PIM: Implement static IGMP joins without an IGMP report #16450
Conversation
nabahr
force-pushed
the
static_joins
branch
4 times, most recently
from
e3653c1 to
b690c0d
Compare
|
ci:rerun |
|
@donaldsharp I agree with the comments here. I would say most of them are as a result of mirroring the static join code which was old/wrong to begin with. Should I go ahead and update the static join parts as well or just leave it alone for now? |
|
I would say update where we can, unfortunately |
|
for the yang stuff let's double check my thoughts with the more yang minded people in our group. I've already asked them to do so. |
|
Looks like the group address in the nb validation does check for reserved group addresses but does not verify it's a multicast address, but the group address type in the yang is |
This will add a static IGMP group that does not rely on an underlying socket join which sends traffic to the cpu unneccesarily. Instead, the groups are joined directly without any IGMP interactions. New command is under interfaces, 'ip igmp static-group ...'. Added an alias for 'ip igmp join ...' to 'ip igmp join-group'. Moved IGMP join groups to new yang list "join-group" and reused the "static-group" list for the IGMP static groups. Signed-off-by: Nathan Bahr <[email protected]>
Signed-off-by: Nathan Bahr <[email protected]>
Copied the existing "join-group" test and modified to test static groups instead. Functionally the same but without IGMP reports. Signed-off-by: Nathan Bahr <[email protected]>
Implement a new
ip igmp static-group ...command separate from theip igmp join ...command. The latter command joins the group via a local socket and generates an IGMP report. The new static-group command will join the group but bypasses the IGMP reports. This is analogous to the Cisco static-group command.Hide the previous
ip igmp join ...command and added a new alias forip igmp join-group ....Updated documentation for the new command and alias.
Added show command for the static-groups separate from the join-groups.