Using `sys.argv[0]` does not quite achieve the effect we want, namely
that any (test-case, $PYTHONHASHSEED) tuple should always default to
the same transport.
For example, the following two examples should always use the same
transport:
$ make PYTHONHASHSEED=1337 test-sh
infamy0:test$ ./case/ietf_system/hostname/test.py
$ make PYTHONHASHSEED=1337 test-sh
infamy0:test$ cd ./case/ietf_system/hostname
infamy0:hostname$ ./test.py
But, since their argv[0]'s are different, they don't.
Therefore use the _last two components_ of argv[0]'s full path
instead. This should:
- Spread the allocation over an entire suite run, since the
penultimate component is usually the test name.
- Keep the allocation stable when test code is modified (which is
common during debugging of failing tests)
- Avoid instabilities stemming from the local storage location (i.e.,
home directory names etc.)
The somewhat clumsy construct of resolving a logical port name to its
physical counterpart has been with us from Infamy's inception:
_, physical_port = env.ltop.xlate(logical_node, logical_port)
This is needlessly verbose. Since devices already have access to the
mappings which applies to it, make them easy to access:
target = env.attach("target")
# Get physical port names via the subscript operator
target["data0"]
* Use new cleanup in infamy to ensure the boot order gets restored
* Use new boot order in operational and use RPC to set boot order
to remove all SSH commands.
If a test need to clean up itself disregarding test status,
it can add an optional test-cleanup phase.
This is useful if the test in some circomstances leave the device in a
bad state. This can be used to restore the unit to its ordinary state.
Furl answers the question "is this URL serving the content I expect?"
Sometimes, we might be asking a server that in the middle of being
reconfigured, is crashing, is buggy, etc. - in which case it might RST
the connection. In that case the answer to the question is "False" and
not "BURN IT ALL DOWN!!"
If the callable threw an exception, then the process running inside
the netns would exit before sending back a value on the tx pipe, which
meant that `rx.recv()` would hang in the original process.
Make sure that we catch any exceptions, and send that instead of the
result over the pipe in those scenarios. Then in the original process,
re-raise the exception.
Fixes the following issue when deleting /infix-services:mdns/interfaces,
which works fine over RESTCONF:
Traceback (most recent call last):
File "/home/jocke/src/infix/test/./case/infix_services/mdns_allow_deny/test.py", line 142, in <module>
dut.delete_xpath("/infix-services:mdns/interfaces")
File "/home/jocke/src/infix/test/infamy/netconf.py", line 418, in delete_xpath
oldd = mod.parse_data_dict(old, no_state=True)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/root/.infix/venv/lib/python3.11/site-packages/libyang/schema.py", line 215, in parse_data_dict
return dict_to_dnode(
^^^^^^^^^^^^^^
File "/root/.infix/venv/lib/python3.11/site-packages/libyang/data.py", line 1277, in dict_to_dnode
result.root().validate(
File "/root/.infix/venv/lib/python3.11/site-packages/libyang/data.py", line 430, in validate
self.validate_all(no_state, validate_present)
File "/root/.infix/venv/lib/python3.11/site-packages/libyang/data.py", line 449, in validate_all
raise self.context.error("validation failed")
libyang.util.LibyangError: validation failed:
Invalid leafref value "e3" - no target instance "/if:interfaces/if:interface/if:name" with the same value.
The patch skips validation of the diff in favor of checking the result
of libyang.xpath_del().
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Replace the existing remove-module-names-pass in the RESTCONF
transport -- which only worked for top-level containers -- with
libyang's standard `.print_dict()` used by the NETCONF transport
implementation.
As a bouns, the implementation of `.get_iface()` is now transport
agnostic.
This patch adds latest symlinks to the curiOS containers to make system
upgrades easier. I.e., a user can now reference the bundled image with:
set image oci-archive:/lib/oci/curios-httpd-latest.tar.gz
So that when they upgrade to the latest Infix, which might include an
update of curiOS httpd, they will get a seamless upgrade also of the
container(s) running.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
This is a specialized namespace, containing two ports,
which (hopefully) acts as completely transparent "bump-on-the-wire"
bridge.
Useful in scenarios where you want to test what happens when the flow
of packets between two nodes is disrupted. E.g., for testing fail-over
behavior in dynamic routing protocols, L2 redundancy, VRRP etc.
Now that we're using the `passthru` mode of MACVLANs, it seems we can
rely on the parent interface's promiscuity going down to 0 as an
indication that the kernel has completed the removal of the interface,
and will accept the creation of a new one.
When used as a context manager, __exit__ will be called even though
__enter__ might throw an exception half way through. So there is no
guarantee that the instance is on the cleanup list when we get to
stop().
This mode let's us receive all packets from the parent device, which
is automatically set in promiscuous mode, and let's the MACVLAN
transmit packets with any SA, which we need to implement a TPMR (or a
bridge or anything else we might dream up)
This is needed to create namespaces with MACVLANs from different
parent interfaces. The first consumer will be a Two-Port Mac
Relay (TPMR), so that we can control the flow of packets between two
nodes.
Story time:
In the days of yore, we generated PCAPs using tcpdump, which will
flush any buffered packets to its output file on reception of
SIGUSR2. Therefore we sent a USR2 before terminating the capture.
At some point, we switched over to use tshark to capture packets for
us, but the SIGUSR2 sayed in place.
It turns out, tshark will just die on USR2, abandoning the dumpcap
child it has created, which means the enclosing network namespace
can't be properly torn down.
Therefore:
Remove skip sending USR2 and just send TERM directly.
SVG is a plain text file, better to see why it changed
than binary PNG. Also fix#744 by adding infix/OEM hash
to test specification
Signed-off-by: Mattias Walström <lazzer@gmail.com>
Mattias Walström