I have a problem of TCP Recv-Q size getting increased after the
subagent sends a SNMP trap to master agent on Net-SNMP platform.
Because of this eventually the subagent stops sending traps further.
Can anyone plz tell me what could be the reason of increment of the
size of Recv-Q of the source socket? plz see below the result of
netstat command.
How can we stop the increment of the size of the Recv-Q?
Please advise.
[root@nas1 app]# tcpdump -i lo port 34968 -s 1500 -w /var/log/
dump1.txt
tcpdump: listening on lo
6 packets received by filter
0 packets dropped by kernel
[root@nas1 app]# tcpdump -v -r /var/log/dump1.txt
08:57:31.379656 localhost.localdomain.34968 > localhost.localdomain.
1055: P [bad
tcp cksum 5bcc!] 2645605356:2645606100(744) ack 2646798913 win 32767
<nop,nop,t
imestamp 29220337 29158580> (DF) [ttl 0] (id 56258, len 796)
08:57:31.379993 localhost.localdomain.1055 > localhost.localdomain.
34968: P [bad
tcp cksum 9416!] 1:753(752) ack 744 win 32767 <nop,nop,timestamp
29220337 29220
337> (DF) [ttl 0] (id 36242, len 804)
08:57:31.380043 localhost.localdomain.34968 > localhost.localdomain.
1055: . [tcp
sum ok] ack 753 win 32767 <nop,nop,timestamp 29220338 29220337> (DF)
[ttl 0] (i
d 56259, len 52)
Asresult receiving queue is growing , getting overflowed and we stop
sending SNMP traps to SMU.
[root@nas1 app]# netstat -ap | grep 1055
tcp 0 0 localhost.localdom:1055 *:*
LISTEN
28586/snmpd
tcp 2640 0 localhost.localdo:34968 localhost.localdom:1055
ESTABLISHED
28633/proSWagntdGSA
tcp 0 0 localhost.localdom:1055 localhost.localdo:34968
ESTABLISHED
28586/snmpd
Thanks