OpenSolaris

Printable Version Enter a New Search
Bug ID 4850407
Synopsis if_mpadm and IPMP DR failure
State 10-Fix Delivered (Fix available in build)
Category:Subcategory network:ipmp
Keywords clearview
Responsible Engineer Peter Memishian
Reported Against s10_30
Duplicate Of
Introduced In solaris_9
Commit to Fix snv_107
Fixed In snv_107
Release Fixed solaris_nevada(snv_107)
Related Bugs 4347223 , 6783149
Submit Date 17-April-2003
Last Update Date 28-January-2009
Description 
I have the following setup:
sc_ipmp0	hme0 and qfe1
ifconfig -a
hme0: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 2
        inet 10.6.17.44 netmask ffffff00 broadcast 10.6.17.255
        groupname sc_ipmp0
        ether 8:0:20:d2:2b:85 
hme0:1: flags=9040843<UP,BROADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NOFAILOVER>
 mtu 1500 index 2
        inet 10.6.17.48 netmask ffffff00 broadcast 10.6.17.255
qfe1: flags=9040843<UP,BROADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NOFAILOVER> m
tu 1500 index 4
        inet 10.6.17.49 netmask ffffff00 broadcast 10.6.17.255
        groupname sc_ipmp0
        ether 8:0:20:cb:8d:29 

Now I want to remove hme0 from the group so I do the following:
if_mpadm -d hme0
ifconfig -a
hme0: flags=89000842<BROADCAST,RUNNING,MULTICAST,IPv4,NOFAILOVER,OFFLINE> mtu 0 
index 2
        inet 0.0.0.0 netmask 0 
        groupname sc_ipmp0
        ether 8:0:20:d2:2b:85 
hme0:1: flags=89040842<BROADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NOFAILOVER,OF
FLINE> mtu 1500 index 2
        inet 10.6.17.48 netmask ffffff00 broadcast 10.6.17.255
qfe1: flags=9040843<UP,BROADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NOFAILOVER> m
tu 1500 index 4
        inet 10.6.17.49 netmask ffffff00 broadcast 10.6.17.255
        groupname sc_ipmp0
        ether 8:0:20:cb:8d:29 
qfe1:1: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 4
        inet 10.6.17.44 netmask ffffff00 broadcast 10.6.17.255

Everything is fine till now. Now to remove hme0 which is already offlined I do
ifconfig hme0 group ""

I get the following message after about 5 seconds:
psimics2# Apr 15 17:44:00 psimics2 in.mpathd[2379]: Successfully failed back to 
NIC hme0

ifconfig -a
hme0: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 2
        inet 10.6.17.44 netmask ffffff00 broadcast 10.6.17.255
        ether 8:0:20:d2:2b:85 
hme0:1: flags=9040842<BROADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NOFAILOVER> mt
u 1500 index 2
        inet 10.6.17.48 netmask ffffff00 broadcast 10.6.17.255
qfe1: flags=9040843<UP,BROADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NOFAILOVER> m
tu 1500 index 4
        inet 10.6.17.49 netmask ffffff00 broadcast 10.6.17.255
        groupname sc_ipmp0
        ether 8:0:20:cb:8d:29
        
I am not sure why a failback happens when hme0 is not in an IPMP group. No 
failback should happen and the IP address should remain on qfe1:1

I had S9U3 installed build 8. This was happening on S9GA also.
Work Around 
N/A
Comments 
N/A