A new test (dot1x) and a new tool (wpa) which was implemented to test 802.1X availability #1093
Conversation
|
I've begun a review; expect a flurry of comments. |
|
Looking forward to it. :-) |
There was a problem hiding this comment.
This is headed in the right direction. Pretty much everything I found in the code is minor.
Additionally:
Replace all occurrences of the old test name to the new one. (There are a few stray references to trace.)
Our convention for naming JSON object members is to separate words with hyphens (e.g., foo-bar rather than foo_bar).
Name the test to be more-specific (e.g., 802.1x or ieee802.1x). Further comments on the test to follow in #1093.
If the driver can be determined based on system information about the interface, remove that parameter from the test and let the tool take care of it.
The test is now renamed to 802.1x. We don't use the driver, so I removed that parameter. |
…ol phase1 and phase2 settings for WPA Supplicant. * Don't pre-populate results we might not need if the test fails. * Print whether we got an address or not in the debug log.
…assid for example we can be sure that the authenticaion was successful. Same goes for IP addresses.
Signed-off-by: Kristofer Hallin <[email protected]>
|
You guys are probably way past this, but I thought I'd share anyway:
https://github.com/perfsonar/pscheduler/tree/master/scripts/PDK
There are some MD files there with instructions, as well as the templating
scripts. Just FYI.
Thanks,
…-Ed
On Fri, Dec 4, 2020 at 9:53 AM Mark Feit ***@***.***> wrote:
***@***.**** commented on this pull request.
------------------------------
In pscheduler-test-dot1x/dot1x/validate.py
<#1093 (comment)>:
> +
+#
+# Validator for "trace" Test
+#
+
+from pscheduler import json_validate
+
+
+def spec_is_valid(json):
+ schema = {
+ "type": "object",
+ "properties": {
+ "schema": {"$ref": "#/pScheduler/Cardinal"},
+ "timeout": {"$ref": "#/pScheduler/Duration"},
+ "interface": {"$ref": "#/pScheduler/String"},
+ "driver": {"$ref": "#/pScheduler/String"},
Sure. Anything that has a fixed set of valid values (e.g., foo and bar
only) should be an enumerated string type that only allows those values.
See how the rtt test's protocol field is validated as an example.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#1093 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ABLYWCYBNWZNOTMWXRPH533STDZXHANCNFSM4UKRGRSA>
.
|
|
I'll try to test this
Thanks for sharing, thats very useful. The plugin_dev script could have saved me quite some time. |
|
Thanks for the feedback, I've tried to fix it and improve thing in general. Also, I'll try to test this thoroughly to make sure everything is still working, there might be some regressions I've missed. |
|
Note for later: 4.4.0 will be making some minor changes to the file layout of packages that will have to be done for this package when it is pulled in. |
Initial pull request for 802.1X test and tool developed at SUNET Sweden. This was initially developed on the 4.3 branch using Python 2.
A few issues I can see right away:
The wpa tool is responsible of both doing authentication with WPA Supplicant and acquire an address using dhclient. We might want to split this into two different tools or rename it.
This requires WPA Supplicant to be configured on the host system with D-Bus support enabled since we are using wpa_cli to configure things. I don't know how things like that is handled for other tools.
When running this, you really want to run it in a separate network namespace, otherwise dhclient will replace the default route if an address is acquired and you risk loosing connectivity to the machine.
Any suggestions for improvements are very welcome.
/Kristofer - SUNET