Is it appropriate to ask colleagues opinions when reviewing a paper?

I think that this practice violates the usual "confidentiality agreement" that presents (formally, or customary) in a review process. However, I would agree that this is a common practice, and for some people it is just hard to work on their own, although technically you are supposed to give your own opinion on a paper you review.

I would say that if you really need to show the paper you review to someone else, you should check the following boxes:

1. make sure the person understands that they are looking on a paper under review, and agrees to maintain the confidentiality, i.e.: not to talk about the ideas from the paper with someone else, not to produce own work based on these results until they are made public, etc.
2. The actual material paper, and the file, do not change hands: do not send the pdf manuscript to your colleagues, do not leave the printed paper with them for a while.
3. Ask specific question(s) about the paper, for which you need second advice, not just a general opinion. Remember, that you should review the manuscript yourself. Definitely do not use the second opinion to shake some work off your shoulders and go.

The safest course of action would be to ask the editor: "My colleague Edna Krabapple has relevant expertise that would help me understand the paper and write a better report. May I share and discuss the manuscript with her?"

If the editor says no, you'd be justified in replying "Since I don't have the expertise to fully understand the paper on my own, I must decline the invitation to review."

This varies by field. In computer science, the members of the "program committee" of a conference are each responsible for reviewing some number of the submissions. However, they also have the power to unilaterally select sub-reviewers (sometimes more than one per paper) to provide input. How this is actually handled varies by venue, but sometimes can be as informal as sending an email to a colleague or graduate student asking them to take a look at the paper and provide input to the committee.