ipmp 配置下出现奇怪问题

zhangyh123

两台sun 服务器，v880 v890 做oracle rac ,
配置了 4块网卡,做ipmp 只有一块网卡出现如下故障,其他内网/外网络ping 都正常,

有没有相关的大虾遇到并解决过?

ps : 系统没有什么负载时候也如此,

$ ping -s v880b-priv
PING v880b-priv: 56 data bytes
64 bytes from v880b-priv (10.0.1.4): icmp_seq=0. time=0.336 ms
64 bytes from v880b-priv (10.0.1.4): icmp_seq=1. time=0.223 ms
64 bytes from v880b-priv (10.0.1.4): icmp_seq=2. time=0.213 ms
64 bytes from v880b-priv (10.0.1.4): icmp_seq=3. time=0.246 ms
ICMP Port Unreachable from gateway v880b-priv (10.0.1.4)
for udp from v890-priv (10.0.1.2) to v880b-priv (10.0.1.4) port 33478
ICMP Port Unreachable from gateway v880b-priv (10.0.1.4)
for udp from v890-priv (10.0.1.2) to v880b-priv (10.0.1.4) port 33477
ICMP Port Unreachable from gateway v880b-priv (10.0.1.4)
for udp from v890-priv (10.0.1.2) to v880b-priv (10.0.1.4) port 33474
64 bytes from v880b-priv (10.0.1.4): icmp_seq=4. time=0.229 ms
ICMP Port Unreachable from gateway v880b-priv (10.0.1.4)
for udp from v890-priv (10.0.1.2) to v880b-priv (10.0.1.4) port 33477
ICMP Port Unreachable from gateway v880b-priv (10.0.1.4)
for udp from v890-priv (10.0.1.2) to v880b-priv (10.0.1.4) port 33478
ICMP Port Unreachable from gateway v880b-priv (10.0.1.4)
for udp from v890-priv (10.0.1.2) to v880b-priv (10.0.1.4) port 33474
64 bytes from v880b-priv (10.0.1.4): icmp_seq=5. time=0.194 ms
64 bytes from v880b-priv (10.0.1.4): icmp_seq=6. time=0.273 ms
64 bytes from v880b-priv (10.0.1.4): icmp_seq=7. time=0.210 ms
64 bytes from v880b-priv (10.0.1.4): icmp_seq=8. time=0.254 ms
ICMP Port Unreachable from gateway v880b-priv (10.0.1.4)
for udp from v890-priv (10.0.1.2) to v880b-priv (10.0.1.4) port 60625
64 bytes from v880b-priv (10.0.1.4): icmp_seq=9. time=0.219 ms
64 bytes from v880b-priv (10.0.1.4): icmp_seq=10. time=0.236 ms
ICMP Port Unreachable from gateway v880b-priv (10.0.1.4)
for udp from v890-priv (10.0.1.2) to v880b-priv (10.0.1.4) port 33478
ICMP Port Unreachable from gateway v880b-priv (10.0.1.4)
for udp from v890-priv (10.0.1.2) to v880b-priv (10.0.1.4) port 33477
ICMP Port Unreachable from gateway v880b-priv (10.0.1.4)
for udp from v890-priv (10.0.1.2) to v880b-priv (10.0.1.4) port 33474
ICMP Port Unreachable from gateway v880b-priv (10.0.1.4)
for udp from v890-priv (10.0.1.2) to v880b-priv (10.0.1.4) port 33477
ICMP Port Unreachable from gateway v880b-priv (10.0.1.4)
for udp from v890-priv (10.0.1.2) to v880b-priv (10.0.1.4) port 33474
ICMP Port Unreachable from gateway v880b-priv (10.0.1.4)
for udp from v890-priv (10.0.1.2) to v880b-priv (10.0.1.4) port 33478
64 bytes from v880b-priv (10.0.1.4): icmp_seq=11. time=0.224 ms
64 bytes from v880b-priv (10.0.1.4): icmp_seq=12. time=0.220 ms
64 bytes from v880b-priv (10.0.1.4): icmp_seq=13. time=0.243 ms
64 bytes from v880b-priv (10.0.1.4): icmp_seq=14. time=0.228 ms
64 bytes from v880b-priv (10.0.1.4): icmp_seq=15. time=0.225 ms
64 bytes from v880b-priv (10.0.1.4): icmp_seq=16. time=0.184 ms
64 bytes from v880b-priv (10.0.1.4): icmp_seq=17. time=0.256 ms
64 bytes from v880b-priv (10.0.1.4): icmp_seq=18. time=0.237 ms


从 v880-priv ping v890-priv 没有这个情况发生

[ 本帖最后由 zhangyh123 于 2009-5-2 14:08 编辑 ]

zhangyh123

cat /etc/hostname.ge0

10.0.1.5  netmask + broadcast + group ipmp2 up  addif 10.0.1.50 deprecated -failover netmask + broadcast + up

cat /etc/hostname.ce0

192.168.2.240 netmask + broadcast + group ipmp1 up  addif 192.168.4.240 deprecated -failover netmask + broadcast + up

cat /etc/hostname.ce1

192.168.2.241 netmask + broadcast + group ipmp1 up  addif 192.168.4.241 deprecated -failover netmask + broadcast + up

zhangyh123

服务器上曾报告如下错误

May  1 01:00:09 v890 in.mpathd[203]: [ID 585766 daemon.error] Cannot meet requested failure detection time of 10000 ms on (inet ce1) new failure detection time for group "ipmp1" is 140966 ms
May  1 01:01:09 v890 in.mpathd[203]: [ID 302819 daemon.error] Improved failure detection time 70483 ms on (inet ce0) for group "ipmp1"
May  1 01:01:10 v890 in.mpathd[203]: [ID 302819 daemon.error] Improved failure detection time 35241 ms on (inet ce1) for group "ipmp1"
May  1 01:01:11 v890 in.mpathd[203]: [ID 302819 daemon.error] Improved failure detection time 17620 ms on (inet ce0) for group "ipmp1"
May  1 01:01:11 v890 in.mpathd[203]: [ID 302819 daemon.error] Improved failure detection time 10000 ms on (inet ce1) for group "ipmp1"
May  1 01:03:01 v890 in.mpathd[203]: [ID 585766 daemon.error] Cannot meet requested failure detection time of 10000 ms on (inet ce0) new failure detection time for group "ipmp1" is 141708 ms
May  1 01:04:01 v890 in.mpathd[203]: [ID 302819 daemon.error] Improved failure detection time 70854 ms on (inet ce1) for group "ipmp1"
May  1 01:04:02 v890 in.mpathd[203]: [ID 302819 daemon.error] Improved failure detection time 35427 ms on (inet ce0) for group "ipmp1"
May  1 01:04:03 v890 in.mpathd[203]: [ID 302819 daemon.error] Improved failure detection time 17713 ms on (inet ce1) for group "ipmp1"
May  1 01:04:04 v890 in.mpathd[203]: [ID 302819 daemon.error] Improved failure detection time 10000 ms on (inet ce0) for group "ipmp1"

qgqceo

有可能是网络的问题。

Oracle的官方诊断描述(http://docs.sun.com/app/docs/doc/816-0211/6m6nc66sm?a=view)：
Cannot meet requested failure detection time of time ms on (inet[6] interface_name) new failure detection is time ms.

The round trip time for ICMP probes is higher than the specified failure detection time. The network is probably congested or the probe targets are loaded. in.mpathd automatically increases the failure detection time to whatever it can achieve under these conditions.

Improved failure detection time time ms on (inet[6] interface_name).
The round trip time for ICMP probes has now decreased and in.mpathd has lowered the failure detection time correspondingly.

故障解释和处理办法：该报错是指，从报文发送到接收返回的时间比指定的检测时间要长。当时的网络很有可能存在堵塞的情况。在这种情况下，进程in.mpathd会自动增加故障检测时间。建议继续观察网络的通信情况，看看是否存在堵塞。

淡然的紫色

回复 1# zhangyh123


    你的问题最后如何解决的？

adambbs

必须要配网关且接到交换机上就ok了

淡然的紫色

回复 6# adambbs


    能具体点吗？我们的oracle rac 私有网络是通过交换机连接的 不是网线直连