vap_profile.py updated to allow the mac address to be used and pass in
debug state on enstanciation of LFRequest
modify_vap.py : added logging, pass debug flag to base classes
Signed-off-by: Chuck SmileyRekiere <chuck.smileyrekiere@candelatech.com>
--port created a conflict with lf_attend_mod_test.py The --port was introduced
a month ago. Will look for other conflicts in the regressions.
Changed lf_attend_mod_test.py to add back in --port
Signed-off-by: Chuck SmileyRekiere <chuck.smileyrekiere@candelatech.com>
More changes all around regression test logic related to the
vap profile. Move some open-coded admin_up/admin_down calls
to base class.
regression test takes the .rc file as primary input,
and cmd-line args will override that.
Signed-off-by: Ben Greear <greearb@candelatech.com>
Add some more notes about external dependencies for this
script. It is meant to run with already-created vlans,
for instance.
Signed-off-by: Ben Greear <greearb@candelatech.com>
On a narrow terminal it is difficult to follow the output. This merges a
series of message lines into one debug output call to help legibility
Signed-off-by: Jed Reynolds <jed@bitratchet.com>
Previous code also insisted that there was at least one success,
but that means that code doing not pass/fail tests would
fail.
Signed-off-by: Ben Greear <greearb@candelatech.com>
This was triggered by a call to admin_up() that was passing self.lfclient_url as an un-named parameter, creating
unanswerable requests formatted like http://localhost:8080/http:/localhost:8080/. There should really only be
one http[s]*:// pattern in a URL, and further instances of colons should be URL encoded to %3A.
Signed-off-by: Jed Reynolds <jed@bitratchet.com>
Logic was broken trying to compare json_response['down'] to "true",
when instead that object is a Boolean variable. So this method
always returned immediately instead of actually waiting until port
was admin up. This was seen testing VAP profile.
Signed-off-by: Ben Greear <greearb@candelatech.com>
Make it look more like macvlans, including logic to set IP addresses,
and to verify qvlans were created properly.
Signed-off-by: Ben Greear <greearb@candelatech.com>
If port cannot be found, then it is down. This fixes race with
creating port and then immediately trying to way for it to go
admin up.
Signed-off-by: Ben Greear <greearb@candelatech.com>
The generic endpoint profile, especially the monitoring code,
was wrong in lots of ways.
Fix this, improve pass/fail and error reporting, and other
improvements for the gen_profile test.
Probably not backwards compat, but I assume due to breakage
no one was using this code anyway.
Signed-off-by: Ben Greear <greearb@candelatech.com>
Let user specify the level of logging desired.
Test implementaiton in example_security_connection.py
Signed-off-by: Ben Greear <greearb@candelatech.com>
In station_profile create method, pay attention to result of
wait_until_ports_appear. If it returns False, we know creation
has failed somehow, so we do not need to do the extra checking
that was in the create() method.
create_l3_stations adds logic to check return code of station
create logic, and exit with error code if it could not
create them as expected.
Signed-off-by: Ben Greear <greearb@candelatech.com>
LFUtils.wait_until_ports_appear: Improve debug by printing existing stations and adding more information to debug string
Signed-off-by: Matthew Stidham <stidmatt@gmail.com>
Default to treat station list as list of EIDS, and use it's
shelf/resource info instead of using resource 1 by default.
Caller may still use resource_id=X if they wish too, and it
will work as it did previously.
Add loop timeout (default is 300 tries with 1-sec sleep, so 5+
minutes). This keeps the method from spinning forever if port
cannot come up.
Signed-off-by: Ben Greear <greearb@candelatech.com>
