This content has been marked as final. Show 4 replies
There are two ways of configuring IPMP. The first one called 'probe based' is by creating test addresses like in your case. The second one called 'kernel based' you don't need test addresses.
Simply put this line on your interfaces configuration file (/etc/hostname.<>) that you want to be part of the IPMP group:
solaris10 group mpath1 up
Then reboot. After reboot, test the IPMP functionality by using command if_mpadm -d <interface> to cause a fail over procedure and ip_mpadm -r <interface> to cause a fail back procedure.
thanks for your these information , it's new to me.
excuse me to ask again , to make sure I understand right:
- with this configuraation Active-Active?
- "" if_mpadm -d <interface> to cause a fail over procedure and ip_mpadm -r <interface> to cause a fail back procedure"" , this is to test the failover, but I the actual failover in case of failure happens automatically?
- which way do you advice me to go with?
thanks for your support
If you have, for example interfaces e1000g0 and e1000g1 that are part of the same IPMP group, if you issue:
if_mpadm -d e1000g0
you will see that the IP data of the e1000g0 will be handled by e1000g1 interface. A logical interface e1000g1:1 will be created with the same IP configuration as e1000g0. The e1000g0 will show the IP 0.0.0.0.
If you give if_mpadm -r e1000g0, the logical interface will be automatically removed and e1000g0 will continue to work normally. Hope I explained it clearly.
are you sure the /etc/hostname.bge0 file is correct ?
not sure what all the plus symbols shown in first post are and also broadcast is misspelled in the addif line
/etc/hostname.bge0it should look like this with bscs-bl and bscs-bl-net0 in /etc/hosts
bscs-bl netmask broadcast + \+
group production up \
addif bscs-bl-net0 deprecated netmask brodcast + -failover up+
bscs-bl netmask + broadcast + group production up \
addif bscs-bl-net0 deprecated netmask + broadcast + -failover up