For some reason 
>>> 'µ'.upper().lower() == 'µ'
False

Note, that
>>> ord('µ')
181

there is another Mu like symbol chr(956), on which this code passes

So, yes, unicode table seems to be a bit inconsistent. In str documentation it worth to emphasize that something.upper().lower() in general not equals to something.

Here is a little more detail on what is occurring:

>>> from unicodedata import *
>>> c = 'µ'
>>> for s in (c, c.lower(), c.upper(), c.lower().upper(), c.upper().lower()):
	print(s, ord(s), name(s), category(s))
	
µ 181 MICRO SIGN Ll
µ 181 MICRO SIGN Ll
Μ 924 GREEK CAPITAL LETTER MU Lu
Μ 924 GREEK CAPITAL LETTER MU Lu
μ 956 GREEK SMALL LETTER MU Ll

Conclusion made: not a bug (not a python bug at least).