网络故障原因，超乎想象

dba对网络知识一般比较薄弱，但影响巨大，如何诊断网络故障呢？


诊断方向：
配置了防火墙
1.网线/网卡/交换机问题
2.UDP receive(rx) buffer sizes设置过小/UDP buffer socket溢出
3.私网性能差并且CPU使用率高，`netstat -s` 出现packet reassembly failures。
4.网络传输坏块（corruption）导致的UDP checksum errors 和/或 send (tx) / receive (rx) transmission errors
5.在通信通道中设置了不匹配的MTU的值

共有26个

SELECT sysdate snap_time,a.inst_ID "INSTANCE", A.VALUE "GC BLOCKS LOST",
 B.VALUE "GC CUR BLOCKS SERVED",
 C.VALUE "GC CR BLOCKS SERVED",
 round(A.VALUE/(B.VALUE+C.VALUE+0.01),2) RATIO
 FROM GV$SYSSTAT A, GV$SYSSTAT B, GV$SYSSTAT C
 WHERE A.NAME='gc blocks lost' AND
 B.NAME='gc current blocks served' AND
 C.NAME='gc cr blocks served' and
 B.INST_ID=a.inst_id AND
  C.INST_ID = a.inst_id
 order by 2;

select a.inst_id "instance", a.value "gc blocks lost",
b.value "gc cur blocks served",
c.value "gc cr blocks served",
a.value/(b.value+c.value+0.01) ratio
from gv$sysstat a, gv$sysstat b, gv$sysstat c
where a.name='gc blocks lost' and
b.name='gc current blocks served' and
c.name='gc cr blocks served' and
b.inst_id=a.inst_id and
c.inst_id = a.inst_id ;

select inst_id, event, total_waits,time_waited
from gv$system_event
where event in ('gc current block lost','gc cr block lost')
order by inst_id, total_waits desc;

参考：
RAC 环境中 gc block lost 和私网通信性能问题的诊断 (Doc ID 1674865.1)RAC 环境中 gc block lost 和私网通信性能问题的诊断 (Doc ID 1674865.1)RAC 环境中 gc block lost 和私网通信性能问题的诊断 (Doc ID 1674865.1)RAC 环境中 gc block lost 和私网通信性能问题的诊断 (Doc ID 1674865.1)RAC 环境中 gc block lost 和私网通信性能问题的诊断 (Doc ID 1674865.1)