Ah, yes, that old confusion. But the problem is df -m reports in 2**20
blocks. So if we divide that by 1000, the result is not the right GB (10**9
blocks). Calling df -HmP probably would be another solution.
-Jongki
On Mon, Feb 7, 2011 at 3:11 AM, Cleber Rosa <cr...@redhat.com> wrote:
> On 02/05/2011 07:22 PM, Jongki Suwandi wrote:
>
> I noticed Host.check_diskpace() in server/hosts/base_classes.py does this:
>
> logging.info('Checking for >= %s GB of space under %s on machine
> %s',
> gb, path, self.hostname)
> df = self.run('df -mP %s | tail -1' % path).stdout.split()
> free_space_gb = int(df[3])/1000.0
> if free_space_gb < gb:
> raise error.AutoservDiskFullHostError(path, gb, free_space_gb)
> else:
> logging.info('Found %s GB >= %s GB of space under %s on
> machine %s',
> free_space_gb, gb, path, self.hostname)
>
> I just want to confirm before I change it, but isn't that supposed to be
> 1024 instead of 1000? The difference between those two values might not be
> much, but it could overinflate the free space by several gigabytes.
>
>
> Jongki,
>
> This is probably due to the fact that one GB is indeed 10**9 bytes (and one
> GiB is 2**30 bytes).
>
> See: http://en.wikipedia.org/wiki/Gibibyte
>
> Cheers.
>
> CR.
>
> Thanks,
> -Jongki
>
>
> _______________________________________________
> Autotest mailing
> listAutotest@test.kernel.orghttp://test.kernel.org/cgi-bin/mailman/listinfo/autotest
>
>
>
_______________________________________________
Autotest mailing list
Autotest@test.kernel.org
http://test.kernel.org/cgi-bin/mailman/listinfo/autotest
