R4.1-build36 vrouter agent core seen @ NHKSyncEntry::NHKSyncEntry(NHKSyncObject*, NextHop const*)

[Sudheendra Rao]

On build36 R4.1 CB mitaka, vrouter-agent core is seen during sanity:

(gdb) bt
#0 0x00007f3f75dc3c37 in raise () from /lib/x86_64-linux-gnu/libc.so.6
#1 0x00007f3f75dc7028 in abort () from /lib/x86_64-linux-gnu/libc.so.6
#2 0x00007f3f75dbcbf6 in ?? () from /lib/x86_64-linux-gnu/libc.so.6
#3 0x00007f3f75dbcca2 in __assert_fail () from /lib/x86_64-linux-gnu/libc.so.6
#4 0x0000000000f8b4b9 in NHKSyncEntry::NHKSyncEntry(NHKSyncObject*, NextHop const*) ()
#5 0x0000000000f8685b in MplsKSyncEntry::Sync(DBEntry*) ()
#6 0x00000000010919b5 in KSyncDBObject::Notify(DBTablePartBase*, DBEntryBase*) ()
#7 0x0000000001474c7a in DBTableBase::RunNotify(DBTablePartBase*, DBEntryBase*) ()
#8 0x00000000014778f8 in DBTablePartBase::RunNotify() ()
#9 0x000000000147346e in DBPartition::QueueRunner::Run() ()
#10 0x00000000015c7157 in TaskImpl::execute() ()
#11 0x00007f3f76992b3a in ?? () from /usr/lib/libtbb.so.2
#12 0x00007f3f7698e816 in ?? () from /usr/lib/libtbb.so.2
#13 0x00007f3f7698df4b in ?? () from /usr/lib/libtbb.so.2
#14 0x00007f3f7698a0ff in ?? () from /usr/lib/libtbb.so.2
#15 0x00007f3f7698a2f9 in ?? () from /usr/lib/libtbb.so.2
#16 0x00007f3f76bae184 in start_thread () from /lib/x86_64-linux-gnu/libpthread.so.0
#17 0x00007f3f75e8737d in clone () from /lib/x86_64-linux-gnu/libc.so.6
(gdb)

The core file is copied under:
/cs-shared/bugs/<bug-id>/

[OpenContrail Admin (ci-admin-f)]

Review in progress for https://review.opencontrail.org/37344
Submitter: Manish Singh (<email address hidden>)

[OpenContrail Admin (ci-admin-f)]

Review in progress for https://review.opencontrail.org/37493
Submitter: Manish Singh (<email address hidden>)

[OpenContrail Admin (ci-admin-f)]

Review in progress for https://review.opencontrail.org/37493
Submitter: Hari Prasad Killi (<email address hidden>)

[OpenContrail Admin (ci-admin-f)]

Review in progress for https://review.opencontrail.org/37493
Submitter: Manish Singh (<email address hidden>)

[OpenContrail Admin (ci-admin-f)]

Review in progress for https://review.opencontrail.org/37710
Submitter: Manish Singh (<email address hidden>)

[OpenContrail Admin (ci-admin-f)]

Review in progress for https://review.opencontrail.org/37728
Submitter: Manish Singh (<email address hidden>)

[OpenContrail Admin (ci-admin-f)]

Review in progress for https://review.opencontrail.org/37710
Submitter: Manish Singh (<email address hidden>)

[OpenContrail Admin (ci-admin-f)]

Reviewed: https://review.opencontrail.org/37710
Committed: http://github.com/Juniper/contrail-controller/commit/dc8557864e15ea3479cb2103dda2aa6a43090874
Submitter: Zuul (<email address hidden>)
Branch: R4.1

commit dc8557864e15ea3479cb2103dda2aa6a43090874
Author: Manish <email address hidden>
Date: Thu Nov 9 12:02:16 2017 +0530

Invalid nexthop in use by mpls.

Allow only multicast peer to rebake all mcast related labels.
In path sync issued for all paths in multicast route, only multicast peer path
should be responsible for rebaking labels and rest should ignore it. This makes
sure all label point to common mcast NH.

Fix is to store labels (both evpn and fmg) in multicast path and rebake for both
of them. Dont only rely on AgentPath label. Multicast Label is derived out of
AgentPath.

Change-Id: Ic3173dea65851e47608425b77997b6490de733e0
partial-bug: #1730191

[OpenContrail Admin (ci-admin-f)]

Review in progress for https://review.opencontrail.org/37960
Submitter: Hari Prasad Killi (<email address hidden>)

[OpenContrail Admin (ci-admin-f)]

Reviewed: https://review.opencontrail.org/37960
Committed: http://github.com/Juniper/contrail-controller/commit/420602d0001c2013c50a518824612fd6419fa1b4
Submitter: Zuul (<email address hidden>)
Branch: master

commit 420602d0001c2013c50a518824612fd6419fa1b4
Author: Manish <email address hidden>
Date: Thu Nov 9 12:02:16 2017 +0530

Invalid nexthop in use by mpls.

Allow only multicast peer to rebake all mcast related labels.
In path sync issued for all paths in multicast route, only multicast peer path
should be responsible for rebaking labels and rest should ignore it. This makes
sure all label point to common mcast NH.

Fix is to store labels (both evpn and fmg) in multicast path and rebake for both
of them. Dont only rely on AgentPath label. Multicast Label is derived out of
AgentPath.

Conflicts:
 src/vnsw/agent/oper/agent_path.cc

closes-bug: #1730191
(cherry picked from dc8557864e15ea3479cb2103dda2aa6a43090874)

Change-Id: I521321a2e45886861f05cbd2f22ed63d65fcfd15