IPMP问题请教

superhjy

本人用两块１００Ｍ网卡做了Ｓolaris内置的ＩＰＭＰ网卡备份（主备模式）测试，但发现以下几个问题，
１．浮动ＩＰ切换后，２～３分钟内出不了外网，就是其他网段的服务器ping不通这个浮动ＩＰ，通过在
本台服务器上找包发现，网关上是有发送ＡＲＰ报文，但奇怪的是，本主机一直没有对其响应，所以导致
２～３分钟的延时．

２．备网卡上的测试ＩＰ在居于网内可以ping通，但也出不了外网，通过抓包发现网关上也有询问测试ＩＰ的
ＡＲＰ包，但其就是不响应．


不知道大家在配置过程中是否也有相同的情况，还请帮忙定位一下，不胜感激啊！


到底是网络问题？
还是ＩＰＭＰ实现机制问题？

yoyosys

应当这两方面的问题都有,我这边这么操作的后果是网卡,时不时的会闪断,但是又不影响正常的数据业务,也为这事,着急中!

superhjy

2楼的兄弟配置的应该是IPMP双主模式的吧?

yoyosys

配置只是当一块网卡当掉的时候,自动切换到另一块网卡进行正常的网络通信.大家没有碰到过这样问题的.我把log贴出来给大家参考一下:
Jun 21 01:08:54 SAPQAS in.mpathd[157]: [ID 594170 daemon.error] NIC failure detected on bge0 of group ipmp1
Jun 21 01:08:54 SAPQAS in.mpathd[157]: [ID 832587 daemon.error] Successfully failed over from NIC bge0 to NIC e1000g0
Jun 21 01:09:03 SAPQAS in.mpathd[157]: [ID 299542 daemon.error] NIC repair detected on bge0 of group ipmp1
Jun 21 01:09:03 SAPQAS in.mpathd[157]: [ID 620804 daemon.error] Successfully failed back to NIC bge0
Jun 23 12:48:53 SAPQAS in.mpathd[157]: [ID 594170 daemon.error] NIC failure detected on e1000g0 of group ipmp1
Jun 23 12:48:53 SAPQAS in.mpathd[157]: [ID 832587 daemon.error] Successfully failed over from NIC e1000g0 to NIC bge0
Jun 23 12:49:02 SAPQAS in.mpathd[157]: [ID 299542 daemon.error] NIC repair detected on e1000g0 of group ipmp1
Jun 23 12:49:02 SAPQAS in.mpathd[157]: [ID 620804 daemon.error] Successfully failed back to NIC e1000g0
Jun 25 02:28:52 SAPQAS in.mpathd[157]: [ID 594170 daemon.error] NIC failure detected on e1000g0 of group ipmp1
Jun 25 02:28:52 SAPQAS in.mpathd[157]: [ID 832587 daemon.error] Successfully failed over from NIC e1000g0 to NIC bge0
Jun 25 02:29:02 SAPQAS in.mpathd[157]: [ID 299542 daemon.error] NIC repair detected on e1000g0 of group ipmp1
Jun 25 02:29:02 SAPQAS in.mpathd[157]: [ID 620804 daemon.error] Successfully failed back to NIC e1000g0

差不多每天都有这样的日志信息

myniker

e.g: IPMP group active-active with interfaces qfe0 and qfe4 with one
physical IP.
Entry of /etc/hostname.qfe0:
<production_IP_host> netmask + broadcast + group ipmp1 up \
addif <test_IP_host> netmask + broadcast + deprecated -failover up
Entry of /etc/hostname.qfe4:
<test_IP_host> netmask + broadcast + group ipmp1 deprecated -failover up
The IPMP group name “ipmp1” is freely chosen in this example!
If the defaultrouter is NOT 100% available please read infodocs
70060 and 70062.
Notes:
• Do not use test IP for normal applications.
• Test IP for all adapters in the same IPMP group must belong to a single IP subnet.
• You need an additional IP for each logical host.
• In the /etc/default/mpathd file, the value of
TRACK_INTERFACES_ONLY_WITH_GROUPS must be yes (default).
• FAILBACK must be yes (default). Bug 6429808.

一般设置正确的话,排除网卡网线ipmp补丁等原因,问题大都是网络网关,防火墙等问题.
强制全双工试试,改改网络探测时间,默认网关是否有限制什么,对于默认网关不能完全开放的情况,可以添加替代主机用来检测网络的脚本.

solaris study

正贴啊