Professional Web Applications Themes

Slow network transfers - SCO

Hi, I've noticed this on a few machines running 5.0.5/6 : transferring over the network is abysmally slow. I've tried rcps, ftps, nfs, rsync and paired netcats. Paired netcats were the fastest and managed a rate of about 2.77Mb/minute over a 100MB/s network, both plugged into the same switch. I've always transferred to a Red Hat box, varying from 7.2 through to 9. When I tried ftp I put both machines on their own 10/100 switch with no other machines connected. I noticed that I would get half a second's activity and then 5 odd seconds of nothing. Ftp transferred ...

  1. #1

    Default Slow network transfers

    Hi,
    I've noticed this on a few machines running 5.0.5/6 : transferring
    over the network is abysmally slow. I've tried rcps, ftps, nfs, rsync
    and paired netcats. Paired netcats were the fastest and managed a rate
    of about 2.77Mb/minute over a 100MB/s network, both plugged into the
    same switch.

    I've always transferred to a Red Hat box, varying from 7.2 through to 9.

    When I tried ftp I put both machines on their own 10/100 switch with
    no other machines connected. I noticed that I would get half a second's
    activity and then 5 odd seconds of nothing. Ftp transferred at about
    0.25 MB/s with 100MB/s cards. Using the same cabling etc RH to RH I got
    no errors reported after transferring an ISO of about 620MB. I don't
    remember how long that took, but it was measured in minutes, not hours.

    My first impression is that SCO has some sort of bandwidth throttle
    in place which I keep hitting.

    I should point out from the start I'm not flamming SCO, I realise
    it's probably something me or my coworkers have or havn't done. However
    it doesn't help when I need to transfer hundreds of MBs about for backup
    purposes.

    --
    Scott Burns
    Mirrabooka Systems

    Tel +61 7 3857 7899
    Fax +61 7 3857 1368

    Scott Guest

  2. #2

    Default Re: Slow network transfers

    In article <2nhwb.507$optus.net.au>,
    Scott Burns <com> wrote:
     
     
     

    That sounds a lot like what happens with a duplex mismatch.
    Or with bad cables running n FDX and one pair not being up to
    snuff.
     

    Nope. And I've seen this in other systems where SCO is nowhere in
    site.

    Check your switch settings if it's a managed switch. Check your
    cables. Then try starting at the lowest level on both machines,
    10Mb/sec HDX and work up. It could be an underperforming NIC
    card and there are many out there. Realtek has been highly
    problematic in the past.


    --
    Bill Vermillion - bv wjv . com
    Bill Guest

  3. #3

    Default Re: Slow network transfers

    Scott Burns wrote: 

    What is the HW, the make and brand of NIC, and the output of
    llistat -l ?

    Mike

    --
    Michael Brown

    The Kingsway Group
    Mike Guest

  4. #4

    Default Re: Slow network transfers

    Mike Brown wrote:
     
    The machine I have available now is a P3/500 with 128Mb RAM, SCSI
    HDDs, and generally me as the only user. My bash prompt turns red when
    the load average hits 1, and it has been at least six months since that
    happened. I've seen slow network transfers when working on client
    machines, mostly of the same vintage, but with twice the RAM. We are
    currently running 5.0.5. We almost always use Intel NICs. We used
    NetGear for a few machines but found they were not up to scratch. The
    curtrent NIC in question is an Intel PRO/100B or PRO/100+ with autosense
    speed/duplex.

    The short llistat version:

    saturn:~# llistat -l > llstat.1
    saturn:~# l test.img
    -rw-r--r-- 1 root other 11798370 Nov 26 10:14 test.img
    saturn:~# time nc -w 1 pluto 9900 < test.img

    real 4m34.680s
    user 0m0.000s
    sys 0m0.720s
    saturn:~# llistat -l > llstat.2
    saturn:~# diff llstat.1 llstat.2
    12,13c12,13
    < In: 176608 0 1011 0 51455461 0
    < Out: 176649 0 345 0 36680514 0
    --- 
    16c16
    < 186 frames received destined for an unbound SAP
    --- 
    23c23
    < Last send time: 591544
    --- 
    saturn:~# uptime
    10:21am up 1:45, 9 users, load average: 0.00, 0.00, 0.00
    saturn:~#

    The long llistat version (llistat.2)

    Device MAC address in use Factory MAC Address
    ------ ------------------ -------------------
    /dev/net0 00:90:27:b2:5a:6e 00:90:27:b2:5a:6e

    Multicast address table
    -----------------------
    01:00:5e:00:00:01

    FRAMES
    Unicast Multicast Broadcast Error Octets Queue Length
    ---------- --------- --------- ------ ----------- ------------
    In: 186061 0 1055 0 52056589 0
    Out: 188207 0 365 0 50732683 0

    DLPI Module Info: 2 SAPs open, 18 SAPs maximum
    191 frames received destined for an unbound SAP

    MAC Driver Info: Media_type: Ethernet
    Min_SDU: 1, Max_SDU: 1514, Address length: 6
    Interface speed: 100 Mbits/sec

    DLPI Restarts Info: Last queue size: 0
    Last send time: 628854
    Restart in progress: 0
    Number of restarts: 0

    Interface Version: MDI 100

    ETHERNET SPECIFIC STATISTICS

    Collision Table - The number of frames successfully transmitted,
    but involved in at least one collision:

    Frames Frames
    ------- -------
    1 collision 0 9 collisions 0
    2 collisions 0 10 collisions 0
    3 collisions 0 11 collisions 0
    4 collisions 0 12 collisions 0
    5 collisions 0 13 collisions 0
    6 collisions 0 14 collisions 0
    7 collisions 0 15 collisions 0
    8 collisions 0 16 collisions 0


    Bad Alignment 0 Number of frames received that were
    not an integral number of octets

    FCS Errors 0 Number of frames received that did
    not pass the Frame Check Sequence

    SQE Test Errors 0 Number of Signal Quality Error Test
    signals that were detected by the
    adapter

    Deferred Transmissions 0 Number of frames delayed on the
    first transmission attempt because
    the media was busy

    Late Collisions 0 Number of times a collision was
    detected later than 512 bits into
    the transmitted frame

    Excessive Collisions 0 Number of frames dropped on
    transmission
    because of excessive collisions

    Internal MAC Transmit 0 Number of frames dropped on
    transmission
    Errors because of errors not covered above

    Carrier Sense Errors 0 Number of times that the carrier sense
    condition was lost when attempting to
    send a frame that was deferred for an
    excessive amount of time

    Frame Too Long 0 Number of frames dropped on reception
    because they were larger than the
    maximum Ethernet frame size

    Internal MAC Receive 0 Number of frames dropped on reception
    Errors because of errors not covered above

    Spurious Interrupts 0 Number of times the adapter interrupted
    the system for an unknown reason

    No STREAMS Buffers 0 Number of frames dropped on reception
    because no STREAMS buffers were
    available

    Underruns/Overruns 7 Number of times the transfer of
    data to or from the frame buffer
    did not complete successfully

    Device Timeouts 0 Number of times the adapter failed to
    respond to a request from the driver

    --
    Scott Burns
    Mirrabooka Systems

    Tel +61 7 3857 7899
    Fax +61 7 3857 1368

    Scott Guest

  5. #5

    Default Re: Slow network transfers

    Bill Vermillion wrote:
     
    >
    >That sounds a lot like what happens with a duplex mismatch.
    >Or with bad cables running n FDX and one pair not being up to
    >snuff.
    >
    >[/ref]
    Bad cabling I can rule out. I've been stung before and it's now the
    first thing I checked. I've used 2 different switches and 2 different
    hubs, all with the same result, but no problems elsewhere. I get full
    through put with the cables on other machines, and they are definately
    Cat5e. The only test I could not do was install Linux on the machine,
    or move the card to another machine for testing.
     
    >
    >Nope. And I've seen this in other systems where SCO is nowhere in
    >site.
    >
    >Check your switch settings if it's a managed switch. Check your
    >cables. Then try starting at the lowest level on both machines,
    >10Mb/sec HDX and work up. It could be an underperforming NIC
    >card and there are many out there. Realtek has been highly
    >problematic in the past.
    >
    >
    >
    >
    >[/ref]
    We almost always use Intel cards. I'll try flipping back to 10/Half
    in the morning if I'm the first one in (machines get shut down every night).

    Sorry about the delay getting back BTW, got stuck on site yesterday...

    --
    Scott Burns
    Mirrabooka Systems

    Tel +61 7 3857 7899
    Fax +61 7 3857 1368

    Scott Guest

  6. #6

    Default Re: Slow network transfers

    In article <iKSwb.586$optus.net.au>,
    Scott Burns <com> wrote: [/ref][/ref]
     [/ref]
     

    Hubs can be strange on devices with autmatic duplex/speed setting.
    And it could be a NIC card. I'm fighting one where some machines
    on the same switch will transfer at maximum rate but going to other
    machines on the same switch transfers fall from just about 10MB
    second [100mbit network] down to 40KB second. These are two
    different type of Unix systems - and both will go at high speed
    when talking to one of their like type on the same side of the
    switch but fail when talking to the other.

    Going outside the switch onto the internet I see transfers
    at high as 600KB [6Mbit/second] as all the machine are on a very
    fast backbone.

    We have a couple of people thinking this through at it makes no
    sense. The problem >may< be something with the 10/100/1Gbit
    on motherboard cards on the G4 or the 2 Xrack machines. They talk
    fine on the Cisco switch. And the BSD and MS machines [Intel
    hardware and Intel NIC on the BSD and some generic on the MS]
    all are fine. And the switch shouldn't be causing this.

    We see the same thing on the other side of the switch and
    with or without a bridgeing/bandwidth-measuring/bandwith-limiting
    devie in or out of the circuit.

    We have a few other things to try and then next step is a network
    yzer.
     
     

    When you said you tried two different hubs are they same
    manufacturer or different brands. Some hubs that are supposed to
    handle the speed translations between 10 and 100 haven't been know
    to have stellar performance.

    Let's hope we both get our problems solved.


    --
    Bill Vermillion - bv wjv . com
    Bill Guest

  7. #7

    Default Re: Slow network transfers

    Bill Vermillion wrote:
     
    Two different hubs, two different switches.
    DLink 16 port 10/100 switch - using this now
    Intel 8 port 10/100 switch
    Acer 8 port 10/100 Hub
    Accton 16 port 10 Hub - this one also had a connection to the 16 port
    DLink so may have confused the NIC into thinking full duplex.

    I might try a crossover cable too, just in case...
     
    Good luck and thanks :)

    --
    Scott Burns
    Mirrabooka Systems

    Tel +61 7 3857 7899
    Fax +61 7 3857 1368

    Scott Guest

  8. #8

    Default Re: Slow network transfers

    Scott Burns wrote: 
    > The machine I have available now is a P3/500 with 128Mb RAM, SCSI
    > HDDs, and generally me as the only user. My bash prompt turns red when
    > the load average hits 1, and it has been at least six months since that
    > happened. I've seen slow network transfers when working on client
    > machines, mostly of the same vintage, but with twice the RAM. We are
    > currently running 5.0.5. We almost always use Intel NICs. We used
    > NetGear for a few machines but found they were not up to scratch. The
    > curtrent NIC in question is an Intel PRO/100B or PRO/100+ with autosense
    > speed/duplex.
    >
    > The short llistat version:
    >
    > saturn:~# llistat -l > llstat.1
    > saturn:~# l test.img
    > -rw-r--r-- 1 root other 11798370 Nov 26 10:14 test.img
    > saturn:~# time nc -w 1 pluto 9900 < test.img
    >
    > real 4m34.680s
    > user 0m0.000s
    > sys 0m0.720s
    > saturn:~# llistat -l > llstat.2
    > saturn:~# diff llstat.1 llstat.2
    > 12,13c12,13
    > < In: 176608 0 1011 0 51455461 0
    > < Out: 176649 0 345 0 36680514 0
    > --- 
    > 16c16
    > < 186 frames received destined for an unbound SAP
    > --- 
    > 23c23
    > < Last send time: 591544
    > --- 
    > saturn:~# uptime
    > 10:21am up 1:45, 9 users, load average: 0.00, 0.00, 0.00
    > saturn:~#
    >
    > The long llistat version (llistat.2)
    >
    > Device MAC address in use Factory MAC Address
    > ------ ------------------ -------------------
    > /dev/net0 00:90:27:b2:5a:6e 00:90:27:b2:5a:6e
    >
    > Multicast address table
    > -----------------------
    > 01:00:5e:00:00:01
    >
    > FRAMES
    > Unicast Multicast Broadcast Error Octets Queue Length
    > ---------- --------- --------- ------ ----------- ------------
    > In: 186061 0 1055 0 52056589 0
    > Out: 188207 0 365 0 50732683 0
    >
    > DLPI Module Info: 2 SAPs open, 18 SAPs maximum
    > 191 frames received destined for an unbound SAP
    >
    > MAC Driver Info: Media_type: Ethernet
    > Min_SDU: 1, Max_SDU: 1514, Address length: 6
    > Interface speed: 100 Mbits/sec
    >
    > DLPI Restarts Info: Last queue size: 0
    > Last send time: 628854
    > Restart in progress: 0
    > Number of restarts: 0
    >
    > Interface Version: MDI 100
    >
    > ETHERNET SPECIFIC STATISTICS
    >
    > Collision Table - The number of frames successfully transmitted,
    > but involved in at least one collision:
    >
    > Frames Frames
    > ------- -------
    > 1 collision 0 9 collisions 0
    > 2 collisions 0 10 collisions 0
    > 3 collisions 0 11 collisions 0
    > 4 collisions 0 12 collisions 0
    > 5 collisions 0 13 collisions 0
    > 6 collisions 0 14 collisions 0
    > 7 collisions 0 15 collisions 0
    > 8 collisions 0 16 collisions 0
    >
    >
    > Bad Alignment 0 Number of frames received that were
    > not an integral number of octets
    >
    > FCS Errors 0 Number of frames received that did
    > not pass the Frame Check Sequence
    >
    > SQE Test Errors 0 Number of Signal Quality Error Test
    > signals that were detected by the
    > adapter
    >
    > Deferred Transmissions 0 Number of frames delayed on the
    > first transmission attempt because
    > the media was busy
    >
    > Late Collisions 0 Number of times a collision was
    > detected later than 512 bits into
    > the transmitted frame
    >
    > Excessive Collisions 0 Number of frames dropped on
    > transmission
    > because of excessive collisions
    >
    > Internal MAC Transmit 0 Number of frames dropped on
    > transmission
    > Errors because of errors not covered above
    >
    > Carrier Sense Errors 0 Number of times that the carrier sense
    > condition was lost when attempting to
    > send a frame that was deferred for an
    > excessive amount of time
    >
    > Frame Too Long 0 Number of frames dropped on reception
    > because they were larger than the
    > maximum Ethernet frame size
    >
    > Internal MAC Receive 0 Number of frames dropped on reception
    > Errors because of errors not covered above
    >
    > Spurious Interrupts 0 Number of times the adapter interrupted
    > the system for an unknown reason
    >
    > No STREAMS Buffers 0 Number of frames dropped on reception
    > because no STREAMS buffers were
    > available
    >
    > Underruns/Overruns 7 Number of times the transfer of
    > data to or from the frame buffer
    > did not complete successfully
    >
    > Device Timeouts 0 Number of times the adapter failed to
    > respond to a request from the driver
    >
    > --
    > Scott Burns
    > Mirrabooka Systems
    >
    > Tel +61 7 3857 7899
    > Fax +61 7 3857 1368[/ref]

    The Underruns/Overruns at 7 looks wrong, but not likely the cause of your problem.
    Download tcpspeed ( a windows test tool ) and run it against the 5.0.5 server.
    You can find it at:
    http://maximized.com/freeware/tcpspeed
    It uses the chargen port on the server, which should test the network stack and
    NIC card. Use a fairly powerfull windows box as the client.

    Mike

    --
    Michael Brown

    The Kingsway Group
    Mike Guest

Similar Threads

  1. Slow Network Logon on NT 4.0 Network
    By Robert Vorwaller in forum Windows Networking
    Replies: 1
    Last Post: July 9th, 07:08 PM

Bookmarks

Posting Permissions

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

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139