bgpd on Edge nodes sometimes crashes and BGP goes temporarily down

Products

• VMware NSX Networking

Issue/Introduction

• You configure BGP on a Tier-0 Gateway
• bgpd on the Edge nodes sometimes crashes and BGP goes temporarily down.
• You might find a coredump as /var/log/core/core.bgpd.*****.gz if coredump is enabled.
• Just before the crash, "show bgp vrf all neighbors json" is logged.

/var/log/syslog

<YYYY-MM-DD>T<HH:MM:SS.SSS>Z <HostName> NSX 2606 - [nsx@6876 comp="nsx-edge" subcomp="agg-service" tid="2944" level="INFO"] ExecCmd would call: /usr/bin/timeout -k 10 10 /usr/bin/vtysh -c "show bgp vrf all neighbors json"

If core dump is disabled,
<YYYY-MM-DD>T<HH:MM:SS.SSS>Z <HostName> NSX 19837 - [nsx@6876 comp="nsx-edge" subcomp="node-mgmt" username="root" level="WARNING"] Process 'bgpd' has crashed but core dumps are disabled

If core dump is enabled,
<YYYY-MM-DD>T<HH:MM:SS.SSS>Z <HostName> NSX 29556 - [nsx@6876 comp="nsx-edge" subcomp="node-mgmt" username="root" level="WARNING"] Core file generated: /var/log/core/core.bgpd.<nnnnnnnnnn>.<nnnnn>.160.6.gz

Environment

VMware NSX-T Datacenter 3.x

VMware NSX 4.0

Cause

NSX Manager periodically invokes "show bgp vrf all neighbors json" on Edge nodes to get BGP status.

There is a known issue that bgpd sometimes crashes when "show bgp vrf all neighbors json" is invoked.

Resolution

Fixed in NSX-T 3.2.3 and NSX 4.0.1.1.

Additional Information

Additional configuration is required to enable coredump of bgpd before NSX-T 3.2.2.1 and NSX 4.1.1.

* NSX-T edge nodes manual core dump generation steps for frr
https://knowledge.broadcom.com/external/article?articleNumber=322548