1. This forum section is a read-only archive which contains old newsgroup posts. If you wish to post a query, please do so in one of our main forum sections (here). This way you will get a faster, better response from the members on Motherboard Point.

[V440] Firmware update, OS choice and ALOM password

Discussion in 'Sun Hardware' started by Bubba, Nov 6, 2011.

  1. Bubba

    Bubba Guest

    Greetings,

    I got V440 and have several questions:

    - as far as I saw, the easiest and most flexible OS would still be
    Solaris, since both OpenBSD and Debian lacked some functionality. Both
    Debian and OBSD had problems with partitioning and boot options, while
    Debian couldn't recognize two additional Quad Gigabit cards and required
    drivers that were not available. Thus, I figured it would be safest to try
    with Solaris - as machine has 2x 1.2 GHz Sparcs and 4GB RAM, would latest
    Solaris 10 work fine (I don't require GUI; compiler sets and web/DB
    related applications will suffice, and I am aware they are available and
    work properly (but not necessarily to fast) on Slowaris. :)

    - the machine was taken from a financial institution and lacked any DASD,
    but I also didn't get any administration documentation. IOW - when I bump
    into ALOM's login part, I have no clue how to log in. Apparently, while
    searching through the Net, "admin" should be username and last 8 *digits*
    of V440's serial number would make a password. Although I assume someone
    tampered with those credentials, my serial number is alphanumeric, so this
    "digit" part largely makes me doubt in that information. Nevertheless, I
    also read it is possible to reset ALOM credentials via Solaris. Since I
    can boot (>ok prompt appears and allows booting from cdrom and disks), is
    this doable?

    - lastly, my ROM version is OpenBoot 4.22.33 and there seems to be an
    upgrade. It also seems Oracle has put much effort to make my life
    miserable and provided numerous steps in order to obtain one with
    requirements I don't understand how to comply to. They require "The
    Customer Support Identifier (or CSI) is the identifier used for the
    support contract for between you and Oracle Support. You may request
    access to one or more, typically you enter a CSI of about 8 digits in
    length." and filling out contact forms... All in all, I wonder, since I
    cannot find change log and would like to upgrade machine with two
    additional CPU's with 8GB RAM, is it worth additional trouble
    and exploration? What is the procedure of upgrading OpenBoot anyway? Can
    it be obtained via 3rd party, in terms of being able to download it
    somewhere else?

    TIA!


    --
    "If you lie to the compiler,
    it will get its revenge."
    Henry Spencer
    2.718281828459045235360287471352662497757247093699959574966967627.com
     
    1. Advertising

  2. Bubba <> writes:
    >I got V440 and have several questions:


    >- as far as I saw, the easiest and most flexible OS would still be
    >Solaris, since both OpenBSD and Debian lacked some functionality...


    Definately. Those other two choices stopped development altogether for
    some point in time. I don't know know if they are currently active or not.
    But yes, Solaris is definately the only thing that would work well on
    that hardware.

    >work properly (but not necessarily to fast) on Slowaris. :)


    Slowaris is what Linux kids claimed until they had to implement all
    the normal locking and control of a modern kernel (instead of flying
    loose and carefree). I find Solaris to be much faster on *equivilent*
    hardware than Linux systems..

    >into ALOM's login part, I have no clue how to log in. Apparently, while
    >searching through the Net, "admin" should be username and last 8 *digits*
    >of V440's serial number would make a password.


    Hmm, I don't remember that being correct. But either way, just reset
    it once you have Solaris running. I wouldn't put too much work into
    getting into ALOM now if you can boot cdrom. And then you just reset
    it later to use it.

    >.. read it is possible to reset ALOM credentials via Solaris. Since I
    >can boot (>ok prompt appears and allows booting from cdrom and disks), is
    >this doable?


    Sure sounds like it. Once I had a few used systems that couldn't boot
    dvd/cdrom and then I had to move a preinstalled system harddisk over
    from another of the same class and get in that way..


    >- lastly, my ROM version is OpenBoot 4.22.33 and there seems to be an
    >upgrade. It also seems Oracle has put much effort to make my life
    >miserable and provided numerous steps in order to obtain one with
    >requirements I don't understand how to comply to.


    Its fairly simple. Get the system up on maintenance for something
    starting around $1200/yr per socket.

    Since I'm guessing that it isn't worth that for you, stay with the
    OpenBoot you have. OpenBoot upgrades are boring, the fixes are
    normally pretty esoteric. The only "feature" I've seen in fairly
    recent Openboot is WANBoot, but I don't know if yours is new enough or
    not. Its not like you spend more than 0.00001% of the time ever messing
    in OpenBoot. Its about as exciting as updating your boot blocks.

    The procedure to do it normally is just installing a patch in Solaris.
    Somebody may or may not have the specific patch you need online, but I
    wouldn't bother searching for it.
     
    1. Advertising

  3. Bubba

    Bubba Guest

    Doug McIntyre's log on stardate 06 stu 2011

    >>- as far as I saw, the easiest and most flexible OS would still be
    >>Solaris, since both OpenBSD and Debian lacked some functionality...

    > Definately. Those other two choices stopped development altogether
    > for some point in time. I don't know know if they are currently
    > active or not. But yes, Solaris is definately the only thing that
    > would work well on that hardware.


    I notticed that, on kinda harder way. :)

    >>work properly (but not necessarily to fast) on Slowaris. :)

    > Slowaris is what Linux kids claimed until they had to implement all
    > the normal locking and control of a modern kernel (instead of flying
    > loose and carefree). I find Solaris to be much faster on *equivilent*
    > hardware than Linux systems..


    Is that correct, for example, in respect of gcc, web performance and
    simmilar things?

    >>into ALOM's login part, I have no clue how to log in. Apparently,
    >>while searching through the Net, "admin" should be username and last
    >>8 *digits* of V440's serial number would make a password.

    > Hmm, I don't remember that being correct. But either way, just reset
    > it once you have Solaris running. I wouldn't put too much work into
    > getting into ALOM now if you can boot cdrom. And then you just reset
    > it later to use it.


    I was afraid there might be something important in ALOM that would
    prevent me from accessing OS or something like that.

    As it appears, I should be good to go to fully functional Solaris.

    One more thing - there are two drives that came with machine, loaded
    with Solaris. Now when I star the installation procedure, I am offered
    an upgrade or "initialization" (I might be wrong, but something like
    that). Since I only skimmed through the instalation procedure and
    haven't reall started it yet, is it possible to wipe/zero-fill/LLF SCSI
    drives from the console, just to avoid any potential fuss with the
    instalation?

    > Sure sounds like it. Once I had a few used systems that couldn't boot
    > dvd/cdrom and then I had to move a preinstalled system harddisk over
    > from another of the same class and get in that way..


    Since machine came from financial institution, I was affraid it might
    have some thorougher security roles. Lucky me, however, I was wrong...
    :)

    >>- lastly, my ROM version is OpenBoot 4.22.33 and there seems to be an
    >>upgrade. It also seems Oracle has put much effort to make my life
    >>miserable and provided numerous steps in order to obtain one with
    >>requirements I don't understand how to comply to.

    > Its fairly simple. Get the system up on maintenance for something
    > starting around $1200/yr per socket.


    Hehe... ;)

    > Since I'm guessing that it isn't worth that for you, stay with the
    > OpenBoot you have. OpenBoot upgrades are boring, the fixes are
    > normally pretty esoteric. The only "feature" I've seen in fairly
    > recent Openboot is WANBoot, but I don't know if yours is new enough
    > or not. Its not like you spend more than 0.00001% of the time ever
    > messing in OpenBoot. Its about as exciting as updating your boot
    > blocks.


    My wories were regarding potential incompatibility with additional
    hardware (more RAM, new CPU cards, etc.); so, it's not worth it?

    > The procedure to do it normally is just installing a patch in
    > Solaris. Somebody may or may not have the specific patch you need
    > online, but I wouldn't bother searching for it.


    It's not like I haven't tried, but I'll just skip it now. :)

    Thank you for your feedback.

    BTW, JFTR, I measured power consumption of V440:

    (ATTN: may wrap)
    http://2.71828182845904523536028747...966967627.com/Sun-V440-Power-Consumption.html

    --
    "If you lie to the compiler,
    it will get its revenge."
    Henry Spencer
    2.718281828459045235360287471352662497757247093699959574966967627.com
     
  4. DoN. Nichols

    DoN. Nichols Guest

    On 2011-11-06, Bubba <> wrote:
    > Doug McIntyre's log on stardate 06 stu 2011


    [ ... ]

    >>>work properly (but not necessarily to fast) on Slowaris. :)

    >> Slowaris is what Linux kids claimed until they had to implement all
    >> the normal locking and control of a modern kernel (instead of flying
    >> loose and carefree). I find Solaris to be much faster on *equivilent*
    >> hardware than Linux systems..

    >
    > Is that correct, for example, in respect of gcc, web performance and
    > simmilar things?


    Well ... I tend to prefer the compiler in Sun's "Studio 12"
    development package -- if you can still get it along with Solaris 10
    since the Oracle takeover, but there are some programs which really need
    gcc to compile. My experience is that one takes longer to compile, but
    produces somewhat faster code. (And I forget which is which at the
    moment.)

    [ ... ]

    > I was afraid there might be something important in ALOM that would
    > prevent me from accessing OS or something like that.


    Only if you are trying to control it remotely (via ALOM). If
    the V440 is like the Sun Fire 280R, you can even pull out the card which
    implements ALOM or RSC (mine had RSC, not ALOM). However, the Sun Fire
    V120 has the LOM chip on the system board, so you can't do that, though
    you can change a jumper and power it up to flush the current settings
    (including passwords) in the LOM.

    Now, if someone had set the firmware password in the OBP, and
    set the security-mode to "command" or "full", you could not boot from
    the CD-ROM at all, and could only boot from a disk with Solaris
    installed. From there, you could reset the security-mode to "none" and
    clear the security-password, and then be free to do what you needed.

    > As it appears, I should be good to go to fully functional Solaris.
    >
    > One more thing - there are two drives that came with machine, loaded
    > with Solaris. Now when I star the installation procedure, I am offered
    > an upgrade or "initialization" (I might be wrong, but something like
    > that). Since I only skimmed through the instalation procedure and
    > haven't reall started it yet, is it possible to wipe/zero-fill/LLF SCSI
    > drives from the console, just to avoid any potential fuss with the
    > instalation?


    If you select "initial install" (I think that is the term, not
    "initialize"), you will newfs all the disk partitions, and if you
    re-partition (select the "custom" option when it gets to asking about
    disk partitions), that will pretty much force a newfs on each changed
    partition. But really, an "initial install" will clear away anything on
    the disk anyway -- though not as thoroughly as it would with the format
    "analyze" options set for one of the modes to "destroy existing data" --
    but I hope that you have plenty of time to twiddle your thumbs while it
    does this.

    If you can boot to Solaris and log in as root, then they proably
    did their own wipe of the disks and did a fresh install of Solaris
    anyway -- whatever version they had on hand. :)

    >> Sure sounds like it. Once I had a few used systems that couldn't boot
    >> dvd/cdrom and then I had to move a preinstalled system harddisk over
    >> from another of the same class and get in that way..

    >
    > Since machine came from financial institution, I was affraid it might
    > have some thorougher security roles. Lucky me, however, I was wrong...
    >:)


    Freshly scrubbed disks with a new install of Solaris -- I would
    *hope*. :) Maybe spare disks which they kept around fully loaded with
    the OS for emergencies, and which they swapped in to send away a "clean"
    system.

    [ ... ]

    >> Since I'm guessing that it isn't worth that for you, stay with the
    >> OpenBoot you have. OpenBoot upgrades are boring, the fixes are
    >> normally pretty esoteric. The only "feature" I've seen in fairly
    >> recent Openboot is WANBoot, but I don't know if yours is new enough
    >> or not. Its not like you spend more than 0.00001% of the time ever
    >> messing in OpenBoot. Its about as exciting as updating your boot
    >> blocks.

    >
    > My wories were regarding potential incompatibility with additional
    > hardware (more RAM, new CPU cards, etc.); so, it's not worth it?


    Usually the upgrades fix minor things with the hardware, and if
    the system has been running quite a while with no problems (in their
    service, not yours) it is probably new enough. I do like to upgrade if
    I can, but Oracle has made it too difficult to get the firmware patches
    these days.

    [ ... ]

    > BTW, JFTR, I measured power consumption of V440:
    >
    > (ATTN: may wrap)
    > http://2.71828182845904523536028747...966967627.com/Sun-V440-Power-Consumption.html


    O.K. Looks like UK power cords. Nice Fluke meters.

    Enjoy,
    DoN.

    --
    Remove oil spill source from e-mail
    Email: <> | Voice (all times): (703) 938-4564
    (too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
    --- Black Holes are where God is dividing by zero ---
     
  5. George

    George Guest

    On 11/06/2011 10:18 PM, DoN. Nichols wrote:
    > On 2011-11-06, Bubba<> wrote:
    >> Doug McIntyre's log on stardate 06 stu 2011

    >
    > [ ... ]
    >
    >>>> work properly (but not necessarily to fast) on Slowaris. :)
    >>> Slowaris is what Linux kids claimed until they had to implement all
    >>> the normal locking and control of a modern kernel (instead of flying
    >>> loose and carefree). I find Solaris to be much faster on *equivilent*
    >>> hardware than Linux systems..

    >>
    >> Is that correct, for example, in respect of gcc, web performance and
    >> simmilar things?

    >
    > Well ... I tend to prefer the compiler in Sun's "Studio 12"
    > development package -- if you can still get it along with Solaris 10
    > since the Oracle takeover, but there are some programs which really need
    > gcc to compile. My experience is that one takes longer to compile, but
    > produces somewhat faster code. (And I forget which is which at the
    > moment.)


    Yes, much code is chock full of Gnuisms that will only compile with gcc.

    Many moons ago, a delegation of Sun engineers gave a dog-and-pony show
    at my place of employment. They said that after developing dtrace, they
    looked at "Slowlaris", found the problems to be system system libraries,
    and worked hard at eliminating the latency. (This is Solaris 10, of
    course.)

    G
     
  6. George <> writes:

    >Yes, much code is chock full of Gnuisms that will only compile with gcc.


    The current crop of Stdio compilers accepts most of the GNUisms.

    Casper
    --
     
  7. Bubba

    Bubba Guest

    DoN. Nichols's log on stardate 07 stu 2011

    >> Is that correct, for example, in respect of gcc, web performance and
    >> simmilar things?

    > Well ... I tend to prefer the compiler in Sun's "Studio 12"
    > development package -- if you can still get it along with Solaris 10
    > since the Oracle takeover, but there are some programs which really
    > need gcc to compile. My experience is that one takes longer to
    > compile, but produces somewhat faster code. (And I forget which is
    > which at the moment.)


    Hehe, well, I'd sacrifice longer compile time (that is, after all, "one
    time job") for faster run time.

    Either way, I need CLI tool, so I believe "Studio 12" or whatever is IDE,
    so it is probably not worth installing just for compiler.

    In any case, I'll check it out, thanks.

    >> I was afraid there might be something important in ALOM that would
    >> prevent me from accessing OS or something like that.

    > Only if you are trying to control it remotely (via ALOM). If
    > the V440 is like the Sun Fire 280R, you can even pull out the card
    > which implements ALOM or RSC (mine had RSC, not ALOM). However, the
    > Sun Fire V120 has the LOM chip on the system board, so you can't do
    > that, though you can change a jumper and power it up to flush the
    > current settings (including passwords) in the LOM.


    I saw some jumpers, yes, and I thought they might be for something like
    clearing settings. However, I couldn't find any sane documentation
    regarding that card per se, so I'll have to give it a try a bit more...

    > If you select "initial install" (I think that is the term, not
    > "initialize"), you will newfs all the disk partitions, and if you
    > re-partition (select the "custom" option when it gets to asking about
    > disk partitions), that will pretty much force a newfs on each changed
    > partition. But really, an "initial install" will clear away anything
    > on the disk anyway -- though not as thoroughly as it would with the
    > format "analyze" options set for one of the modes to "destroy
    > existing data" -- but I hope that you have plenty of time to twiddle
    > your thumbs while it does this.


    Oh, that's great news.

    Well, I'll just keep it rollin' during the night...

    > Freshly scrubbed disks with a new install of Solaris -- I would
    > *hope*. :) Maybe spare disks which they kept around fully loaded
    > with the OS for emergencies, and which they swapped in to send away a
    > "clean" system.


    I don't really care, I just want to wipe them clean, even if they had
    human readable text files with PIN's. :)

    > O.K. Looks like UK power cords. Nice Fluke meters.


    Nope, two times. :) It's Croatia (Schuko -
    http://en.wikipedia.org/wiki/Schuko) and ammeter's noware near Fluke :D;
    but for el cheapo Ebay purchase, it does its job pretty well. :)

    --
    "If you lie to the compiler,
    it will get its revenge."
    Henry Spencer
    2.718281828459045235360287471352662497757247093699959574966967627.com
     
  8. Guest

    Bubba <> wrote:
    > DoN. Nichols's log on stardate 07 stu 2011
    >
    >>> Is that correct, for example, in respect of gcc, web performance and
    >>> simmilar things?

    >> Well ... I tend to prefer the compiler in Sun's "Studio 12"
    >> development package -- if you can still get it along with Solaris 10
    >> since the Oracle takeover, but there are some programs which really
    >> need gcc to compile. My experience is that one takes longer to
    >> compile, but produces somewhat faster code. (And I forget which is
    >> which at the moment.)

    >
    > Hehe, well, I'd sacrifice longer compile time (that is, after all, "one
    > time job") for faster run time.
    >
    > Either way, I need CLI tool, so I believe "Studio 12" or whatever is IDE,
    > so it is probably not worth installing just for compiler.
    >
    > In any case, I'll check it out, thanks.


    FWIW, I almost never use the IDE to compile, but it is well worth having
    it if you need to debug.



    --
    Jim Pennino

    Remove .spam.sux to reply.
     
  9. Bubba <> writes:
    >Doug McIntyre's log on stardate 06 stu 2011
    >> ... I find Solaris to be much faster on *equivilent*
    >> hardware than Linux systems..


    >Is that correct, for example, in respect of gcc, web performance and
    >simmilar things?


    Sure. But the comparisons are usually extremely lopsided. Ie. they
    take a 12 year old Sun server, and run it up against a modern Intel
    Xeon, and trump that the Xeon CPU is faster, disk is faster, etc.

    >I was afraid there might be something important in ALOM that would
    >prevent me from accessing OS or something like that.


    Only if it doesn't let you boot without a password.

    >As it appears, I should be good to go to fully functional Solaris.


    >One more thing - there are two drives that came with machine, loaded
    >with Solaris. Now when I star the installation procedure, I am offered
    >an upgrade or "initialization" (I might be wrong, but something like
    >that). Since I only skimmed through the instalation procedure and
    >haven't reall started it yet, is it possible to wipe/zero-fill/LLF SCSI
    >drives from the console, just to avoid any potential fuss with the
    >instalation?


    If you do an initial install, vs. a an upgrade, you'll have the option
    to redo the drive layout, and it'll newfs all your new partitions
    right over anything old.

    >Since machine came from financial institution, I was affraid it might
    >have some thorougher security roles. Lucky me, however, I was wrong...
    >:)


    In my experience, the bigger the institution, the more lax the
    security/knowledge/sys admin experience.

    >>>.OpenBootPROM Version <<<

    >My wories were regarding potential incompatibility with additional
    >hardware (more RAM, new CPU cards, etc.); so, it's not worth it?


    It is a remote possibility that you could require a new OBP to support
    the newest CPUs, I don't know. I search for the release notes for 142707-01,
    but didn't really see anything that called out to me as required for
    newer CPUs than you may have. If you buy CPU upgrades from a decent
    dealer, they may just help you out if something is required.

    >BTW, JFTR, I measured power consumption of V440:


    Looks like about right from what I remember of those class of machines.
    It got worse after these, then better after those.
     
  10. DoN. Nichols

    DoN. Nichols Guest

    On 2011-11-07, Bubba <> wrote:
    > DoN. Nichols's log on stardate 07 stu 2011
    >
    >>> Is that correct, for example, in respect of gcc, web performance and
    >>> simmilar things?

    >> Well ... I tend to prefer the compiler in Sun's "Studio 12"
    >> development package -- if you can still get it along with Solaris 10
    >> since the Oracle takeover, but there are some programs which really
    >> need gcc to compile. My experience is that one takes longer to
    >> compile, but produces somewhat faster code. (And I forget which is
    >> which at the moment.)

    >
    > Hehe, well, I'd sacrifice longer compile time (that is, after all, "one
    > time job") for faster run time.


    With the tradeoff that if you are doing a lot of compiles in
    debugging, the faster compile is a benefit until you are ready for
    production. (So write your code so it will compile on both. :)

    Here is the comparison (using the old dhrystone benchmark) with
    the two compilers on a Sun Blade 2000 with dual 1.2 GHz CPUs:

    ======================================================================
    * MACHINE MICROPROCESSOR OPERATING COMPILER DHRYSTONES/SEC.
    * TYPE SYSTEM NO REG REGS
    * -------------------------- ------------ ----------- ---------------
    * SB-2K US-III dual 1.2 GHZ Solaris 10 cc 3086419 3086419
    * SB-2K US-III dual 1.2 GHZ Solaris 10 gcc 3355704 3355704
    ======================================================================

    So -- since gcc produced the faster run time (in this one very
    limited benchmark), then the cc with Studio 12 must have been the faster
    compile time.

    with lots of other machine results snipped out.

    > Either way, I need CLI tool, so I believe "Studio 12" or whatever is IDE,
    > so it is probably not worth installing just for compiler.


    There are a bunch of CLI compilers in the Studio 12 package, and
    NetBeans is the accompanying IDE, which I have never used.


    ======================================================================
    /opt/SUNWspro/bin:
    analyzer@ cxref@ fbe@ rtc_patch_area@ tcov@
    b2m@ dbx@ fdumpmod@ rxm@ tha@
    bcheck@ dem@ fpp@ rxs@ uil2xd@
    bil2xd@ dmake@ fpr@ sbcleanup@ uninstaller@
    binopt@ dumpstabs@ fpversion@ sbenter@ version-5.0@
    c++filt@ dwarfdump@ fsplit@ sbquery@ version@
    c89@ ellcc@ gil2xd@ sbtags@ visu@
    c99@ er_archive@ gnuattach@ smallxd@ visuroot@
    cb@ er_cp@ gnuclient@ smctl@ whatdir@
    cc-5.0@ er_export@ gnudoit@ sparcv9/ xdcapture@
    cc@ er_kernel@ gvim@ ss_attach@ xdconfig@
    CC@ er_mv@ indent@ sunas@ xdesigner@
    CCadmin@ er_print@ libsunperf_check@ sunc89@ xdhelp@
    cflow@ er_rm@ lint@ sunc99@ xdrecord@
    checkjava@ er_src@ lock_lint@ suncc@ xdreplay@
    collect@ etags@ ootags@ sunCC@ xdroot@
    cscope@ f77@ prepare_system@ sunf77@ xdtosj@
    ctc@ f90@ ptclean@ sunf90@ xemacs-mule@
    ctcr@ f95@ rcs-checkin@ sunf95@ xemacs@
    ctrace@ fbe-4.0@ rdtimgr@ sunstudio@
    ======================================================================

    These are all duplicated by more links in /usr/bin. So lots of CLI
    compilers, including several flavors of C, and several versions of
    FORTRAN as well as tools like indent, cflow, cscope, and dbx (the
    debugger).

    BTW Also beware that some programs may not absolutely require gcc
    to compile, but will look like it, because they require a gnu
    (or similar) version of make.

    BTW Also notice several gnu programs are in the above list.

    > In any case, I'll check it out, thanks.


    Enjoy.

    >>> I was afraid there might be something important in ALOM that would
    >>> prevent me from accessing OS or something like that.

    >> Only if you are trying to control it remotely (via ALOM). If
    >> the V440 is like the Sun Fire 280R, you can even pull out the card
    >> which implements ALOM or RSC (mine had RSC, not ALOM). However, the
    >> Sun Fire V120 has the LOM chip on the system board, so you can't do
    >> that, though you can change a jumper and power it up to flush the
    >> current settings (including passwords) in the LOM.

    >
    > I saw some jumpers, yes, and I thought they might be for something like
    > clearing settings. However, I couldn't find any sane documentation
    > regarding that card per se, so I'll have to give it a try a bit more...


    Is the ALOM on the V440 a separate card, like the RCS on the Sun
    Fire 280R? If so, you can simply pull the card and get it all out of
    the way until you have an installed OS, and then can reinstall the ALOM
    card and clean the user list and associated passwords in there.

    >> If you select "initial install" (I think that is the term, not
    >> "initialize"), you will newfs all the disk partitions, and if you
    >> re-partition (select the "custom" option when it gets to asking about
    >> disk partitions), that will pretty much force a newfs on each changed
    >> partition. But really, an "initial install" will clear away anything
    >> on the disk anyway -- though not as thoroughly as it would with the
    >> format "analyze" options set for one of the modes to "destroy
    >> existing data" -- but I hope that you have plenty of time to twiddle
    >> your thumbs while it does this.

    >
    > Oh, that's great news.
    >
    > Well, I'll just keep it rollin' during the night...


    O.K. The initial install does not take that long, even with
    manual repartitioning, but if you go into format and ask for a
    destructive surface check, that will take forever with 72 GB or larger
    disks. But for your purposes, just letting the initial install do its
    newfs on each partition should be fine.

    >> Freshly scrubbed disks with a new install of Solaris -- I would
    >> *hope*. :) Maybe spare disks which they kept around fully loaded
    >> with the OS for emergencies, and which they swapped in to send away a
    >> "clean" system.

    >
    > I don't really care, I just want to wipe them clean, even if they had
    > human readable text files with PIN's. :)


    If you really fear that such is there (and are worried about it
    being there), then the format with the surface check option. O.K. Here
    is the first menu in format:

    ======================================================================
    FORMAT MENU:
    disk - select a disk
    type - select (define) a disk type
    partition - select (define) a partition table
    current - describe the current disk
    format - format and analyze the disk
    repair - repair a defective sector
    label - write label to the disk
    analyze - surface analysis
    defect - defect list management
    backup - search for backup labels
    verify - read and display labels
    save - save new disk/partition definitions
    inquiry - show vendor, product and revision
    volname - set 8-character volume name
    !<cmd> - execute <cmd>, then return
    quit
    format>
    ======================================================================

    And the choice which I was trying to remember is "analyze". And its
    sub-menu is:

    ======================================================================
    ANALYZE MENU:
    read - read only test (doesn't harm SunOS)
    refresh - read then write (doesn't harm data)
    test - pattern testing (doesn't harm data)
    write - write then read (corrupts data)
    compare - write, read, compare (corrupts data)
    purge - write, read, write (corrupts data)
    verify - write entire disk, then verify (corrupts data)
    print - display data buffer
    setup - set analysis parameters
    config - show analysis parameters
    !<cmd> - execute <cmd> , then return
    quit
    analyze>
    ======================================================================

    The "purge" choice would probably be the best one, and it will take (by
    default) two passes, each with bit patterns which are the compliment of
    the previous one. A third pass with yet another bit pattern would be
    even better, if you have the time to spare. After this, nobody except
    perhaps the intelligence agencies would be able to show what was on
    there before. You can tune the bit patterns among othe things with the
    "setup" menu entry.

    >> O.K. Looks like UK power cords. Nice Fluke meters.

    >
    > Nope, two times. :) It's Croatia (Schuko -


    O.K. Certainly not US ones -- including based on the voltage
    shown.

    > http://en.wikipedia.org/wiki/Schuko) and ammeter's noware near Fluke :D;
    > but for el cheapo Ebay purchase, it does its job pretty well. :)


    That is what matters. The voltmeter is a Fluke, though, is it
    not? I've lost the URL so I can't go back and check.

    Good Luck,
    DoN.

    --
    Remove oil spill source from e-mail
    Email: <> | Voice (all times): (703) 938-4564
    (too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
    --- Black Holes are where God is dividing by zero ---
     
  11. Bubba

    Bubba Guest

    's log on stardate 07 stu 2011

    > FWIW, I almost never use the IDE to compile, but it is well worth
    > having it if you need to debug.


    Seldom is my code non portable, so I tend to use ANSI C (C90) and Visual
    Studio for that purposes. Thus, Sun should only do the numbers... :)

    --
    "If you lie to the compiler,
    it will get its revenge."
    Henry Spencer
    2.718281828459045235360287471352662497757247093699959574966967627.com
     
  12. Bubba

    Bubba Guest

    DoN. Nichols's log on stardate 08 stu 2011

    >> Hehe, well, I'd sacrifice longer compile time (that is, after all,
    >> "one time job") for faster run time.

    > With the tradeoff that if you are doing a lot of compiles in
    > debugging, the faster compile is a benefit until you are ready for
    > production. (So write your code so it will compile on both. :)
    > Here is the comparison (using the old dhrystone benchmark) with
    > the two compilers on a Sun Blade 2000 with dual 1.2 GHz CPUs:


    /snip

    > So -- since gcc produced the faster run time (in this one very
    > limited benchmark), then the cc with Studio 12 must have been the
    > faster compile time.


    I survived compiling of PHP, Apache, MySQL, samba, GMP and several
    others on m68k 33MHz. :)

    (ATTN: may wrap)
    http://2.718281828459045235360287471352662497757247093699959574966967627.com/ddaveuptime.png

    Trust me, there's nothing that can surprise me any more regarding speed
    of computation. :)

    That's why these machines sit in the basement and have SSH connections.
    So you let it compile in the morning, and come back after lunch - et
    voila! :)

    >> Either way, I need CLI tool, so I believe "Studio 12" or whatever is
    >> IDE, so it is probably not worth installing just for compiler.

    > There are a bunch of CLI compilers in the Studio 12 package, and
    > NetBeans is the accompanying IDE, which I have never used.


    /snip

    Well, in any case, I won't have any Solaris specific compilation. Since
    gcc is de facto C compiler standard, I'll just cut myself some slack
    and simply install and forget about it. :)

    >> In any case, I'll check it out, thanks.

    > Enjoy.


    I will, when I remove machine phisicaly, since it is unberable loud. I
    must admit I don't recall such loud machine in quite some time...

    > Is the ALOM on the V440 a separate card, like the RCS on the Sun
    > Fire 280R?


    Yup.

    > If so, you can simply pull the card and get it all out of the way
    > until you have an installed OS, and then can reinstall the ALOM card
    > and clean the user list and associated passwords in there.


    I did that, but got no output from DB9 RS232 port (ttyb). Did I miss
    something?

    Lack of quality hardware documentation about that server really
    frustrates me. You know like old Sun's had lots of info about switching
    jumpers to change from RS232 to RS422 and simmilar stuff, I can't find
    any refference about jumpers and hardware settings for V440.

    Sun Fire V440 Server Parts Installation and Removal Guide is somewhat
    close to what I seek, but I really don't need 20 pictures of how to
    remove memory from slot... :(

    > O.K. The initial install does not take that long, even with
    > manual repartitioning, but if you go into format and ask for a
    > destructive surface check, that will take forever with 72 GB or
    > larger disks. But for your purposes, just letting the initial
    > install do its newfs on each partition should be fine.


    Well, he's got whole week, since I have plenty of day (real) job to do,
    so actual fiddling with the machine will have to wait for the weekend

    > If you really fear that such is there (and are worried about it
    > being there), then the format with the surface check option. O.K.
    > Here is the first menu in format:


    /snip

    > The "purge" choice would probably be the best one, and it will take
    > (by default) two passes, each with bit patterns which are the
    > compliment of the previous one. A third pass with yet another bit
    > pattern would be even better, if you have the time to spare. After
    > this, nobody except perhaps the intelligence agencies would be able
    > to show what was on there before. You can tune the bit patterns
    > among othe things with the "setup" menu entry.


    Great, thanks for thorough explanation. However, don't worry, after
    first change of data (bits), only dear God, Chuck Norris and time
    machine can whiteness what was on the hard drive. Everything else is
    just purely, plainly - bullshit.

    >>> O.K. Looks like UK power cords. Nice Fluke meters.

    >> Nope, two times. :) It's Croatia (Schuko -

    > O.K. Certainly not US ones -- including based on the voltage
    > shown.


    That's right. :)

    >> http://en.wikipedia.org/wiki/Schuko) and ammeter's noware near Fluke
    >> :D; but for el cheapo Ebay purchase, it does its job pretty well. :)

    > That is what matters. The voltmeter is a Fluke, though, is it
    > not? I've lost the URL so I can't go back and check.


    (ATTN: may wrap)
    http://2.71828182845904523536028747...966967627.com/Sun-V440-Power-Consumption.html

    --
    "If you lie to the compiler,
    it will get its revenge."
    Henry Spencer
    2.718281828459045235360287471352662497757247093699959574966967627.com
     
  13. Bubba

    Bubba Guest

    Doug McIntyre's log on stardate 07 stu 2011

    >>> ... I find Solaris to be much faster on *equivilent*
    >>> hardware than Linux systems..

    >>Is that correct, for example, in respect of gcc, web performance and
    >>simmilar things?

    > Sure. But the comparisons are usually extremely lopsided. Ie. they
    > take a 12 year old Sun server, and run it up against a modern Intel
    > Xeon, and trump that the Xeon CPU is faster, disk is faster, etc.


    Regarding that, is it only my imagination, or it seems that Sun machines
    are not near market ration that they used to be, regardless of Niagara and
    whole multithreading story? In addition, after Oracles (hostile) takeover,
    I would guess that will be even bigger hardware downfall...

    >>Since machine came from financial institution, I was affraid it might
    >>have some thorougher security roles. Lucky me, however, I was
    >>wrong...

    > In my experience, the bigger the institution, the more lax the
    > security/knowledge/sys admin experience.


    That's true, but they usually outsource guys for serious stuff. :)

    >>>>.OpenBootPROM Version <<<

    >>My wories were regarding potential incompatibility with additional
    >>hardware (more RAM, new CPU cards, etc.); so, it's not worth it?

    > It is a remote possibility that you could require a new OBP to
    > support the newest CPUs, I don't know. I search for the release notes
    > for 142707-01, but didn't really see anything that called out to me
    > as required for newer CPUs than you may have. If you buy CPU upgrades
    > from a decent dealer, they may just help you out if something is
    > required.


    Luckily, I decided to go with the option of upgrading the number of CPU's,
    not change the entire platform for faster ones, so that wouldn't be a
    problem anyway.

    However, it is disappointing one cannot find change log or any other
    relevant document about versions of OBP... :\

    >>BTW, JFTR, I measured power consumption of V440:

    > Looks like about right from what I remember of those class of
    > machines. It got worse after these, then better after those.


    Well, according to this:

    http://en.wikipedia.org/wiki/SPARC#SPARC_microprocessor_specifications

    TDP (I guess?) should be 80W. And that would count for fastest CPU. Since
    I do have 300 MHz slower versions (1.28 GHz), I would guess their power
    consumption would be around 65-70W. I wonder, where did almost 150W go if
    that chart is correct...

    I'll try removing one CPU and test with additional CPU's when they arrive
    just to check out how much (approximately) does one CPU card consumes.

    --
    "If you lie to the compiler,
    it will get its revenge."
    Henry Spencer
    2.718281828459045235360287471352662497757247093699959574966967627.com
     
  14. Bubba <> writes:
    > I will, when I remove machine phisicaly, since it is unberable
    > loud. I must admit I don't recall such loud machine in quite some time...


    Any enterprise server in my facility is equally as loud. They only
    care about moving air in and out, and plan on having things like this
    housed in a data center. No need to get softer fans, other than
    perhaps a power saving benefit.

    > I did that, but got no output from DB9 RS232 port (ttyb). Did I miss
    > something?


    Console is always ttya on a Sun system. On your specific model (as
    well as most newer ones), that is called 'Serial MGMT'. That also
    bypasses the ALOM on your model box in general, unless they specific
    tied all console in/out through the ALOM. Then removing the ALOM card
    should revert console output back to 'Serial MGMT'.


    > Lack of quality hardware documentation about that server really
    > frustrates me. You know like old Sun's had lots of info about
    > switching
    > jumpers to change from RS232 to RS422 and simmilar stuff, I can't find
    > any refference about jumpers and hardware settings for V440.


    There aren't that many jumpers any longer, or the ones that remain are
    for emergancy field tech use, not for any normal ops.

    I find this doc to be better overall.. Perhaps this does better for
    you?

    http://download.oracle.com/docs/cd/E19088-01/v440.srvr/816-7728-10/816-7728-10.pdf

    Your doc name is pretty specialized to removing and installing components..

    Oracle does still have the full doc library open and online.

    http://download.oracle.com/docs/cd/E19088-01/v440.srvr/index.html


    >However, it is disappointing one cannot find change log or any other
    >relevant document about versions of OBP... :\


    While the Oracle portal is the official way.. There's others out there..

    http://wesunsolve.net/patch/id/142707-01

    >TDP (I guess?) should be 80W. And that would count for fastest CPU. Since
    >I do have 300 MHz slower versions (1.28 GHz), I would guess their power
    >consumption would be around 65-70W. I wonder, where did almost 150W go if
    >that chart is correct...


    While CPU is the single most largest power draw now, you still have
    hard drives, fans, frame buffers, NICs, other expansion cards, etc.
    Hard drives can be 12-20W a piece. Fans add up there. Chipset on
    motherboard adds up. Etc.
     
  15. Guest

    Bubba <> wrote:
    > 's log on stardate 07 stu 2011
    >
    >> FWIW, I almost never use the IDE to compile, but it is well worth
    >> having it if you need to debug.

    >
    > Seldom is my code non portable, so I tend to use ANSI C (C90) and Visual
    > Studio for that purposes. Thus, Sun should only do the numbers... :)


    IMHO debugging with Visual Studio is primative compared to the Solaris
    Studio tools.

    YMMV.



    --
    Jim Pennino

    Remove .spam.sux to reply.
     
  16. Bubba

    Bubba Guest

    's log on stardate 08 stu 2011

    >>> FWIW, I almost never use the IDE to compile, but it is well worth
    >>> having it if you need to debug.

    >> Seldom is my code non portable, so I tend to use ANSI C (C90) and
    >> Visual Studio for that purposes. Thus, Sun should only do the
    >> numbers... :)

    > IMHO debugging with Visual Studio is primative compared to the
    > Solaris Studio tools.
    > YMMV.


    I haven't tried Sun Studio so thoroughly, but you are aware that calling
    VS primitive in any aspect is a very strong and serious statement that I
    personally believe is hardly applicable comparing to anything.

    --
    "If you lie to the compiler,
    it will get its revenge."
    Henry Spencer
    2.718281828459045235360287471352662497757247093699959574966967627.com
     
  17. Bubba

    Bubba Guest

    Doug McIntyre's log on stardate 08 stu 2011

    >> I will, when I remove machine phisicaly, since it is unberable
    >> loud. I must admit I don't recall such loud machine in quite some
    >> time...

    > Any enterprise server in my facility is equally as loud. They only
    > care about moving air in and out, and plan on having things like this
    > housed in a data center. No need to get softer fans, other than
    > perhaps a power saving benefit.


    No argument about loud enterprise servers, but this one has a specific...
    hissing sound, due to those two "turbins" that cool CPU's...

    >> I did that, but got no output from DB9 RS232 port (ttyb). Did I miss
    >> something?

    > Console is always ttya on a Sun system.


    Except whene there is no ttya? :)

    http://www.spectra.com/pdfs/sunfirev440.pdf - ttya is RJ45 serial
    management on ALOM card - DB9 on V440's motherboard is plane RS232...

    > On your specific model (as well as most newer ones), that is called
    > 'Serial MGMT'. That also bypasses the ALOM on your model box in
    > general, unless they specific tied all console in/out through the
    > ALOM. Then removing the ALOM card should revert console output back
    > to 'Serial MGMT'.


    As I've said, I already tried removing ALOM card (it seems to be inserted
    in some kind of proprietary slot) and connected cable to ttyb. Nothing
    happened.

    > There aren't that many jumpers any longer, or the ones that remain
    > are for emergancy field tech use, not for any normal ops.
    > I find this doc to be better overall.. Perhaps this does better for
    > you?
    > http://download.oracle.com/docs/cd/E19088-01/v440.srvr/816-7728-10/816
    > -7728-10.pdf
    > Your doc name is pretty specialized to removing and installing
    > components..


    This is not particularly useful to anyone having more than 3 months of
    experience with any kind of hardware.

    IIRC, there are 3 sets of jumpers on ALOM card. I am pretty sure one is
    used for "removing" data from NVRAM or wherever it is written. I can't
    seem to find any reference on what they are for, at least not officially
    from SUN.

    > Oracle does still have the full doc library open and online.
    > http://download.oracle.com/docs/cd/E19088-01/v440.srvr/index.html


    And again, except for several Solaris related facts, pretty useless, in
    general, comparing to, for example, IBM's Redbooks and datasheets.

    >>However, it is disappointing one cannot find change log or any other
    >>relevant document about versions of OBP... :\

    > While the Oracle portal is the official way.. There's others out
    > there..
    > http://wesunsolve.net/patch/id/142707-01


    Great site, thank you!

    For the love of god, did you look at the bugs list that latest patch
    repairs? Naturally, 142707-01.zip is unavailable for download via Oracle
    site, and I find no reference of it on Google.

    Do you thing that perhaps people maintaining that site (I see that they
    have IRC) could provide that patch, unofficially and without liability,
    naturally?

    >>TDP (I guess?) should be 80W. And that would count for fastest CPU.
    >>Since I do have 300 MHz slower versions (1.28 GHz), I would guess
    >>their power consumption would be around 65-70W. I wonder, where did
    >>almost 150W go if that chart is correct...

    > While CPU is the single most largest power draw now, you still have
    > hard drives, fans, frame buffers, NICs, other expansion cards, etc.
    > Hard drives can be 12-20W a piece. Fans add up there. Chipset on
    > motherboard adds up. Etc.


    SCSI power consumption is known and does not vary much, but yes, coming to
    think of, 8 memory modules (ECC Reg.), chipset(s), 2x Quad Gigabit
    NIC's... It adds up quickly. :)

    --
    "If you lie to the compiler,
    it will get its revenge."
    Henry Spencer
    2.718281828459045235360287471352662497757247093699959574966967627.com
     
  18. Guest

    Bubba <> wrote:
    > 's log on stardate 08 stu 2011
    >
    >>>> FWIW, I almost never use the IDE to compile, but it is well worth
    >>>> having it if you need to debug.
    >>> Seldom is my code non portable, so I tend to use ANSI C (C90) and
    >>> Visual Studio for that purposes. Thus, Sun should only do the
    >>> numbers... :)

    >> IMHO debugging with Visual Studio is primative compared to the
    >> Solaris Studio tools.
    >> YMMV.

    >
    > I haven't tried Sun Studio so thoroughly, but you are aware that calling
    > VS primitive in any aspect is a very strong and serious statement that I
    > personally believe is hardly applicable comparing to anything.


    Unless VS has been vastly improved since the last time I used it, I have
    to stand by my opinion and am sorry that it offends you somehow, but for
    serious debugging Solaris Studio has every other debugger I've ever used
    beat by a wide margin.


    --
    Jim Pennino

    Remove .spam.sux to reply.
     
  19. Bubba <> writes:
    >Except whene there is no ttya? :)


    There is, it is 'Serial MGMT', and is NOT a DE9. It is an RJ45.

    >As I've said, I already tried removing ALOM card (it seems to be inserted
    >in some kind of proprietary slot) and connected cable to ttyb. Nothing
    >happened.


    Not surprising, ttyb is not used by anything console related.
    ttya (AKA Serial MGMT) is. You will have to connect to it with the
    correct RJ45 serial cable (same as used by Cisco, Juniper, almost
    every network hardware vendor).
     
  20. Bubba

    Bubba Guest

    Doug McIntyre's log on stardate 08 stu 2011

    >>Except whene there is no ttya? :)

    > There is, it is 'Serial MGMT', and is NOT a DE9. It is an RJ45.


    And it is exactly the one I'm using, but it's on ALOM card!

    >>As I've said, I already tried removing ALOM card (it seems to be
    >>inserted in some kind of proprietary slot) and connected cable to
    >>ttyb. Nothing happened.

    > Not surprising, ttyb is not used by anything console related.
    > ttya (AKA Serial MGMT) is. You will have to connect to it with the
    > correct RJ45 serial cable (same as used by Cisco, Juniper, almost
    > every network hardware vendor).


    :) Well, as I've mentioned in posts early, I am connected to ALOM card,
    that is ttya, and that's where my initial issue started - I was unsure
    whether inability to login to ALOM had anything to do with normal
    functionality of whole machine, since only way to access it (without frame
    buffer) is ttya *on ALOM card*.

    --
    "If you lie to the compiler,
    it will get its revenge."
    Henry Spencer
    2.718281828459045235360287471352662497757247093699959574966967627.com
     
    1. Advertising

Want to reply to this thread or ask your own question?

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. Michael Wuest
    Replies:
    0
    Views:
    1,545
    Michael Wuest
    Jan 25, 2004
  2. Jonathan Sturges

    V440 ALOM console problem

    Jonathan Sturges, Feb 21, 2005, in forum: Sun Hardware
    Replies:
    2
    Views:
    627
    Jonathan Sturges
    Feb 22, 2005
  3. Mario
    Replies:
    2
    Views:
    1,594
    Mario
    Aug 23, 2008
  4. Doug McIntyre

    Re: Resetting ALOM password

    Doug McIntyre, Sep 14, 2010, in forum: Sun Hardware
    Replies:
    1
    Views:
    1,369
    Doug McIntyre
    Sep 14, 2010
  5. alacky3
    Replies:
    15
    Views:
    1,027
    Albert van der Horst
    Jun 1, 2011
Loading...

Share This Page