You can use a list-comprehension with enumerate(). With thanks to @MikeScotty, a performance improvement would be to calculate the smallest value first.
Here's the code:
mn = min(A)
[i for i,e in enumerate(A) if e == mn]
which gives:
[2, 5]
which are the indexes of the 1s in A - not [2, 8]
To prove this is faster, here's the minx wrapper:
>>> def minx(l):
...     print("called")
...     return min(l)
... 
>>> [i for i,e in enumerate(A) if e == minx(A)]
called
called
called
called
called
called
called
called
called
[2, 5]
And some timings using timeit:
>>> timeit.timeit("[i for i,e in enumerate(A) if e == min(A)]", globals=locals())
5.9568054789997404
>>> timeit.timeit("[i for i,e in enumerate(A) if e == 1]", globals=locals())
1.397674421001284