Mcast population during cluster reboot 73/84373/8
authormanojna v <manojna.vijayakrishna@ericsson.com>
Fri, 13 Sep 2019 11:18:13 +0000 (16:48 +0530)
committerChetan Arakere Gowdru <chetan.arakere@altencalsoftlabs.com>
Thu, 3 Oct 2019 06:18:32 +0000 (06:18 +0000)
commite60734472e9b58a2358fce1c80d7b0e1f8e0c314
treef165d1eebd5a4c8efb951d552e0822d59a626f5c
parenta86dd4ab5f0652fad745b124ecfdb4e2ed02872d
Mcast population during cluster reboot

these changes make sure that during cluster reboot scenario,
mcast remains intact as it was before cluster reboot.

as the elan interfaces gets replayed mcast gets reconstructed from
scratch and as the final elan interface is added mcast will become
identical to what it was before cluster reboot.

During this period where mcast starts with 1 bucket and reaches to
the correct no of buckets the broadcast traffic will be affected.

while doing mcast update, instead of replacing the group with new set
of buckets , first read existing buckets and add the new buckets to
existing buckets.

Change-Id: I203174ff845eb66a3a6fc685bbbd74966f91e718
Signed-off-by: manojna v <manojna.vijayakrishna@ericsson.com>
elanmanager/impl/src/main/java/org/opendaylight/netvirt/elan/cache/ConfigMcastCache.java [new file with mode: 0644]
elanmanager/impl/src/main/java/org/opendaylight/netvirt/elan/l2gw/jobs/McastUpdateJob.java
elanmanager/impl/src/main/java/org/opendaylight/netvirt/elan/l2gw/utils/ElanL2GatewayMulticastUtils.java
elanmanager/impl/src/main/java/org/opendaylight/netvirt/elan/l2gw/utils/ElanRefUtil.java