upgrade-env fails when LC_ALL is empty

[sryabin]

Step for reproduce:

[root@fuel bin]# LC_ALL="" octane upgrade-env 1
DEPRECATION WARNING: /etc/fuel/client/config.yaml exists and will be used as the source for settings. This behavior is deprecated. Please specify the path to your custom settings file in the FUELCLIENT_CUSTOM_SETTINGS environment variable.
2015-09-25 12:48:38 INFO octane.util.ssh Creating new SSH connection to node 2
2015-09-25 12:48:38 INFO paramiko.transport Connected (version 2.0, client OpenSSH_6.6.1p1)
2015-09-25 12:48:39 INFO paramiko.transport Authentication (publickey) successful!
2015-09-25 12:48:39 INFO paramiko.transport Secsh channel 1 opened.
2015-09-25 12:48:39 INFO octane.util.subprocess Started process bash[at node-2]: bash -c '. /root/openrc;' 'keystone tenant-list | ' 'awk -F\| '"'"'\$2 ~ /id/{print \$3}'"'"' | tr -d \ '
2015-09-25 12:48:39 INFO octane.util.subprocess Process bash finished with return value 0
Traceback (most recent call last):
  File "/usr/lib64/python2.6/logging/__init__.py", line 799, in emit
    stream.write(fs % msg.encode("UTF-8"))
UnicodeDecodeError: 'ascii' codec can't decode byte 0xd0 in position 63: ordinal not in range(128)

[Yuriy Taraday (yorik-sar)]

Here's what leads to this problem:

- some IOError happens;
- IOError gets string associated with errno value based on current locale (which just happend to be ru_RU.UTF-8), gets UTF-8-encoded str object;
- cliff catches this IOError in "err" variable and does "log.error(err)";
- some dozen layers down we end up in StreamHandler.emit method;
- it calls "formatter.format(msg)" (msg is still IOError here) and gets that str object;
- seeing that it's not a unicode, emit() then does "'%s\n' % msg" which succeeds in adding a newline to the string;
- then it passes it to stream.write();
- stream has 'UTF-8' encoding set, so first thing it does is "self.encode(object, self.errors)" which expects object to be a unicode string, but it's not, so it tries implicitly convert it with ASCII encdoding and blows up;
- emit() has nothing better to do but to try "msg.encode('utf-8')" in exception handler, which obviously blows up as well;
- we have new bug in Launchpad.

We should probably convert all messages passed to emit() to Unicode with UTF-8 encoding (or current system encoding) ourselves.

[Dmitry Pyzhov (dpyzhov)]

We passed SCF for 8.0. Moving bug to 9.0.

[Bug Checker Bot (bug-checker)]

(This check performed automatically)
Please, make sure that bug description contains the following sections filled in with the appropriate data related to the bug you are describing:

actual result

expected result

steps to reproduce

For more detailed information on the contents of each of the listed sections see https://wiki.openstack.org/wiki/Fuel/How_to_contribute#Here_is_how_you_file_a_bug