SR-IOV: match on PCI address, don't do runtime config #31
Conversation
fnordahl
force-pushed
the
bug/1908351
branch
3 times, most recently
from
8b3f34e
to
2ca8bc6
Compare
lib/charms/ovn_charm.py
Outdated
| # configuration of Virtual Functions (VFs) in the system. | ||
| # | ||
| # The charm does not do run-time configuration of VFs as this | ||
| # would be detrimental to any instances consuming the VFs. In some | ||
| # confiugrations it would also break NIC firmware LP: #1908351. |
There was a problem hiding this comment.
s/confiugrations/configurations/
| @@ -61,7 +61,7 @@ def test_dpdk_device(self): | |||
| self.assertEquals(self.target.dpdk_device.driver, 'fakedriver') | |||
|
|
|||
| def test_sriov_device(self): | |||
| self.assertDictEqual(self.target.sriov_device, | |||
| self.assertDictEqual(self.target.sriov_device(), | |||
There was a problem hiding this comment.
This is testing the legacy behavior. I guess we could have one more such tiny test covering the new get_map()?
There was a problem hiding this comment.
So, testing the innards of the SRIOV context belongs to the unit tests in charm helpers. If anything we should probably remove this test here.
There was a problem hiding this comment.
Yes I thought about that, on the other hand this kind of tests is good to catch non-backward-compatible changes done to charmhelpers. You could imagine someone changing the behavior of the called function and changing the corresponding test in charmhelpers, so that it's passing there in charmhelpers and lands. It would then turn red here and make obvious that something has to be done in this layer.
I don't mind too much though, as you like.
There was a problem hiding this comment.
The test mocks the context in its entirety, so all it tests is that when you access the property you get a object that comes from a given location. The return value is defined in the test itself.
Populate the sriov-netplan-shim interfaces.yaml config with PCI address of interfaces. This is necessary as the VFs may be configured at a time in the boot sequence where interface naming has not settled yet. Stop doing runtime reconfiguration of SR-IOV, in addition to being detrimental to any virtual machine instance consuming the VF this will break NIC firmware in some configurations. Merge after: juju/charm-helpers#549 Closes-Bug: #1908351
fnordahl
force-pushed
the
bug/1908351
branch
from
2ca8bc6
to
dfe80a3
Compare
Populate the sriov-netplan-shim interfaces.yaml config with PCI address of interfaces. This is necessary as the VFs may be configured at a time in the boot sequence where interface naming has not settled yet. Stop doing runtime reconfiguration of SR-IOV, in addition to being detrimental to any virtual machine instance consuming the VF this will break NIC firmware in some configurations. Merge after: juju/charm-helpers#549 openstack-charmers#31 Change-Id: I397a2a82501c565724d6fce6b49f7b68d57d47fd Closes-Bug: #1908351
Populate the sriov-netplan-shim interfaces.yaml config with PCI
address of interfaces. This is necessary as the VFs may be
configured at a time in the boot sequence where interface naming
has not settled yet.
Stop doing runtime reconfiguration of SR-IOV, in addition to being
detrimental to any virtual machine instance consuming the VF this
will break NIC firmware in some configurations.
Merge after:
juju/charm-helpers#549
Closes-Bug: #1908351