cloud-init-dev team mailing list archive

[Dan Watkins <daniel.watkins@xxxxxxxxxxxxx>]

Diff comments:

> === modified file 'cloudinit/config/cc_disk_setup.py'
> --- cloudinit/config/cc_disk_setup.py	2015-07-22 19:14:33 +0000
> +++ cloudinit/config/cc_disk_setup.py	2015-10-19 14:33:26 +0000
> @@ -343,13 +345,16 @@
>  
>  
>  def get_mbr_hdd_size(device):
> -    size_cmd = [SFDISK_CMD, '--show-size', device]
> -    size = None
> +    size_cmd = [SFDISK_CMD, '-l', device]
>      try:
> -        size, _err = util.subp(size_cmd)
> +        output, _ = util.subp(size_cmd)
>      except Exception as e:
>          raise Exception("Failed to get %s size\n%s" % (device, e))
>  
> +    match = re.search(r'(?P<sector_count>\d+) sectors', output)
> +    if match is None:
> +        raise Exception("Failed to get %s size\n%s" % (device, e))
> +    size = match.groupdict()['sector_count']
>      return int(size.strip())

I considered both of those options, but didn't go for them because it isn't clear to me that sfdisk expects to use 512-byte sectors.

For example, if we had a disk that sfdisk treats as having 1024 byte sectors, all of our partition definitions would be out by a factor of two.

I have opted for parsing the output of sfdisk, because that seems like the best way of finding out how sfdisk views the disk in question.

Given all the stuff you do with curtin, you certainly have more knowledge in this area than me.  Am I missing something that would make this line of reasoning incorrect?

>  
>  


-- 
https://code.launchpad.net/~daniel-thewatkins/cloud-init/lp1460715/+merge/274897
Your team cloud init development team is requested to review the proposed merge of lp:~daniel-thewatkins/cloud-init/lp1460715 into lp:cloud-init.