Re: [PATCH 4/5] Display drive model and size in MB in partitioning UI (#460697)

[Steffen Maier <maier linux vnet ibm com>]

On 07/07/2009 04:12 AM, David Cantrell wrote:
> On Fri, 3 Jul 2009, Steffen Maier wrote:
>> On 07/03/2009 08:39 AM, David Cantrell wrote:
>>> When a partition table is unreadable, anaconda displays a
>>> window explaining the situation and asks you if you would
>>> like to format the disk for use.  The existing message would
>>> only give the device node name (e.g., /dev/sda47).  This
>>> patch adds the drive model name and capacity to the message
>>> so confused users might know which disk anaconda is talking
>>> about.
>>
>>> diff --git a/storage/devicetree.py b/storage/devicetree.py
>>> index 08846f3..c199212 100644
>>> --- a/storage/devicetree.py
>>> +++ b/storage/devicetree.py
>>> @@ -126,10 +126,18 @@ def questionInitializeDisk(intf=None, name=None):
>>
>>> +                _("Error processing drive %s (%s  %-0.fMB).\n\n"
>>>                    "Maybe it needs to be reinitialized.  "
>>
>>> +                  "YOU WILL LOSE ALL DATA ON THIS DRIVE!")
>>> +                % (name, dev.model, dev.getSize(),),
>>
>> Just one little comment regarding disks on s390. With DASDs the model
>> and hence the size are unfortunately often the same for disks in the
>> system. Thus they would still not be distinguishable. What do you think
>> about adding the symlink name of /dev/disk/by-path as well? This would
>> unambiguously identify all kinds of disk (DASD and zFCP on s390).
> 
> The 'name' value should already contain that information, unless I'm
> mistaken.
> For example, on my system, I might see:
> 
>     Error processing drive sda (ATA SAMSUNG MMCRE64G  61052MB).
> 
> So you would still see dasda and so on for s390.

I'll take the autopartitioner drive list as an example of multiple ones similar to the one case you changed:

iw/autopart_type.py:487:PartitionTypeWindow::_fillBootStore()
        for disk in self.storage.disks:
            if disk.name not in self.anaconda.id.bootloader.drivelist:
                continue
            dispstr = "%s %8.0f MB %s" %(disk.name, disk.size, disk.model)

On s390 you often have multiple (if not many) drives of the same type and size, which looks like this if you have some "model 9" DASDs:

dasda 7043 MB (IBM S390 DASD drive)
dasdb 7043 MB (IBM S390 DASD drive)
dasdc 7043 MB (IBM S390 DASD drive)
dasdd 7043 MB (IBM S390 DASD drive)
...

If I add zFCP LUNs I get more entries like the following:

sda   5114 MB   IBM 2107900
sdb   5114 MB   IBM 2107900
sdc   5114 MB   IBM 2107900
sdd   5114 MB   IBM 2107900
...

Depending on the DASD parameter specified in parmfile or linuxrc.s390, and on FCP_* parameters specified in parmfile or zfcp parameter in kickstart or the "add zFCP" dialog for advanced storage configuration in the partitioner screen the drives are activated and get assigned kernel device names.

>From a usability point of view, they are no longer distinguishable (at least if they are all of the same type and size). Especially with many disk devices, the user can hardly remember the mapping he implicitly chose on activating the disks to deduce which actual drive is behind the kernel device name dasdX or sdX.

Therefore I was suggesting to (additionally?) use the disk device path as found as symlinks under /dev/disk/by-path. Then things would look like this:

dasda 7043 MB (IBM S390 DASD drive) ccw-0.0.beef
dasdb 7043 MB (IBM S390 DASD drive) ccw-0.0.abcd
dasdc 7043 MB (IBM S390 DASD drive) ccw-0.0.1234
dasdd 7043 MB (IBM S390 DASD drive) ccw-0.0.42fe
...
sda   5114 MB   IBM 2107900         ccw-0.0.3c1b-zfcp-0xaaaabbbb0300c562:0x401040eb00000000
sdb   5114 MB   IBM 2107900         ccw-0.0.3c1c-zfcp-0xaaaabbbb0300c562:0x401040eb00000000
sdc   5114 MB   IBM 2107900         ccw-0.0.4444-zfcp-0xaaaabbbb0300c042:0x4010beef00000000
sdd   5114 MB   IBM 2107900         ccw-0.0.5a6c-zfcp-0xaaaabbbb03000001:0x401a1a1a00000000
...

For dasd this includes the device number which is the main identification used on s390 for such drives.
For zFCP this includes the device number of the FCP adaptor, the WWPN, and the LUN which are the three main identification parts of FCP logical units.
For other architectures this includes whatever is necessary to uniquely identify which disk is behind the kernel device name.

Steffen

Linux on System z Development

IBM Deutschland Research & Development GmbH
Vorsitzender des Aufsichtsrats: Martin Jetter
Geschäftsführung: Erich Baier
Sitz der Gesellschaft: Böblingen
Registergericht: Amtsgericht Stuttgart, HRB 243294