Hi Jack,
Jack Schwartz wrote:
> Hi everyone.
>
> I did a little more experimenting...
>
> ls only shows the size difference between compressed and uncompressed
> files when fiocompress -c (not -cm) is used. du shows the difference
> whether -c or -cm is used.
>
> When I run du on the build_data/bootroot and the mounted
> sparc.microroot, I see a difference, so compression is being done, at
> least on some files.
>
> The "file" command doesn't show a difference between compressed or
> uncompressed files, unfortunately, so I can't tell if a small file is
> compressed or not. Du reports sizes in 512 or larger units.
>
> Regarding the devfsadm issue, it looks like the lock file
> (/etc/dev/.devfsadm-dev.lock) is processed as the svc repository.db file
> is. Being that repository.db is a large enough file to see that it is
> not being compressed, I conclude that the devfsadm lock file isn't
> either. Based on this, what I have done is working.
>
> While it is true that the error I saw earlier dealt with access of the
> devfsadm lock file, it was because the root file system is read only.
> Not sure how to get around this...
My guess is that the image you generate contains live-fs-root
without fix for
4230 Bootroot assembly changes for SPARC ai images
The fix makes sure that ramdisk is remounted as read-write for Sparc,
since the command is slightly different comparing to the one used
for x86 - the excerpt from fixed live-fs-root script shows:
[...]
echo "\rRemounting root read/write" > /dev/msglog
ISA_INFO=`/sbin/uname -p`
if [ $ISA_INFO = "sparc" ]
then
# ramdisk is differently labeled in sparc vs x86
/sbin/mount -o remount,rw /devices/ramdisk-root:a /
else
# x86 label
/sbin/mount -o remount,rw /devices/ramdisk:a /
fi
[...]
Thank you,
Jan
>
> Thanks,
> Jack
>
>
> On 12/20/08 09:40, Jack Schwartz wrote:
>> Hi Karen.
>>
>> Not sure if you are online... I'm in the office (for just a few more
>> minutes) and heard your disk spin up...
>>
>> I have a question: how can I tell whether or not a file has been
>> fiocompressed. should it be reflected in the file size reported by ls,
>> by the output of some command, that fiocompress -d gives an error or
>> not, ???
>>
>> I can boot my sparc image, but the files which Jan said not to
>> fiocompress still give the same errors. E.g.: when I do a devfsadm it
>> still says the lock file is on a read-only file system. Yet I didn't
>> compress that file, so it should be updatable by devfsadm ... or is
>> there something else I'm not doing right?
>>
>> I'll be leaving shortly, but coming back around 2 for about an hour.
>>
>> I'd like to post my code review, but want to get past this testing first.
>>
>> Thanks,
>> Jack
>> _______________________________________________
>> caiman-discuss mailing list
>> caiman-discuss at opensolaris.org
>> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
>>
>
> _______________________________________________
> caiman-discuss mailing list
> caiman-discuss at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
