Promise 20276 FastTrak IDE RAID1 Hot Swap - Linux

This is a discussion on Promise 20276 FastTrak IDE RAID1 Hot Swap - Linux ; Hi, Has anybody had any experience with handling a server with a Promise FastTrak IDE RAID1 Hot Swap configuration e.g. 20276? I have finally succeeded in creating a bootable two disk RAID1 array with Redhat 9 that is fault tolerant. ...

+ Reply to Thread
Results 1 to 3 of 3

Thread: Promise 20276 FastTrak IDE RAID1 Hot Swap

  1. Promise 20276 FastTrak IDE RAID1 Hot Swap

    Hi,

    Has anybody had any experience with handling a server with a Promise
    FastTrak IDE RAID1 Hot Swap configuration e.g. 20276?

    I have finally succeeded in creating a bootable two disk RAID1 array
    with Redhat 9 that is fault tolerant.

    But I am challenged by a couple of issues that I hope can be resolved
    somehow:

    1) Re-building an array after a typical disk failure appears to
    require a system shutdown and a 70 minute operation under BIOS control
    for two 120GByte disks.

    2) I cannot find a trace in any log files for the following console
    message that is printed in case of a single drive failure:

    logFastTrak| 1|4123|Error|Disk Offline: Maxtor 4R120L0 (Channel
    1/Master)

    In this scenario hot swapping is not possible even if the harware
    supports IDE hot swapping fully.


    For Windows, Promise supplies a program called "FastCheck Monitoring
    Utility" which appears to handle 1) via a rebuild function and 2) via
    log file configuration.

    Are there any Linux distributions that support this harware (Redhat
    doesn't)? In that case I would be prepared to switch.


    Many thanks,
    Bernard



  2. Re: Promise 20276 FastTrak IDE RAID1 Hot Swap

    Bernard wrote:

    > Hi,
    >
    > Has anybody had any experience with handling a server with a Promise
    > FastTrak IDE RAID1 Hot Swap configuration e.g. 20276?


    I've dealt with a stack of Promise chipsets. I usually find it faster
    and overall cheaper to replace/exchange it with hardware based around
    the Siig or 3Ware controllers: Promise sucks really hard, especially for
    Linux support (which they claim but do not provide) and for actual
    hardware reliability.


    > 2) I cannot find a trace in any log files for the following console
    > message that is printed in case of a single drive failure:
    >
    > logFastTrak| 1|4123|Error|Disk Offline: Maxtor 4R120L0 (Channel
    > 1/Master)


    What kernel or modified kernel are you using?

    > For Windows, Promise supplies a program called "FastCheck Monitoring
    > Utility" which appears to handle 1) via a rebuild function and 2) via
    > log file configuration.
    >
    > Are there any Linux distributions that support this harware (Redhat
    > doesn't)? In that case I would be prepared to switch.


    Promise doesn't publish up-to-date drivers, or even admit that there are
    chip numbers associated with their different controllers. Don't use them.


  3. Re: Promise 20276 FastTrak IDE RAID1 Hot Swap

    Hi Nico,

    Thanks for your advice. Buying one of the cards that you are
    suggesting is probably the next thing I will do.

    What I found with the Promise 20276 is that when I pull out one drive
    where both are perfectly in sync i.e. the array is not "critical",
    then the system automatically and instantly boots into the Promise
    BIOS, with a big warning screen.

    I must admit that this is not the typical failure condition, because
    in effect there was no failure before I pulled out the drive, but one
    might want to use the RAID feature to replace a drive for perfectly
    valid reasons, e.g. preventative maintenance, backup, or disk upgrade.

    I should say that I DO have hot-swappable IDE hardware interfaces
    between the drives and the controller.

    My kernel is a Redhad 9 Linux kernel 2.4.20-8

    I vaguely remember that somewhere logging can be configured to include
    messages from various programs. I did not configure anything so that
    logFastTrak messages would be included.

    A burning question: How would you non-destructively simulate a
    harddisk error so that you can verify that all your proccesses are in
    place for an actual recovery situation?

    Thanks again,
    Bernard

    Nico Kadel-Garcia wrote:

    >Bernard wrote:
    >
    >> Hi,
    >>
    >> Has anybody had any experience with handling a server with a Promise
    >> FastTrak IDE RAID1 Hot Swap configuration e.g. 20276?

    >
    >I've dealt with a stack of Promise chipsets. I usually find it faster
    >and overall cheaper to replace/exchange it with hardware based around
    >the Siig or 3Ware controllers: Promise sucks really hard, especially for
    >Linux support (which they claim but do not provide) and for actual
    >hardware reliability.
    >
    >
    >> 2) I cannot find a trace in any log files for the following console
    >> message that is printed in case of a single drive failure:
    >>
    >> logFastTrak| 1|4123|Error|Disk Offline: Maxtor 4R120L0 (Channel
    >> 1/Master)

    >
    >What kernel or modified kernel are you using?
    >
    >> For Windows, Promise supplies a program called "FastCheck Monitoring
    >> Utility" which appears to handle 1) via a rebuild function and 2) via
    >> log file configuration.
    >>
    >> Are there any Linux distributions that support this harware (Redhat
    >> doesn't)? In that case I would be prepared to switch.

    >
    >Promise doesn't publish up-to-date drivers, or even admit that there are
    >chip numbers associated with their different controllers. Don't use them.



+ Reply to Thread