Ken, I'm still a little confused. Are you trying to boot an already installed OS, or are you trying to boot from cd? If the former is true, what does the partition table on the drive look like? BTW, the sun serial console should work at 9600/8/n/1 on a null modem cable connected to your PC's serial port (or even an old vt100 terminal). Can you share with me the output of probe-scsi? If you get me a little more info, I can probably help you out. Like I said, I run linux on sparc at home (and have for years), and I used to be a sparc / solaris system admin a few years back. Anyway, good luck. --Adam On Tue, March 20, 2007 9:40 pm, ken jones said:
Doug et al
I found my cross over cables. printenv told me to set up at 9600bps ASCI 8 none 1. I have tried VT100 and auto-detect. When I turn on the sparc it tells me that it is using ttya for input. However, my hyperterminal sees nothing. If I slow down the expected clock rate to 2400 I get to letters "pp" when I startup the sparc. I do not get "ok" and I get no response when I hit the "Enter" key. Too bad, since that connection would allow me to show you what my terminal is saying.
I did get a good response from probe-scsi and probe-scsi-all. It saw both hard drives.
I am now trying test-all. Last time it just hung but now it seems to be correctly testing the terminal screen.
OKAY! I know something. Self test failed a Transceiver check. It says: Transceiver check -- failed Selftest Failed. Return Code = -1 Testing sbus@1f, 0/SUNW, fas@e 8800000 CE DMA fill from address ffef8000 for 80 bytes Testing sbus@1f, 0/SUNW fdtwo@f, 1400000 Testing floppy disk system. A formatted disk should be in the drive. No diskette or incorrect format. Self test failed. Return code = -1
I wondered what is the "transceiver test" which is failing. The word "transceiver" usually is referring to I/O. I was correct. I installed the ethernet cable and the transceiver test passed.
I know one thing now. The CDROM reader does not work. Part of test-all is to check for a formatted disk in the reader. I put a known good disk in, but the SUN does not see it. That does not explain why it will not boot up correctly. It should note that there is no disk and boot to installed linux.
My latest status is to try booting over my office network. I gave my home router/DHCP server the new sparc's ethernet address and assigned an IP address. I have no idea whether the sparc is configured to expect a DHCP address. I do not see where I can configure that.
Ken Jones
Doug Mildram wrote:
doug> I know the Sparc a bit.
give me a call hell, i am trashing sparc's here, incl. 1 ultra1
also: sun keyboard is a big factor. If it's seen at powerup, its the console (monitor or no monitor)
if keyb not attached on powerup, the com1 equivalent /dev/ttya (DIN or 25pin, varies) becomes the console.
doug 508-243-6580 cell
MORE RANDOM TIPS TO MYSELF YEARS AGO: on a SPARC instead of BIOS we have "eeprom" or "rom prompt"....
There's no "F2" or "DELETE" hint during bootup, but there is one "dangerous" way to stop either the OS or the boot-up.
L1-a or STOP-A is like a ctl-Z to the OS...or an interrupt a bootup.
eeprom prompt is "ok" ok>
If you accidentally STOP the OS by mistake you can get it to go BACK UP WITH
ok> go or maybe ok> continue
normally we do NOT DO MUCH with the eeprom settings, AND, it is possible to tweak eeprom settings from either situation: 1) from the ok> prompt with "setenv" 2) from SOLARIS with # eeprom
So you can see all the settings with a simple anyuser% /usr/sbin/eeprom
Syntax is slightly different. I wont talk about UNIX "eeprom" anymore except to say there is a man page.
BACK TO THE ok> PROMPT you can...most notably:
ok> setenv auto-boot? false or ok> setenv auto-boot? true
the question mark is not a "wild character". It just implies that it's a true/false parameter.
false means it will *NOT* boot to solaris... on power up. true means "normal"...will try to load solaris from boot-device
a UNIX "reboot" command will override this setting.
-----boot "device" in eeprom NORMALLY the eeprom setting for boot-device has a few words and "disk" should be the first word.
ok> printenv boot-device
may return disk diskbrd diskisp disksoc net I only know what "disk" and "net" means..... "net" is useless unless you have another "diskless server" set up.
ok> probe-scsi (can you see the CD drive?) Target 0 Unit 0 Disk .....
maybe not . but if the CD is on an external scsi bus you may need to say
ok> probe-scsi-all (can you see the CD drive?)
pci@1f,4000/scsi@3,1 (a 2nd controller for the external scsi port) Target 6 Unit 0 removable read only device TOSHIBA..... Unit 0 Disk .....
pci@1f,4000/scsi@3 Target 0 Unit 0 Disk .....
BACK TO installation. hopefully you simply say:
with solaris2.7 CD in the drive,
ok> boot cdrom
didnt work? it must have been looking for
ok> devalias
shows
cdrom /pci@1f,4000/scsi@3/disk@6,0:f
BUT I HAD A CD drive on the 3,1 controller so I can say:
boot /pci@1f,4000/scsi@3,1/disk@6,0:f
AND THIS WORKED on this particular machine.
(if you see the CD light come on, and eventually see some kernel string like
SunOS 5.7 Generic_106541-17 sun4u sparc SUNW,Ultra-Enterprise
THEN YOU KNOW you got the right command.
Next you hope your CD and CD drive dont have too many read errors.
ok>
SPECIAL BOOT ARGUMENT......to rebuild /dev* entries.
lets say you ADD HARDWARE like a 2nd disk.
UNIX% halt ...add a disk if you can...then you'd want to
ok> boot -r (THIS Reconfigures /dev and /devices, so it would add /dev/???/??? for the new disk)
SPECIAL BOOT ARGUMENT......to stop at singleuser.
ok> boot -s (would stop at runlevel 1 for the default kernel ..on disk?)
But even then (unlike most unix's) it prompts you for the root password, i think. So dont forget it. ------------------doug
_______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug