When we use tcpdump command on a particular interface it is flapping.
Posted: 2018/02/05 09:17:41
Ingress in the name of the interface which we are using in our environment.
Summary:
The ingress flapping was happening when we use tcpdump command on that particular interface.
tcpdump logs ended with :
tcpdump: pcap_loop: The interface went down
while running this command on the particular OA, the ingress interface was flapping. After stopping the command, it goes back to normal state UP.
Issue might be on the RPM libpcap-1.5.3-9.el7.x86_64 used by tcpdump to capture packets which causes the packet-loss-when-using-pcap-loop
Observations:
tcpdump: pcap_loop: The interface went down
79 packets captured
99 packets received by filter
0 packets dropped by kernel
And the Ingress interface was flapping up/down
oa6.chn6us1c1.cdn ~# ip a | grep ingress
9: ingress@egress: <BROADCAST,MULTICAST> mtu 9000 qdisc noqueue state DOWN qlen 1000
oa6.chn6us1c1.cdn ~# ip a | grep ingress
9: ingress@egress: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc noqueue state UP qlen 1000
inet 10.50.42.86/24 brd 10.50.42.255 scope global ingress
Because of flapping of ingress interface, more number of cc errors shows in logs
# tail -f /usr/cachelogic/log/vxoa_cook_live_info.log
2018-01-26T15:52:39.242Z [Ing10:1388] INF [muxid=1000069/515K/udp://235.7.8.41:10007;ingress] 75 CC discontinuities in the last 10
Is this any kind of Bug or something present in the package "libpcap-1.5.3-9.el7.x86_64"? This issue was not reported in previous versions of the package. We face this issue only after upgrading the system to the above mentioned package version.
Summary:
The ingress flapping was happening when we use tcpdump command on that particular interface.
tcpdump logs ended with :
tcpdump: pcap_loop: The interface went down
while running this command on the particular OA, the ingress interface was flapping. After stopping the command, it goes back to normal state UP.
Issue might be on the RPM libpcap-1.5.3-9.el7.x86_64 used by tcpdump to capture packets which causes the packet-loss-when-using-pcap-loop
Observations:
tcpdump: pcap_loop: The interface went down
79 packets captured
99 packets received by filter
0 packets dropped by kernel
And the Ingress interface was flapping up/down
oa6.chn6us1c1.cdn ~# ip a | grep ingress
9: ingress@egress: <BROADCAST,MULTICAST> mtu 9000 qdisc noqueue state DOWN qlen 1000
oa6.chn6us1c1.cdn ~# ip a | grep ingress
9: ingress@egress: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc noqueue state UP qlen 1000
inet 10.50.42.86/24 brd 10.50.42.255 scope global ingress
Because of flapping of ingress interface, more number of cc errors shows in logs
# tail -f /usr/cachelogic/log/vxoa_cook_live_info.log
2018-01-26T15:52:39.242Z [Ing10:1388] INF [muxid=1000069/515K/udp://235.7.8.41:10007;ingress] 75 CC discontinuities in the last 10
Is this any kind of Bug or something present in the package "libpcap-1.5.3-9.el7.x86_64"? This issue was not reported in previous versions of the package. We face this issue only after upgrading the system to the above mentioned package version.