahead     = 256 (on)
     geometry      = 39761/16/63, sectors = 40079088, start = 0
    
    /dev/hdf:
     multcount     = 16 (on)
     IO_support    =  1 (32-bit)
     unmaskirq     =  1 (on)
     using_dma     =  1 (on)
     keepsettings  =  0 (off)
     readonly      =  0 (off)
     readahead     = 256 (on)
     geometry      = 65535/16/63, sectors = 78165360, start = 0

Page 2 of 2 FirstFirst 1 2

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •