Fix healthcheck.py script to accept the loopback IP

[charlescng]

IPs that are considered loopback IPs were not added to the addresses list. This caused the healthcheck.py script to not run.

@thomas-mangin

[coveralls]

Coverage remained the same at 45.663% when pulling 29d3227 on charlescng:master into 933d9ad on Exa-Networks:master.

[thomas-mangin]

Thank you very much for this patch, I will wait for @vincentbernat to comment before pulling.

[vincentbernat]

Well, I don't understand the problem nor the fix. is_loopback is true for 127.0.0.0/8-like addresses. We don't want to advertise those. You either need to add some non-loopback addresses on lo or use --ip to tell which IP you want to advertise.

[charlescng]

It was failing during the loopback IP retrieval.

processes/healthcheck.py --config healthcheck.conf --ip 192.168.99.2
DEBUG[healthcheck] Retrieve loopback IP addresses
ERROR[healthcheck] Uncatched exception: No loopback IP found
Traceback (most recent call last):
  File "processes/healthcheck.py", line 428, in <module>
    setup_ips(options.ips, options.label)
  File "processes/healthcheck.py", line 248, in setup_ips
    existing = set(loopback_ips(label))
  File "processes/healthcheck.py", line 242, in loopback_ips
    raise RuntimeError("No loopback IP found")
RuntimeError: No loopback IP found

I did not run into this when adding the --no-ip-setup flag to my setup, which is what I want to do regardless.

I don't think I quite follow the method's intentions but close this PR if you're happy with the code as-is.

Thanks for looking into it.

[vincentbernat]

OK, I understand the problem better. I have opened another PR (#217) with the appropriate fix.