Set linkStaleAge to 15 seconds: prevent link down event during mastership transition in HA scenario
Set allowDuplicatedIp to false: cleanup IP when a new container is launched with same IP but different MAC
Change-Id: Ic2a9f87179de367c2ff84b232ca7f8ca3e152715
This commit fixes the bug that prevent initial redirection from happening if the host is added without IP and get updated with an IP later
Change-Id: Ic4e90763e38eff94b1613d90f943f76f5285cf94
- Configurable pseudowires supporting untagged-untagged, single-single, double-double tagged traffic.
- Commands for listing, adding and removing pseudowires.
- Initial support for link failures.
- Pseudowires also configurable by network configuration.
- Tested with ofdpa_3.0.5.
Known limitations that I plan to fix soon :
- Adding pseudowires from configuration is inconvenient because we need to ammend new pws to the
existing configuration. We should create a REST API for adding/removing/listing pws and abandond the
network configuration.
- Spine fabric switches have rules matching a special mpls tag used for pw traffic for a specific leaf.
However, this rules redirect to an "indirect" group for forwarding traffic. If the resulting port there
is no mechanism as of now to handle this. We should use the MPLS ECMP Groups of ofdpa, however they are
not functional. Thus, we need to inject logic into the application to handle this case.
Change-Id: Ia85cf4514ebab627fc6ed5a19ad9f6cdc67dc24c
Matching VLAN_VID in TMAC table is mandatory on QMX switches
XGS switches support optional VLAN_VID matching so this patch is effective for both platforms
In addition,
- Take native vlan into account when assigning VLAN for untagged multicast traffic
- Fix a bug that puts wrong eth_type in TMAC table
Change-Id: I19caf2d6d69096a96b75cb528b6ed37b28d7d988
Instead of early return, simply ignore link-local IP matching if interface MAC is not configured (e.g. spine port)
Piggybacked in this commit:
- Refactor the ICMPv6 handler
Change-Id: I44e4daac067262b63e6726b4b4d7e3426e0f7fdb
Note: this is only a temporary fix that still doesn't offer full dynamic interface config capability
Change-Id: Ieb2a9316821d3e8a3ad9257b2980e14ddbafec4d
In addition
- Carry all alternatives in ROUTE_UPDATED and ROUTE_REMOVED event
- More unit tests
- More debug messages
- Fix routing table out-of-sync issue in MockRoutingService
- Fix populateRoute/revokeRoute mastership issue
Change-Id: I22d537625b570b09ecd4e22e6e14bb1ee27f8bcb
- reentrant lock was not being used correctly
- fixHashGroup in group handler was not updating global store correctly
- linkUp was not being noted in seenLinks if configuration came after switches connected
- serialization error in global objective store due to missing kryo for Sets
- damaged routepath computation was not taking pair-devs into account
- switch failures were leading to improper ecmpSpg graph updates, and missed hash-group changes
- implemented more next-objective verification as group sub-system can go out-of-sync with objective-store
Change-Id: If3cfdd715e9b69820894b49def31f75ceb748863
Two things:
- In dual (paired) ToR scenarios it is possible to have the same outport
in multiple buckets in a hash group, as long as they have different labels.
When adding buckets this was taken into account. But when removing buckets,
only outport was being checked. This bug fix ensures that labels are checked
as well when removing buckets.
- In dual ToR scenarios, getting the right set of hash buckets proved difficult
with existing 'retryHash' mechanism. Repealed and replaced with a bucket corrector
mechanism that periodically corrects the hash group buckets when the topology
has been stable for the last 10 secs. Required the introduction of a VERIFY
operation in Next Objectives. Also added a cli command to trigger this
operation manually.
Change-Id: Ib0d2734060fadc6e7a4bd0d75f3409e194413a97
Also include following refactoring
- Refactor the way we generate ARP probe
- Remove some unused code
Change-Id: I96b1c47bd5731b7b38ef4d19a941d231b5d0054c
