Jan Damborsky wrote:
> Hi Jean,
>
> my apologies that I didn't get back to you earlier with comments,
> it took me a while to go through those changes :-)
>
> Since the fix has been integrated, I have just commented on things
> which I think should be addressed anyway.
> If those comments are relevant, do you think separate bugs should
> be filed to track them ?
Jan, thanks for the comments. My responses are inline.
Jean
>
> Thank you,
> Jan
>
>
> [1] installadm-common.sh
> ------------------------
>
>
> 455 printf "\tmodule
> /${BootLofs}/platform/i86pc/$ISADIR/boot_archive\n" >> ${tmpmenu}
> ->
> 455 printf "\tmodule$
> /${BootLofs}/platform/i86pc/\$ISADIR/boot_archive\n" >> ${tmpmenu}
>
> - '$' has to be appended to 'module' keyword, otherwise GRUB will not
> expand '$ISADIR'
> - '$' in '$ISADIR' has to be prefixed with backlash in order to avoid
> variable expansion
Any thoughts as to why our testing didn't find this? Mary did both 32
and 64 bit installs.
I'll definitely file a bug and fix it.
>
>
> [2] setup-sparc.sh
> ------------------
>
> I have created install service using new Sparc AI image and - it was
> set it up
> as the one with global scope:
>
> # installadm create-service -n 1003s-jeanm -s
> /export/home/ai/iso/ai_sparc_jeanm.iso /export/home/ai/images/1003s-jeanm
> Setting up the target image at /export/home/ai/images/1003s-jeanm ...
> Registering the service 1003s-jeanm._OSInstall._tcp.local
> Service discovery fallback mechanism set up
> Creating SPARC configuration file
> #
>
> Then I tried to create different Sparc service and saw the complaint
> from grep:
>
> # installadm create-service -n 1003s-24_26-126 -s
> /export/home/ai/iso/ai_sparc_24-26_126.iso
> /export/home/ai/images/1003s-24_26-126
> Setting up the target image at /export/home/ai/images/1003s-24_26-126 ...
> Registering the service 1003s-24_26-126._OSInstall._tcp.local
> Service discovery fallback mechanism set up
> Creating SPARC configuration file
> grep: can't open
> /export/home/ai/images/1003s-jeanm/boot/platform/install.conf
> Service is currently being used by SPARC clients which have not
> explicitly been associated with another service via the
> 'create-client' subcommand.
> To select service 1003s-24_26-126 for those SPARC clients, use the
> following commands:
> /usr/bin/rm -f /etc/netboot/wanboot.conf
> /usr/bin/ln -s 1003s-24_26-126/wanboot.conf /etc/netboot
>
> It seems it might be caused by the fact that location of boot_archive
> changed but function get_service_with_global_scope() was not
> modified to reflect this.
Correct. I'll file a bug and fix this. I believe I need backward
compatibility in
this code, correct?
>
>
> [3] ai_post_boot_archive_pkg_image_mod
> --------------------------------------
>
> I have compared Sparc AI images with and without the fix
> and they differ in size - new one is bigger:
>
> # du -sk /export/home/ai/iso/ai_sparc_ai_err_126_final.iso
> 286935 /export/home/ai/iso/ai_sparc_ai_err_126_final.iso
>
> # du -sk /export/home/ai/iso/ai_sparc_jeanm.iso
> 316272 /export/home/ai/iso/ai_sparc_jeanm.iso
>
> And comparing size and content of 'platform' directories:
>
> Old image (boot & platform):
>
> # du -sk /export/home/ai/images/1003s-ai-err-126-final/platform/
> 13841 /export/home/ai/images/1003s-ai-err-126-final/platform/
> # find /export/home/ai/images/1003s-ai-err-126-final/platform/ -type
> f | wc -l
> 10
>
> # du -sk /export/home/ai/images/1003s-ai-err-126-final/boot/
> 166934 /export/home/ai/images/1003s-ai-err-126-final/boot/
> # find /export/home/ai/images/1003s-ai-err-126-final/boot/ -type f |
> wc -l
> 5
>
>
> New image (everything in boot/platform):
>
>
> # du -sk /export/home/ai/images/1003s-jeanm/boot/platform/
> 209239 /export/home/ai/images/1003s-jeanm/boot/platform/
> # find /export/home/ai/images/1003s-jeanm/boot/platform/ -type f | wc -l
> 383
>
>
> My guess is that it is because lines 86-91 are removed
> from original code. Is it intentional or could we keep the
> content of 'platform' minimal as it was in done original script ?
I'll look at this and file a bug as needed. I don't think we can do
exactly what lines
86-91 did because of the restructure but something \with that general
idea is needed.
Jean
>
>
> jeanm wrote:
>> I'd like Jan and Alok to please review the following:
>>
>> http://cr.opensolaris.org/~jeanm/slim_8347_6092/
>>
>> The bugs addressed are the following:
>>
>> 6092 <http://defect.opensolaris.org/bz/show_bug.cgi?id=6092> Rename
>> bootroot to boot_archive in DC
>> 8347 <http://defect.opensolaris.org/bz/show_bug.cgi?id=8347> Move
>> boot archive from /boot/x86
>> <http://defect.opensolaris.org/bz/show_bug.cgi?id=86>.microroot to
>> /platform/i86
>> <http://defect.opensolaris.org/bz/show_bug.cgi?id=86>pc/boot_archive
>>
>>
>>
>> Testing:
>>
>> LiveCD build and install via virtual box
>>
>> And Mary helped with testing doing:
>> AI bare metal install on 64 bit system
>>
>> AI bare metal install on 32 bit system -
>>
>> XVM PV AI install on 32 bit system
>>
>> XVM PV AI install on 64 bit system
>>
>> AI install on sun4u and sun4v
>>
>> Jean
>>
>>
>> _______________________________________________
>> caiman-discuss mailing list
>> caiman-discuss at opensolaris.org
>> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
>
