ArchiveOrangemail archive

Caiman project discussion


caiman-discuss.opensolaris.org
(List home) (Recent threads) (350 other OpenSolaris lists)

Subscription Options

  • RSS or Atom: Read-only subscription using a browser or aggregator. This is the recommended way if you don't need to send messages to the list. You can learn more about feed syndication and clients here.
  • Conventional: All messages are delivered to your mail address, and you can reply. To subscribe, send an email to the list's subscribe address with "subscribe" in the subject line, or visit the list's homepage here.
  • This list contains about 27,768 messages, beginning Jan 2007
  • This list doesn't seem to be active
Report the Spam
This button sends a spam report to the moderator. Please use it sparingly. For other removal requests, read this.
Are you sure? yes no

Pre-OBP 4.17.1 legacy sun4u (off-media /platform/sun4u/wanboot)

Ad
Martin Bochnig 1240838957Mon, 27 Apr 2009 13:29:17 +0000 (UTC)
Hey dear Caiman crowd, did you meanwhile find a fix for

http://www.opensolaris.org/jive/thread.jspa?m...
http://mail.opensolaris.org/pipermail/caiman-...


<<<<<<QUOTE>>>>>>
Currently we only support sun4u systems that are capable of using
wanboot with network-boot-arguments arguments. YOu can find out by
running eeprom.  This is supported on all sun4v. For sun4u, I know they
are supported on ultra45, sb2500, sb100, sb150, v215, v245, v480, v240
... etc.  As long as the OBP is 4.17.1 or higher, they will work.

  For other sun4u systems that is not capable of doing wanboot from
network, there are bugs in wanboot itself.  After the wanboot bugs are
fixed, we will need to try to see whether it is possible to get those
other sun4u machine to work with AI. This will not be ready for 2009.1H.

6797441 - wanboot fails to configure network device when booted from media
6805094 -  wanboot install from snv and s10u6/s10u7_04 fails after
loading miniroot.

I will suggest that you add yourself to the bug interest lists. There
will be bug fixes available later and then we can try to come up with
some workaround to see whether they will work.
<<<<<</QUOTE>>>>>>

??

I burnt a few nights and finally (after almost getting mad, due to the
lack of proper debugging for wanboot, other than snoop -v'ing) found
the somewhat hidden solution.
If you like to test, I can upload it right today. Or is this bug already fixed
(Infinitely printing "Requesting Ethernet address for 192.168.111.65"
 (The install and  dhcp server in the same subnet / no router and no
proxy, whatever mixture of those setups ... )  ??


Now I have it working on every legacy sun4u workstation I tested so
far, for example Blade 100 or 150 with pre 4.17.1-OBP, Ultra 10 with
3.1x OBP, SunFire 280erver R with 4.16.4, SunBlade 1000/2000 with
4.16.4 (same board as Blade 1000, but loads another OBP-subimage if
RSC is detected during POST, which does make this box'es behavior
indeed quite different in various points, even with exactly the same
OBP version).

Versionitis is not unimportant because at different parts in OS/Net,
not only in promif but also in the TCP/IP stack, assumptions (and
sometimes mis-assumptions) are being made, what behavior the
underlying OBP level might or or might not feature.


Regards,
%martin
Dave Miner 1240842660Mon, 27 Apr 2009 14:31:00 +0000 (UTC)
Martin Bochnig wrote: > Hey dear Caiman crowd, did you meanwhile find a fix for > > http://www.opensolaris.org/jive/thread.jspa?m... > http://mail.opensolaris.org/pipermail/caiman-... > > > <<<<<<QUOTE>>>>>> > Currently we only support sun4u systems that are capable of using > wanboot with network-boot-arguments arguments. YOu can find out by > running eeprom. This is supported on all sun4v. For sun4u, I know they > are supported on ultra45, sb2500, sb100, sb150, v215, v245, v480, v240 > ... etc. As long as the OBP is 4.17.1 or higher, they will work. > > For other sun4u systems that is not capable of doing wanboot from > network, there are bugs in wanboot itself. After the wanboot bugs are > fixed, we will need to try to see whether it is possible to get those > other sun4u machine to work with AI. This will not be ready for 2009.1H. > > 6797441 - wanboot fails to configure network device when booted from media
This was fixed in build 111. Dave
Martin Bochnig 1240842825Mon, 27 Apr 2009 14:33:45 +0000 (UTC)
On Mon, Apr 27, 2009 at 5:31 PM, Dave Miner wrote: > Martin Bochnig wrote: >> >> Hey dear Caiman crowd, did you meanwhile find a fix for >> >> http://www.opensolaris.org/jive/thread.jspa?m... >> >> http://mail.opensolaris.org/pipermail/caiman-... >> >> >> <<<<<<QUOTE>>>>>> >> Currently we only support sun4u systems that are capable of using >> wanboot with network-boot-arguments arguments. YOu can find out by >> running eeprom. ?This is supported on all sun4v. For sun4u, I know they >> are supported on ultra45, sb2500, sb100, sb150, v215, v245, v480, v240 >> ... etc. ?As long as the OBP is 4.17.1 or higher, they will work. >> >> ?For other sun4u systems that is not capable of doing wanboot from >> network, there are bugs in wanboot itself. ?After the wanboot bugs are >> fixed, we will need to try to see whether it is possible to get those >> other sun4u machine to work with AI. This will not be ready for 2009.1H. >> >> 6797441 - wanboot fails to configure network device when booted from media > > This was fixed in build 111.
No, it doesn't work with 111 LiveDVD's on-disk /platform/sun4u/wanboot. %martin
Martin Bochnig 1240843056Mon, 27 Apr 2009 14:37:36 +0000 (UTC)
On Mon, Apr 27, 2009 at 5:33 PM, Martin Bochnig wrote: >>> 6797441 - wanboot fails to configure network device when booted from media >> >> This was fixed in build 111. > > > No, it doesn't work with 111 LiveDVD's on-disk /platform/sun4u/wanboot.
Err, not really LiveDVD (although it alwas did offer an xterm since Solaris 2.3). What I meant was simply the SXCE SPARC Install-DVD, booted via boot cdrom -o dhcp -F wanboot - install Doesn't work with 111. Where is that fix's diff?
> > %martin >
Martin Bochnig 1240843807Mon, 27 Apr 2009 14:50:07 +0000 (UTC)
On Mon, Apr 27, 2009 at 5:37 PM, Martin Bochnig wrote: > On Mon, Apr 27, 2009 at 5:33 PM, Martin Bochnig wrote: >>>> 6797441 - wanboot fails to configure network device when booted from media >>> >>> This was fixed in build 111. >> >>
Dave, so you are referring to http://bugs.opensolaris.org/view_bug.do;jsess... ? Then that bug describes 2 problems in one description. Whereas the 2nd thing was the actual problem. Namely not that you have that problem on a SF280R only, namely there that you run into "Last Trap: Fast Data Access MMU Miss" if you try to boot wanboot from CD or DVD. The real problem is only mentioned in half a sentence at the bottom of the page, after the so called workaround (not for the real, namely the 2nd [!] problem): http://bugs.opensolaris.org/view_bug.do;jsess... """Unfortunately, if you try to use go to do wanboot install, it still fails.""" That's the actual problem, and it still doesn't even seem to have a bug-id. It applies to almost all legacy sun4u systems, not only to the SF280R. Maybe nobody tested.
>> No, it doesn't work with 111 LiveDVD's on-disk /platform/sun4u/wanboot. > > > Err, not really LiveDVD (although it alwas did offer an xterm since > Solaris 2.3). > What I meant was simply the SXCE SPARC Install-DVD, booted via > boot cdrom -o dhcp -F wanboot - install > > Doesn't work with 111. > Where is that fix's diff? > >> >> %martin >> >
Martin Bochnig 1240844715Mon, 27 Apr 2009 15:05:15 +0000 (UTC)
Ok, if nobody believes me that this bug exists, maybe I had a 2 weeks
long dream.
On Mon, Apr 27, 2009 at 5:50 PM, Martin Bochnig wrote: > On Mon, Apr 27, 2009 at 5:37 PM, Martin Bochnig wrote: >> On Mon, Apr 27, 2009 at 5:33 PM, Martin Bochnig wrote: >>>>> 6797441 - wanboot fails to configure network device when booted from media >>>> >>>> This was fixed in build 111. >>> >>> > > > Dave, so you are referring to > http://bugs.opensolaris.org/view_bug.do;jsess... > ?? > > Then that bug describes 2 problems in one description. > Whereas the 2nd thing was the actual problem. > Namely not that you have that problem on a SF280R only, namely there > that you run into ?"Last Trap: Fast Data Access MMU Miss" if you try > to boot wanboot from CD or DVD. > The real problem is only mentioned in half a sentence at the bottom of > the page, after the so called workaround (not for the real, namely the > 2nd [!] problem): > > http://bugs.opensolaris.org/view_bug.do;jsess... > > """Unfortunately, if you try to use go to do wanboot install, it still fails.""" > That's the actual problem, and it still doesn't even seem to have a bug-id. > It applies to almost all legacy sun4u systems, not only to the SF280R. > > Maybe nobody tested. > > >>> No, it doesn't work with 111 LiveDVD's on-disk /platform/sun4u/wanboot. >> >> >> Err, not really LiveDVD (although it alwas did offer an xterm since >> Solaris 2.3). >> What I meant was simply the SXCE SPARC Install-DVD, booted via >> boot cdrom -o dhcp -F wanboot - install >> >> Doesn't work with 111. >> Where is that fix's diff? >> >>> >>> %martin >>> >> >
Martin Bochnig 1240846773Mon, 27 Apr 2009 15:39:33 +0000 (UTC)
Many other legacy sun4u pre-4.17.1 platforms behave differently.
I recommend you to try this on Ultra 2, 5/10, 30, 60, 80, Blade 1000/2000.
Then you see what I mean.


The SF280R works indeed, does not infinitely wait for ARP to respond,
which never happens, simply because the install client fails to make
an ARP broadcast on certain Workstation / OBP combinations, so: No
broadcast, no response, endlessly loop through timeouts!!!
(although I still have the minor "Last Trap: Fast Data Access MMU
Miss" error on the SF280R, still, with b111 SXCE Install-DVD, but this
problem can easily be avoided by booting wanboot from hdd, so I wonder
how 111 has fixed any of the 2 bugs).

But so does a pre 4.17.1 Blade 100/150 work well (no ARP misbehavior).
It works by accident, only on them, not on many others, due to OBP
version related mis-assumptions.
Doesn't have Sun any of those so called legacy boxes for a quick test?
As long as nobody confirms the bug's existance, I won't fix it.
> On Mon, Apr 27, 2009 at 5:50 PM, Martin Bochnig wrote: >> On Mon, Apr 27, 2009 at 5:37 PM, Martin Bochnig wrote: >>> On Mon, Apr 27, 2009 at 5:33 PM, Martin Bochnig wrote: >>>>>> 6797441 - wanboot fails to configure network device when booted from media >>>>> >>>>> This was fixed in build 111. >>>> >>>> >> >> >> Dave, so you are referring to >> http://bugs.opensolaris.org/view_bug.do;jsess... >> ?? >> >> Then that bug describes 2 problems in one description. >> Whereas the 2nd thing was the actual problem. >> Namely not that you have that problem on a SF280R only, namely there >> that you run into ?"Last Trap: Fast Data Access MMU Miss" if you try >> to boot wanboot from CD or DVD. >> The real problem is only mentioned in half a sentence at the bottom of >> the page, after the so called workaround (not for the real, namely the >> 2nd [!] problem): >> >> http://bugs.opensolaris.org/view_bug.do;jsess... >> >> """Unfortunately, if you try to use go to do wanboot install, it still fails.""" >> That's the actual problem, and it still doesn't even seem to have a bug-id. >> It applies to almost all legacy sun4u systems, not only to the SF280R. >> >> Maybe nobody tested. >> >> >>>> No, it doesn't work with 111 LiveDVD's on-disk /platform/sun4u/wanboot. >>> >>> >>> Err, not really LiveDVD (although it alwas did offer an xterm since >>> Solaris 2.3). >>> What I meant was simply the SXCE SPARC Install-DVD, booted via >>> boot cdrom -o dhcp -F wanboot - install >>> >>> Doesn't work with 111. >>> Where is that fix's diff? >>> >>>> >>>> %martin >>>> >>> >> >
mary ding 1240851802Mon, 27 Apr 2009 17:03:22 +0000 (UTC)
Martin and Dave:

I am the original bug filer for the wanboot bugs. I found the problem on 
both the 280R and Sun blade 2000 in nevada and s10 updates.

When snv_111 was available, I did try to boot the DVD media and I was 
able to do a regular wanboot install on the sunblade 2000 and confirmed 
the bug was fixed.

This is fixed in snv_111 and this is from line2-sb2k:

Probing system devices


Sun Blade 1000 (2 X UltraSPARC-III+) , No Keyboard
Copyright 1998-2004 Sun Microsystems, Inc.  All rights reserved.
OpenBoot 4.16.4, 1024 MB memory installed, Serial #51680977.
Ethernet address 0:3:ba:14:96:d1, Host ID: 831496d1.



Rebooting with command: boot cdrom -o prompt -F wanboot - install
Boot device: /pci at 8,700000/scsi at 6/disk at 6,0:f  File and args: -o prompt 
-F wanboot - install
<time unavailable> wanboot info: WAN boot messages->console
<time unavailable> wanboot info: configuring /pci at 8,700000/network at 5,1:


<time unavailable> wanboot info: Default net-config-strategy: manual

boot>  router-ip=10.6.35.1

boot> hostname=line2-sb2k

boot> host-ip=10.6.35.141

boot> subnet-mask=255.255.255.0

boot> bootserver= http://10.6.35.8/cgi-bin/wanboot-cgi

boot> list

host-ip:      10.6.35.141
subnet-mask:  255.255.255.0
router-ip:    10.6.35.1
hostname:     line2-sb2k
http-proxy:   UNSET
client-id:    UNSET
aes:          *HIDDEN*
3des:         *HIDDEN*
sha1:         *HIDDEN*
bootserver:   http://10.6.35.8/cgi-bin/wanboot-cgi

boot>  go

<time unavailable> wanboot progress: wanbootfs: Read 370 of 370 kB (100%)
<time unavailable> wanboot info: wanbootfs: Download complete
Tue Mar 31 02:34:15 wanboot progress: miniroot: Read 208182 of 208182 kB 
(100%)
Tue Mar 31 02:34:15 wanboot info: miniroot: Download complete
SunOS Release 5.11 Version snv_111 64-bit
Copyright 1983-2009 Sun Microsystems, Inc.  All rights reserved.
Use is subject to license terms.
WARNING: Last shutdown is later than time on time-of-day chip; check date.
strplumb: open /devices/pseudo/clone at 0:sd failed: 19
Configuring /dev
WANBOOT: Setting up install
Network interface was configured manually.
10.6.35.141
Reading ZFS config: done.
Serial console, reverting to text install
WANBOOT: using non-graphical install
Beginning system identification...
Searching for configuration file(s)...
Using sysid configuration file http://10.6.35.8/config/sysidcfg
Search complete.
Discovering additional network configuration...
Completing system identification...
Starting remote procedure call (RPC) services: done.


The 280R was not available for testing until now. I will try snv_113 and 
see whether it works.  If it still fails, then we need
Martin Bochnig wrote: > Ok, if nobody believes me that this bug exists, maybe I had a 2 weeks > long dream. > > > > On Mon, Apr 27, 2009 at 5:50 PM, Martin Bochnig wrote: >> On Mon, Apr 27, 2009 at 5:37 PM, Martin Bochnig wrote: >>> On Mon, Apr 27, 2009 at 5:33 PM, Martin Bochnig wrote: >>>>>> 6797441 - wanboot fails to configure network device when booted from media >>>>> This was fixed in build 111. >>>> >> >> Dave, so you are referring to >> http://bugs.opensolaris.org/view_bug.do;jsess... >> ? >> >> Then that bug describes 2 problems in one description. >> Whereas the 2nd thing was the actual problem. >> Namely not that you have that problem on a SF280R only, namely there >> that you run into "Last Trap: Fast Data Access MMU Miss" if you try >> to boot wanboot from CD or DVD. >> The real problem is only mentioned in half a sentence at the bottom of >> the page, after the so called workaround (not for the real, namely the >> 2nd [!] problem): >> >> http://bugs.opensolaris.org/view_bug.do;jsess... >> >> """Unfortunately, if you try to use go to do wanboot install, it still fails.""" >> That's the actual problem, and it still doesn't even seem to have a bug-id. >> It applies to almost all legacy sun4u systems, not only to the SF280R. >> >> Maybe nobody tested. >> >> >>>> No, it doesn't work with 111 LiveDVD's on-disk /platform/sun4u/wanboot. >>> >>> Err, not really LiveDVD (although it alwas did offer an xterm since >>> Solaris 2.3). >>> What I meant was simply the SXCE SPARC Install-DVD, booted via >>> boot cdrom -o dhcp -F wanboot - install >>> >>> Doesn't work with 111. >>> Where is that fix's diff? >>> >>>> %martin >>>> > _______________________________________________ > caiman-discuss mailing list > caiman-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/...
Martin Bochnig 1240910664Tue, 28 Apr 2009 09:24:24 +0000 (UTC)
Folks,

it is more than idiotic, but with b111 all the previous off-media
wanboot problems are indeed fixed.
No eternal "Requesting Ethernet address for <IPv4 of Install-server>
is printed to screen anymore, repeated every 10 minutes.
Everything works.

Somehow I confused the (unlabeled, as my CD-pen dried out a month ago)
SXCE 110, 106 and 111 DVD's. I have too many media laying on my desk.


It is really idiotic that I was so proud for having fixed the ARP
broadcast bug in ethernet.c, while Sun has done that already and I
could have saved the time.

On the SF280R it still does not work, I get the Trap MMU error. But it
is likely that this has to do with the unsupported 3rd party La Cie
SCSI DVD-RAM drive which is in it, rather than the default Toschiba
SD-M1401 DVD-ROM drive which has a Sun part # and which shipped in
most SCSI based boxes of the 1999-2004 era. It must be related to the
drive's firmware. Most generic SCSI drives did just work, some did
not.

Now I can look into the next issue, now of course everything should be
way less dramatic, as the boot does work now. On all sun4u workstation
models (except U1), which I all just tested.

The next problem is simply, that the booted 111 AI image complains on
all machines:

"Couldn't determine automated install image.
Requesting maintenance mode."
It looks like wanboot may not provide enough parameters to the booted
client install environment.

Now I must check if I get the same on a Blade 100 with OBP 4.17.1 when
I boot directly via OBP's wanboot, rather than off-media's
/platform/sun4u/wanboot.

Another (very minor) issue had been last week, at first, that I needed
to change one char in the instal- and dhcp- server's
/etc/netboot/wanboot.conf from

boot_file=/export/home/martin/Desktop/osol-0906-111-ai-sparc/platform/sun4v/wanboot

to

boot_file=/export/home/martin/Desktop/osol-0906-111-ai-sparc/platform/sun4u/wanboot

Did I miss a sun4u-specific option during installadm create service?
Or is such an option not yet implemented in installadm?

Sorry for yesterday's confusion.
My personal fix for the ARP-broadcast problem was in ethernet.c, which
wanboot is linked against. Now I cannot expect it to check the OS/Net
hg log messages, which set of diffs Sun has developed to fix the
problem.

Final question: If I would change a fe compile flags and minor plaform
related things in OS/Net to revive UltrSPARC_I (64-bit) support, would
you ship it???
Or can I only do this if I redistribute my "own" version of SPARC-Indiana?

Thanks.
Martin Bochnig 1240912618Tue, 28 Apr 2009 09:56:58 +0000 (UTC)
On Tue, Apr 28, 2009 at 12:24 PM, Martin Bochnig wrote: > Now I can look into the next issue, now of course everything should be > way less dramatic, as the boot does work now. On all sun4u workstation > models (except U1), which I all just tested. > > The next problem is simply, that the booted 111 AI image complains on > all machines: > > "Couldn't determine automated install image.
Err, damn laptop keyboard and touchscreen, two words got missing. I do not want to report incomplete error messages. The precise error message is. "Couldn't determine source of automated install image Requesting System Maintenance mode ..." I guess this one is not difficult to fix (properly). Also a workaround should be possible after I log in into System Maintenance milestone. Which command must I call? TIA.
Martin Bochnig 1240913461Tue, 28 Apr 2009 10:11:01 +0000 (UTC)
Ok, I found it.
It is just in the LiveMedia-framework - derived (originally by Moinak
Ghosh, now AI-adopted) /lib/svc/method/live-fs-root SMF script  after
line 184 or after 253, according to grep.
If it is a bug, I respond later with a solution.
Maybe it was simply a user error by myself (during installadm create-service).
On Tue, Apr 28, 2009 at 12:56 PM, Martin Bochnig wrote: > On Tue, Apr 28, 2009 at 12:24 PM, Martin Bochnig wrote: >> Now I can look into the next issue, now of course everything should be >> way less dramatic, as the boot does work now. On all sun4u workstation >> models (except U1), which I all just tested. >> >> The next problem is simply, that the booted 111 AI image complains on >> all machines: >> >> "Couldn't determine automated install image. > > > Err, damn laptop keyboard and touchscreen, two words got missing. > I do not want to report incomplete error messages. > The precise error message is. > > "Couldn't determine source of automated install image > Requesting System Maintenance mode ..." > > > I guess this one is not difficult to fix (properly). > Also a workaround should be possible after I log in into System > Maintenance milestone. > Which command must I call? > > TIA. > > -- > %martin > > >> Requesting maintenance mode." >> It looks like wanboot may not provide enough parameters to the booted >> client install environment. >> >> Now I must check if I get the same on a Blade 100 with OBP 4.17.1 when >> I boot directly via OBP's wanboot, rather than off-media's >> /platform/sun4u/wanboot. >
Martin Bochnig 1240914531Tue, 28 Apr 2009 10:28:51 +0000 (UTC)
On Tue, Apr 28, 2009 at 1:11 PM, Martin Bochnig wrote: > Ok, I found it. > It is just in the LiveMedia-framework - derived (originally by Moinak > Ghosh, now AI-adopted) /lib/svc/method/live-fs-root SMF script ?after > line 184 or after 253, according to grep. > If it is a bug, I respond later with a solution. > Maybe it was simply a user error by myself (during installadm create-service).
Ok, at least on some boxes (right now tested it on Ultra80), /lib/svc/methos/live-fs-root fails at line 261 /sbin/dhcpinfo Rootpath. The NIC drivers are present and attached (on the U80 it is hme). But nothing except lo0 got plumbed. Manually doing "ifconfig -a plumb" does work, however. This makes everything else a small problem. Because the former can have hard-to-find reasons, if and where it rarely happens. Fortunately not here. Now let's see, if $Rootpath is actually empty ...
Martin Bochnig 1240915597Tue, 28 Apr 2009 10:46:37 +0000 (UTC)
On Tue, Apr 28, 2009 at 1:28 PM, Martin Bochnig wrote: > On Tue, Apr 28, 2009 at 1:11 PM, Martin Bochnig wrote: >> Ok, I found it. >> It is just in the LiveMedia-framework - derived (originally by Moinak >> Ghosh, now AI-adopted) /lib/svc/method/live-fs-root SMF script ?after >> line 184 or after 253, according to grep. >> If it is a bug, I respond later with a solution. >> Maybe it was simply a user error by myself (during installadm create-service). > > > Ok, at least on some boxes (right now tested it on Ultra80), > /lib/svc/methos/live-fs-root fails at line 261 > /sbin/dhcpinfo Rootpath. > > The NIC drivers are present and attached (on the U80 it is hme). > But nothing except lo0 got plumbed. > Manually doing "ifconfig -a plumb" does work, however. > This makes everything else a small problem. Because the former can > have hard-to-find reasons, if and where it rarely happens. Fortunately > not here. > > Now let's see, if $Rootpath is actually empty ...
Yep. Five lines earlier (line 256) /sbin/dhcpagent -a fails (maybe it should have a longer timeout value, as extra option?). As no NIC got configured for dhcp, line 261 can only fail, too. Because if hme0 is not even plumbed, not to mention up, obviously line 261 MEDIA='/sbin/dhcpinfo Rootpath' cannot contact the dhcp server to get the dhcp $Rootpath variable from the server. Question is now, why line 256 doesn't work in the first place. Testing Blade 150 and a few other boxes now, to get more comparision experiences ...
Jan Damborsky 1240921391Tue, 28 Apr 2009 12:23:11 +0000 (UTC)
Martin Bochnig wrote: > On Tue, Apr 28, 2009 at 1:28 PM, Martin Bochnig wrote: >> On Tue, Apr 28, 2009 at 1:11 PM, Martin Bochnig wrote: >>> Ok, I found it. >>> It is just in the LiveMedia-framework - derived (originally by Moinak >>> Ghosh, now AI-adopted) /lib/svc/method/live-fs-root SMF script after >>> line 184 or after 253, according to grep. >>> If it is a bug, I respond later with a solution. >>> Maybe it was simply a user error by myself (during installadm create-service). >> >> Ok, at least on some boxes (right now tested it on Ultra80), >> /lib/svc/methos/live-fs-root fails at line 261 >> /sbin/dhcpinfo Rootpath. >> >> The NIC drivers are present and attached (on the U80 it is hme). >> But nothing except lo0 got plumbed. >> Manually doing "ifconfig -a plumb" does work, however. >> This makes everything else a small problem. Because the former can >> have hard-to-find reasons, if and where it rarely happens. Fortunately >> not here. >> >> Now let's see, if $Rootpath is actually empty ...
Martin, please be aware tha Rootpath DHCP option is no longer used by AI images built from latest install bits and not populated by the latest installadm(1M) tools. Usage of Rootpath was removed as part of fix for bug 8130. For more details please see flag day sent to this mailing list with following subject: Flag day: Automated Installer and Distribution Constructor - Sparc - fix for bug 8130 Those changes will appear in build 2009.06-04 which I assume will be available as 111b. Thank you, Jan
Martin Bochnig 1240922386Tue, 28 Apr 2009 12:39:46 +0000 (UTC)
On Tue, Apr 28, 2009 at 3:23 PM, Jan Damborsky wrote: > Martin Bochnig wrote: >> >> On Tue, Apr 28, 2009 at 1:28 PM, Martin Bochnig wrote: >>> >>> On Tue, Apr 28, 2009 at 1:11 PM, Martin Bochnig >>> wrote: >>>> >>>> Ok, I found it. >>>> It is just in the LiveMedia-framework - derived (originally by Moinak >>>> Ghosh, now AI-adopted) /lib/svc/method/live-fs-root SMF script ?after >>>> line 184 or after 253, according to grep. >>>> If it is a bug, I respond later with a solution. >>>> Maybe it was simply a user error by myself (during installadm >>>> create-service). >>> >>> Ok, at least on some boxes (right now tested it on Ultra80), >>> /lib/svc/methos/live-fs-root fails at line 261 >>> /sbin/dhcpinfo Rootpath. >>> >>> The NIC drivers are present and attached (on the U80 it is hme). >>> But nothing except lo0 got plumbed. >>> Manually doing "ifconfig -a plumb" does work, however. >>> This makes everything else a small problem. Because the former can >>> have hard-to-find reasons, if and where it rarely happens. Fortunately >>> not here. >>> >>> Now let's see, if $Rootpath is actually empty ... > > Martin, > > please be aware tha Rootpath DHCP option is no longer used > by AI images built from latest install bits and not populated > by the latest installadm(1M) tools. Usage of Rootpath was removed > as part of fix for bug 8130. For more details please see flag day sent > to this mailing list with following subject: > > Flag day: Automated Installer and Distribution Constructor - Sparc - fix for > bug 8130 > > Those changes will appear in build 2009.06-04 which I assume will be > available > as 111b. > > Thank you, > Jan
Hello Jan, I have to thank you. Yes, I didn't read the flagday message, because I had that ARP problem (and ignored pretty much everything else in life). For nothing, as it turned out, b111 .... Ok, thank you for the flagday pointer. Next I must read that. Although: In this case the problem I described a few hours ago only happens, where off-media /platform/sun4u/wanboot is used. The Blade 100 with OBP 4.17.1 (when booted via boot net:dhcp) is theonly of my machines which has a new enough OBP to boot AI without off-media-/platform/sun4u/wanboot. And presumably it is no coincidence, that she is the only one, which does not have the dhcpagent -a problem. So it seems, not the type of NIC was the reason, but indeed the boot method. I tested U30, U60, SB100 with 4.17.1, SB150 (=100) without 4.17.1, now comes the Blade 2000 and then the SF280R. On a side note: I found that the Blade 100 which does have OBP 4.17.1 (the other does not, for comparison reasons) CANNOT boot into off-media wanboot. I mean, why should it. It's not needed. But at least interesting to witness. If somebody is interested, I stand up and connect a serial console to that box ... Regards, %martin
mary ding 1240934418Tue, 28 Apr 2009 16:00:18 +0000 (UTC)
Martin:

We had a Sunblade100 with 512 MB memory and OBP 4.17.1, I can do osol 
install on there without any problem and it works.

Even with the latest AI nightly build, I have problems getting osol 
install to work on older platforms like Sunblade 2000 and 280R.  This is 
being looked at.
Martin Bochnig wrote: > On Tue, Apr 28, 2009 at 3:23 PM, Jan Damborsky wrote: >> Martin Bochnig wrote: >>> On Tue, Apr 28, 2009 at 1:28 PM, Martin Bochnig wrote: >>>> On Tue, Apr 28, 2009 at 1:11 PM, Martin Bochnig >>>> wrote: >>>>> Ok, I found it. >>>>> It is just in the LiveMedia-framework - derived (originally by Moinak >>>>> Ghosh, now AI-adopted) /lib/svc/method/live-fs-root SMF script after >>>>> line 184 or after 253, according to grep. >>>>> If it is a bug, I respond later with a solution. >>>>> Maybe it was simply a user error by myself (during installadm >>>>> create-service). >>>> Ok, at least on some boxes (right now tested it on Ultra80), >>>> /lib/svc/methos/live-fs-root fails at line 261 >>>> /sbin/dhcpinfo Rootpath. >>>> >>>> The NIC drivers are present and attached (on the U80 it is hme). >>>> But nothing except lo0 got plumbed. >>>> Manually doing "ifconfig -a plumb" does work, however. >>>> This makes everything else a small problem. Because the former can >>>> have hard-to-find reasons, if and where it rarely happens. Fortunately >>>> not here. >>>> >>>> Now let's see, if $Rootpath is actually empty ... >> Martin, >> >> please be aware tha Rootpath DHCP option is no longer used >> by AI images built from latest install bits and not populated >> by the latest installadm(1M) tools. Usage of Rootpath was removed >> as part of fix for bug 8130. For more details please see flag day sent >> to this mailing list with following subject: >> >> Flag day: Automated Installer and Distribution Constructor - Sparc - fix for >> bug 8130 >> >> Those changes will appear in build 2009.06-04 which I assume will be >> available >> as 111b. >> >> Thank you, >> Jan > > > > Hello Jan, I have to thank you. > Yes, I didn't read the flagday message, because I had that ARP problem > (and ignored pretty much everything else in life). > For nothing, as it turned out, b111 .... > > Ok, thank you for the flagday pointer. Next I must read that. > Although: In this case the problem I described a few hours ago only > happens, where off-media /platform/sun4u/wanboot is used. > The Blade 100 with OBP 4.17.1 (when booted via boot net:dhcp) is > theonly of my machines which has a new enough OBP to boot AI without > off-media-/platform/sun4u/wanboot. And presumably it is no > coincidence, that she is the only one, which does not have the > dhcpagent -a problem. So it seems, not the type of NIC was the reason, > but indeed the boot method. I tested U30, U60, SB100 with 4.17.1, > SB150 (=100) without 4.17.1, now comes the Blade 2000 and then the > SF280R. > > On a side note: I found that the Blade 100 which does have OBP 4.17.1 > (the other does not, for comparison reasons) CANNOT boot into > off-media wanboot. I mean, why should it. It's not needed. But at > least interesting to witness. If somebody is interested, I stand up > and connect a serial console to that box ... > > Regards, > %martin > _______________________________________________ > caiman-discuss mailing list > caiman-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/...
Dave Miner 1240934902Tue, 28 Apr 2009 16:08:22 +0000 (UTC)
> Final question: If I would change a fe compile flags and minor plaform > related things in OS/Net to revive UltrSPARC_I (64-bit) support, would > you ship it??? > Or can I only do this if I redistribute my "own" version of SPARC-Indiana? >
This project isn't the place to ask that question; none of us is going to sponsor such work. I'd suggest it's more appropriate for the ON community list. Dave
Robert Hanke 1262180352Wed, 30 Dec 2009 13:39:12 +0000 (UTC)
Hello:


Is there more current information on the issue of an OpenSolaris SPARC AI install on pre 4.17.1 OPB, specifically on sun4u platforms?

We still have several Ultra5 and Ultra10 systems in production use. They may be long out of production (ours are all circa 2001), but they are functional for everything but Flash. More importantly, as is typical with Sun systems, they are not only usable but still even operational so many years later.

I think I read an install path forward, but if resolved, the instructions at 

OpenSolaris Automated Installer Guide
http://dlc.sun.com/osol/docs/content/2009.06/...

still do not recognize the possibility.

Further, opensolaris.org and opensolaris.com search functions did not usefully respond with queries such as Ultra10 or sun4u. Google did, sigh.

My best hopes for Sun, its employees, and the Sun community in the new year!

Thanks.


mechsys
Ad
Home | About | Privacy