[blivet:master 17/17] Make TmpFS._resizable False.

David Lehman dlehman at redhat.com
Wed Jan 7 19:12:48 UTC 2015


On 01/07/2015 01:00 PM, Martin Kolman wrote:
> On Tue, 2015-01-06 at 15:38 -0500, mulhern wrote:
>> This doesn't change current operation at all, since it will be set to False
>> in constructor anyway. But even if it weren't there are some problems:
>>
>> * It will crash in FS.doResize() due to lack of resizefsUnit field.
>> I have another patch set in which I address that, however...
>> * Most filesystems do not need to be mounted to be resized, but tmpfs
>> does because it uses mount. This means it does not fit well into the
>> framework of resizing actions and so forth that the other filesystems
>> fit into. This problem needs to be thought about fairly deeply.
>> * TmpFS is not really able to find its current size. This is a tricky thing
>> to do; it seems people tend to use some combination of du and df.
> I think they probably know what they are doing when using blivet to
> resize a volume. If we followed the same logic we would need to disable
> resize for btrfs as well, because df also often reports very misleading
> values with df in many cases due to snapshots, reflinks, metadata block
> exhaustion and general madness.

We don't support resize of btrfs :)


More information about the anaconda-patches mailing list