x2100 crappity crap - Solaris

This is a discussion on x2100 crappity crap - Solaris ; Thought I'd get a bargain with the x2100. But first I run into the SMDC/SP woes; mainly, console redirection does not work -- once Solaris takes over redirect stops. ok whatever, I can use a terminal server like the old ...

+ Reply to Thread
Results 1 to 15 of 15

Thread: x2100 crappity crap

  1. x2100 crappity crap

    Thought I'd get a bargain with the x2100. But first I run into the
    SMDC/SP woes; mainly, console redirection does not work -- once
    Solaris takes over redirect stops. ok whatever, I can use a terminal
    server like the old days.

    Then, I find that if I pull a hard drive, the system locks up for a
    few minutes. wtf? Then, if I return the hard drive, it isn't
    recognized! What kind of RAS is that??

    Oh, what about the stupid LED that doesn't go out? I guess the 4100
    has that problem also.

    And all the other issues, like USB keyboard going away, etc.

    I guess if Sun wants to sell these they need to have the BIOS (as
    opposed to OBP) but ugh! Would EFI be better?

    -frank

  2. Re: x2100 crappity crap

    Frank Cusack wrote:
    > Thought I'd get a bargain with the x2100. But first I run into the
    > SMDC/SP woes; mainly, console redirection does not work -- once
    > Solaris takes over redirect stops. ok whatever, I can use a terminal
    > server like the old days.
    >
    > Then, I find that if I pull a hard drive, the system locks up for a
    > few minutes. wtf? Then, if I return the hard drive, it isn't
    > recognized! What kind of RAS is that??
    >
    > Oh, what about the stupid LED that doesn't go out? I guess the 4100
    > has that problem also.
    >
    > And all the other issues, like USB keyboard going away, etc.
    >
    > I guess if Sun wants to sell these they need to have the BIOS (as
    > opposed to OBP) but ugh! Would EFI be better?
    >
    > -frank


    Yeah, sun hardware is pretty junior.

  3. Re: x2100 crappity crap

    In article ,
    Frank Cusack wrote:
    >Thought I'd get a bargain with the x2100. But first I run into the
    >SMDC/SP woes; mainly, console redirection does not work -- once
    >Solaris takes over redirect stops. ok whatever, I can use a terminal
    >server like the old days.


    Did you configure the serial port as Solaris console?
    eeprom output-device=ttya
    eeprom input-device=ttya
    eeprom console=ttya

    >Then, I find that if I pull a hard drive, the system locks up for a
    >few minutes. wtf? Then, if I return the hard drive, it isn't
    >recognized! What kind of RAS is that??


    Are these mirrored disks? Looks like a bug in the RAID 1 controller.

    >Oh, what about the stupid LED that doesn't go out? I guess the 4100
    >has that problem also.
    >
    >And all the other issues, like USB keyboard going away, etc.


    No SRDBs for the x2100 on SunSolve yet.

    >I guess if Sun wants to sell these they need to have the BIOS (as
    >opposed to OBP) but ugh! Would EFI be better?


    A selling point for these systems is the investment protection.
    If Solaris doesn't work out, then you can swap in Windows Server
    or RHAS. I don't think Microsoft or Redhat have committed to
    EFI.

    John
    groenveld@acm.org

  4. Re: x2100 crappity crap

    * John D Groenveld:

    > I don't think Microsoft or Redhat have committed to
    > EFI.


    Not on x86/x64 yet. But they also have Itanium versions of their
    Enterprise software, and Itanium uses EFI...

    Benjamin

  5. Re: x2100 crappity crap

    On Thu, 27 Apr 2006 14:39:14 +0000, John D Groenveld wrote:

    > In article ,
    > Frank Cusack wrote:
    >>Thought I'd get a bargain with the x2100. But first I run into the
    >>SMDC/SP woes; mainly, console redirection does not work -- once
    >>Solaris takes over redirect stops. ok whatever, I can use a terminal
    >>server like the old days.

    >
    > Did you configure the serial port as Solaris console?
    > eeprom output-device=ttya
    > eeprom input-device=ttya
    > eeprom console=ttya
    >
    >>Then, I find that if I pull a hard drive, the system locks up for a
    >>few minutes. wtf? Then, if I return the hard drive, it isn't
    >>recognized! What kind of RAS is that??

    >
    > Are these mirrored disks? Looks like a bug in the RAID 1 controller.
    >


    No HW RAID on the X2100 (last time I looked), so may be an SVM issue

    >>Oh, what about the stupid LED that doesn't go out? I guess the 4100
    >>has that problem also.
    >>
    >>And all the other issues, like USB keyboard going away, etc.

    >
    > No SRDBs for the x2100 on SunSolve yet.
    >
    >>I guess if Sun wants to sell these they need to have the BIOS (as
    >>opposed to OBP) but ugh! Would EFI be better?

    >
    > A selling point for these systems is the investment protection.
    > If Solaris doesn't work out, then you can swap in Windows Server
    > or RHAS. I don't think Microsoft or Redhat have committed to
    > EFI.
    >
    > John
    > groenveld@acm.org


    --
    Bruce

    "The internet is a huge and diverse community and
    not every one is friendly"
    http://www.ytc1.co.uk


  6. Re: x2100 crappity crap

    On Thu, 27 Apr 2006 14:39:14 +0000 (UTC) groenvel@cse.psu.edu (John D Groenveld) wrote:
    > In article ,
    > Frank Cusack wrote:
    >>Thought I'd get a bargain with the x2100. But first I run into the
    >>SMDC/SP woes; mainly, console redirection does not work -- once
    >>Solaris takes over redirect stops. ok whatever, I can use a terminal
    >>server like the old days.

    >
    > Did you configure the serial port as Solaris console?
    > eeprom output-device=ttya
    > eeprom input-device=ttya
    > eeprom console=ttya


    No. I was trying to use the video->SMDC console redirection, to
    avoid having to use the serial port. This is documented as not
    working. It does work for the bios, but stops when Solaris boots.

    Of course (?) if I set Solaris to use the serial console, and use
    video->serial console redirection, then it works. But then I need
    a terminal server.

    I think the breakage is due to Solaris resetting bge0, which is
    shared with the SP.

    >>Then, I find that if I pull a hard drive, the system locks up for a
    >>few minutes. wtf? Then, if I return the hard drive, it isn't
    >>recognized! What kind of RAS is that??

    >
    > Are these mirrored disks? Looks like a bug in the RAID 1 controller.


    SVM

    Can you verify that this works for you?

    >>Oh, what about the stupid LED that doesn't go out? I guess the 4100
    >>has that problem also.
    >>
    >>And all the other issues, like USB keyboard going away, etc.

    >
    > No SRDBs for the x2100 on SunSolve yet.


    The release notes (819-3722-13) document these issues. Ah ... now
    I see the SATA issue there, hadn't noticed it before. !#@%@!%
    Never would have bought the thing in the first place if I'd read
    that more carefully.

    Any insiders have any ideas on when/if these issues might get fixed?
    It's a good server (at the right price), it's just that a lot of
    hardware RAS features don't work under Solaris! blah.

    -frank

  7. Re: x2100 crappity crap

    Bruce Porter wrote:
    > On Thu, 27 Apr 2006 14:39:14 +0000, John D Groenveld wrote:
    >> In article ,
    >> Frank Cusack wrote:
    >>>Then, I find that if I pull a hard drive, the system locks up for a
    >>>few minutes. wtf? Then, if I return the hard drive, it isn't
    >>>recognized! What kind of RAS is that??

    >>
    >> Are these mirrored disks? Looks like a bug in the RAID 1 controller.
    >>

    >
    > No HW RAID on the X2100 (last time I looked), so may be an SVM issue


    No, it is more a problem of the IDE driver which has quite large timeout
    values in case of disk errors. I have seen it several times on my machine
    with a dying disk drive: Even though the other disks (all ok) are on another
    controller they will get affected also:

    This is the failing disk:

    15:02:03 ata_command: select failed DRDY 0x0 CMD 0x91 F 0x0 N 0x3f S 0x0 H 0xe CL 0x0 CH 0x0
    15:02:03 ata_init_dev_parms: failed
    15:02:03 WARNING: /pci@0,0/pci-ide@6/ide@0 (ata0):
    15:02:03 timeout: early timeout, target=0 lun=0
    15:02:03 WARNING: /pci@0,0/pci-ide@6/ide@0 (ata0):
    15:02:03 timeout: early timeout, target=1 lun=0
    15:02:03 WARNING: /pci@0,0/pci-ide@6/ide@0/cmdk@0,0 (Disk2):
    15:02:03 Error for command 'write sector' Error Level: Informational
    15:02:03 Sense Key: aborted command
    15:02:03 Vendor 'Gen-ATA ' error code: 0x3
    15:02:03 WARNING: md: d101: write error on /dev/dsk/c0d0s6

    Trying to revive it in the evening...

    22:13:22 WARNING: /pci@0,0/pci-ide@6/ide@0 (ata0):
    22:13:22 timeout: early timeout, target=0 lun=0
    22:13:22 WARNING: /pci@0,0/pci-ide@6/ide@0/cmdk@0,0 (Disk2):
    22:13:22 Error for command 'read sector' Error Level: Informational
    22:13:22 Sense Key: aborted command
    22:13:22 Vendor 'Gen-ATA ' error code: 0x3

    .... caused a timeout on the second disk on the same bus ...

    22:13:57 WARNING: /pci@0,0/pci-ide@6/ide@0 (ata0):
    22:13:57 timeout: abort request, target=1 lun=0
    22:13:57 WARNING: /pci@0,0/pci-ide@6/ide@0 (ata0):
    22:13:57 timeout: abort device, target=1 lun=0
    22:13:57 WARNING: /pci@0,0/pci-ide@6/ide@0 (ata0):
    22:13:57 timeout: reset target, target=1 lun=0
    22:13:57 WARNING: /pci@0,0/pci-ide@6/ide@0 (ata0):
    22:13:57 timeout: reset bus, target=1 lun=0
    22:14:35 ata_command: select failed DRDY 0x0 CMD 0x91 F 0x0 N 0x3f S 0x0 H 0xe CL 0x0 CH 0x0
    22:14:35 ata_init_dev_parms: failed
    22:14:35 WARNING: /pci@0,0/pci-ide@6/ide@0/cmdk@1,0 (Disk3):
    22:14:35 Error for command 'read sector' Error Level: Informational
    22:14:35 Sense Key: aborted command
    22:14:35 Vendor 'Gen-ATA ' error code: 0x3

    .... and also timeouts on another unrelated IDE controller (the one above was
    PATA, this one is SATA):

    22:14:35 WARNING: /pci@0,0/pci-ide@7/ide@0 (ata2):
    22:14:35 timeout: early timeout, target=0 lun=0
    22:14:35 WARNING: /pci@0,0/pci-ide@7/ide@0/cmdk@0,0 (Disk4):
    22:14:35 Error for command 'read sector' Error Level: Informational
    22:14:35 Sense Key: aborted command
    22:14:35 Vendor 'Gen-ATA ' error code: 0x3
    22:14:35 WARNING: md: d11: read error on /dev/dsk/c2d0s0

    .... and also all of my other disks.

    Result: All my md's were in state Maintenance, both mirror halves either
    in Maintenance or Last Erred, I had to run a lot of "metareplace -e
    dXX cXdYsZ" - which takes some time on 400 GB mirrored (~5 hours to bring
    the mirror back to state Ok)


    It should get better in the future with a native SATA driver.


    --
    Daniel

  8. Re: x2100 crappity crap

    * Andrew Gabriel:

    >>> I don't think Microsoft or Redhat have committed to
    >>> EFI.

    >> Not on x86/x64 yet. But they also have Itanium versions of their
    >> Enterprise software, and Itanium uses EFI...

    >
    > Intel's next generation of its Centrino notebook platform
    > due early next year is rumoured to be EFI.


    I doubt it because MS clearly stated that Vista won't support EFI, and
    in Longhorn server its only the IA64 variant that supports EFI.

    > Apple is already
    > using it.


    Apple isn't selling Windows but MacOS X which of course has no problems
    with EFI. And it took some real effort to get Windowsxp running on the
    intel Macs (one of this effort was retrofitting a BIOS emulation in the
    system firmware)...

    Benjamin

  9. Re: x2100 crappity crap

    Frank Cusack writes:

    >No. I was trying to use the video->SMDC console redirection, to
    >avoid having to use the serial port. This is documented as not
    >working. It does work for the bios, but stops when Solaris boots.


    >I think the breakage is due to Solaris resetting bge0, which is
    >shared with the SP.


    Correct; known issue.

    >Any insiders have any ideas on when/if these issues might get fixed?
    >It's a good server (at the right price), it's just that a lot of
    >hardware RAS features don't work under Solaris! blah.


    We're working hard to fix all those issues.

    But the bleeding edge were fixes happen first is Solaris Nevada.

    E.g., the sideband NIC bug was fixed in Nevada around 3 months ago;
    the bug id is 6337341.

    But it's also patched in 122028-02. Have you tried that patch?



    Casper
    --
    Expressed in this posting are my opinions. They are in no way related
    to opinions held by my employer, Sun Microsystems.
    Statements on Sun products included here are not gospel and may
    be fiction rather than truth.

  10. Re: x2100 crappity crap

    On Fri, 28 Apr 2006 11:41:08 +0200, Benjamin Gawert wrote:

    > * Andrew Gabriel:
    >



    >> Apple is already
    >> using it.

    >
    > Apple isn't selling Windows but MacOS X which of course has no problems
    > with EFI. And it took some real effort to get Windowsxp running on the
    > intel Macs (one of this effort was retrofitting a BIOS emulation in the
    > system firmware)...
    >


    I gather running any OS other than MacOS on an Intel based Apple is
    difficult.
    Because "They made it so"

    --
    Bruce

    "The internet is a huge and diverse community and
    not every one is friendly"
    http://www.ytc1.co.uk


  11. Re: x2100 crappity crap

    On 28 Apr 2006 13:49:35 GMT Casper H.S. Dik wrote:
    > We're working hard to fix all those issues.
    >
    > But the bleeding edge were fixes happen first is Solaris Nevada.


    huh. Well, I'm not going to use this hardware for anything now, so
    that gives me a good reason/way to check out solaris express. goody!

    > E.g., the sideband NIC bug was fixed in Nevada around 3 months ago;
    > the bug id is 6337341.
    >
    > But it's also patched in 122028-02. Have you tried that patch?


    Thanks, i'll give it a whirl before going to nevada.

    -frank

  12. Re: x2100 crappity crap

    On Fri, 28 Apr 2006 11:41:08 +0200 Benjamin Gawert wrote:
    > Apple isn't selling Windows but MacOS X which of course has no
    > problems with EFI. And it took some real effort to get Windowsxp
    > running on the intel Macs (one of this effort was retrofitting a BIOS
    > emulation in the system firmware)...


    Huh, well if Apple can do EFI and write a BIOS emulation then so can Sun!
    Or maybe Sun can do an OF/OBP thing and add a BIOS emulation layer. It
    sure would enhance the x86 hardware offering.

    -frank

  13. Re: x2100 crappity crap

    On Fri, 28 Apr 2006, Frank Cusack wrote:

    > Huh, well if Apple can do EFI and write a BIOS emulation then so can Sun!
    > Or maybe Sun can do an OF/OBP thing and add a BIOS emulation layer. It
    > sure would enhance the x86 hardware offering.


    Agreed. Having to deal with the bletcherousness of the BIOS after being
    used to OBP is a bit of a shock. Scotty's expression in Star Trek IV
    comes to mind: "BIOS? How quaint!".

    --
    Rich Teer, SCNA, SCSA, OpenSolaris CAB member

    President,
    Rite Online Inc.

    Voice: +1 (250) 979-1638
    URL: http://www.rite-group.com/rich

  14. Re: x2100 crappity crap

    * Bruce Porter:

    > I gather running any OS other than MacOS on an Intel based Apple is
    > difficult.
    > Because "They made it so"


    No. It's just that Apple decided not to rely on the ancient BIOS but on
    EFI which is much more flexible and versatile. The problem is that not
    much operating systems support EFI _and_ x86 at the same time...

    Benjamin

  15. Re: x2100 crappity crap

    On Thu, 27 Apr 2006 19:06:11 +0000 (UTC) "Daniel Rock" wrote:
    > It should get better in the future with a native SATA driver.


    So does S10_U2 not have a native SATA driver? I just installed and tested
    and it has the same problems (with the same syslog messages) as before.

    From Casper's post, I'd expected U2 to work a lot better.

    -frank

+ Reply to Thread