suse 10.3 slow disk maybe no DMA - Suse

This is a discussion on suse 10.3 slow disk maybe no DMA - Suse ; Hi, I installed SUSE 10.3 on nforce4 board asus a8n premium. It has 64-bit AMD X2 3800+ processor. It has 4 GB memory and ATI x600 card. First I was surprised that disks becomes /dev/sdX something instead /dev/hdb, which is ...

+ Reply to Thread
Results 1 to 3 of 3

Thread: suse 10.3 slow disk maybe no DMA

  1. suse 10.3 slow disk maybe no DMA

    Hi,

    I installed SUSE 10.3 on nforce4 board asus a8n premium. It has 64-bit
    AMD X2 3800+ processor. It has 4 GB memory and ATI x600 card.

    First I was surprised that disks becomes /dev/sdX something instead
    /dev/hdb, which is my Linux-disk. It is seagate PATA 250GB disk. Hdparm
    does not work for it

    hdparm /dev/sdb

    /dev/sdb:
    IO_support = 0 (default 16-bit)
    HDIO_GET_UNMASKINTR failed: Inappropriate ioctl for device
    HDIO_GET_DMA failed: Inappropriate ioctl for device
    HDIO_GET_KEEPSETTINGS failed: Inappropriate ioctl for device
    readonly = 0 (off)
    readahead = 256 (on)
    geometry = 30401/255/63, sectors = 488397168, start = 0

    this probably because there is now different driver for ata. How can I
    disable it ? I was thinking to put kernel 2.6.18 vanilla to it but I am
    not sure what all changes I have to make to other files. fstab surely
    and then menu.lst in /boot/grub. What else ? what about the modules.conf?

    Speed test gives this
    hdparm -tT /dev/sdb

    /dev/sdb:
    Timing cached reads: 1330 MB in 2.00 seconds = 665.20 MB/sec
    Timing buffered disk reads: 88 MB in 3.00 seconds = 29.31 MB/sec

    I think these are little on slow side - or what ?

    I run VMWare inside this machine and those virtual machines win2003
    server and winxp have become unusable because the disk trashes all the
    time - in other words it is very slow and light is burning whole time.

    Please someone who knows help me out with this. I would like to use suse
    10.3 but I might go back to 10.2 if that's the only option.

    Also what is the role of sdparm which have been commented to be used
    instead of hdparm? I did not find anythin to do with DMA in it.

    Also I noticed that when copying cd to hard disk with command

    dd if=/dev/sr0 of=/root/disk_image.iso

    I only get about 1GB of the whole disk which contains about 3.7GB (suse
    10.3 DVD). I haven't tested this very much yet.

    Best Regards
    Kari




  2. Re: suse 10.3 slow disk maybe no DMA

    Some being by the name of Kari Laine managed to piece together the
    following:

    > /dev/sdb:
    > Timing cached reads: * 1330 MB in *2.00 seconds = 665.20 MB/sec
    > Timing buffered disk reads: * 88 MB in *3.00 seconds = *29.31 MB/sec
    >


    Mine which are scsi (RAID 5) give me the following:

    /dev/sda:
    Timing cached reads: 1580 MB in 2.00 seconds = 789.81 MB/sec
    Timing buffered disk reads: 108 MB in 3.07 seconds = 35.23 MB/sec

    So I don't think your disk is slow, certainly compared to mine, but then
    again my xSeries RAID controller is supposed to be slow - so they say. As
    for VMWare I have 4 VMWare machines running on there, but all running some
    flavour of Linux, and not windows. In saying that I have a pair of HT Xeons
    running in there as well. I don't think it's slow on any VM that's running,
    one of them is running Oracle 10GR2 and doing a very nice job of it too.

    Have you checked out things like iostats, sar, mpsar, vmstat and the like?
    Any errors in /var/log/messages which may point to something else?

    --
    I'm a dyslexic satanist, I worship the drivel... must be why I'm here

  3. Re: suse 10.3 slow disk maybe no DMA

    TGGA wrote:
    > /dev/sda:
    > Timing cached reads: 1580 MB in 2.00 seconds = 789.81 MB/sec
    > Timing buffered disk reads: 108 MB in 3.07 seconds = 35.23 MB/sec
    >
    > So I don't think your disk is slow, certainly compared to mine, but then
    > again my xSeries RAID controller is supposed to be slow - so they say. As
    > for VMWare I have 4 VMWare machines running on there, but all running some
    > flavour of Linux, and not windows. In saying that I have a pair of HT Xeons
    > running in there as well. I don't think it's slow on any VM that's running,
    > one of them is running Oracle 10GR2 and doing a very nice job of it too.

    thanks for prompt reply
    >
    > Have you checked out things like iostats, sar, mpsar, vmstat and the like?

    Well, I could use some quidance how the check these...


    > Any errors in /var/log/messages which may point to something else?

    I don't know why I did not look there first ...
    There are lot of errors concerning my disk at /dev/sdb like these
    Oct 28 22:51:16 asusamd smartd[4055]: Device: /dev/sdc, SMART Prefailure
    Attribute: 1 Raw_Read_Error_Rate changed from 102 to 103
    Oct 28 22:51:16 asusamd smartd[4055]: Device: /dev/sdd, SMART Prefailure
    Attribute: 1 Raw_Read_Error_Rate changed from 105 to 100
    Oct 28 22:51:16 asusamd smartd[4055]: Device: /dev/sdd, SMART Usage
    Attribute: 195 Hardware_ECC_Recovered changed from 63 to 64
    Oct 28 22:51:16 asusamd smartd[4055]: Device: /dev/sde, SMART Prefailure
    Attribute: 1 Raw_Read_Error_Rate changed from 107 to 109
    Oct 28 22:51:16 asusamd smartd[4055]: Device: /dev/sde, SMART Usage
    Attribute: 190 Temperature_Celsius changed from 47 to 48
    Oct 28 22:51:16 asusamd smartd[4055]: Device: /dev/sde, SMART Usage
    Attribute: 194 Temperature_Celsius changed from 53 to 52
    Oct 28 22:51:16 asusamd smartd[4055]: Device: /dev/sde, SMART Usage
    Attribute: 195 Hardware_ECC_Recovered changed from 62 to 61
    Oct 28 22:51:16 asusamd smartd[4055]: Device: /dev/sdf, SMART Prefailure
    Attribute: 1 Raw_Read_Error_Rate changed from 87 to 88
    Oct 28 22:51:16 asusamd smartd[4055]: Device: /dev/sdf, SMART Usage
    Attribute: 190 Temperature_Celsius changed from 65 to 66
    Oct 28 22:51:16 asusamd smartd[4055]: Device: /dev/sdf, SMART Usage
    Attribute: 194 Temperature_Celsius changed from 35 to 34
    Oct 28 22:51:16 asusamd smartd[4055]: Device: /dev/sdf, SMART Usage
    Attribute: 195 Hardware_ECC_Recovered changed from 63 to 62

    So I have to change the disk - bummer

    Thank you I get back to your guys after I change the disk.


    Best Regards
    Kari




+ Reply to Thread