I am new to the Celerra NX4 and have some questions that will clear up a few things for me:
Can high availability be achieved using active/active on the data movers or only when using active/passive?
Reading the documentation it states that I need to configure IP addresses for the storage processors for using NST for array maintenance. I don¿t see these management ports only 1 control station mgmt port? Are these addresses suppose to be in the same VLAN as the control station mgmt IP?
If high availibilty can be achieved is dependent on the point of view. F.e. for a network issue, it can (depending on the configuration), but if a datamover is failing, active/active is no longer HA.
The Controlstation will use proxy arp and forward the IPs of the storage processors using its mgmt port into your customer network. If your network is configured for VLAN, then the SP IPs should also be part of the Controlstation mgmt VLAN.
Thanks for the quick response so its standard practice to use active/passive confiuration for HA? Does the data movers not work the same way as the a clariion storage system which is active/active - If one processor fails the luns trespass too the other proc?
Regarding network configuration on the data mover - if I use FSN's i need fsn device that maps ports on both data movers so that if one data mover fails the FSN device will fail over too the standby blade? Is this correct?
Yes, active/passive is standard practice, the datamover do not work the same way like SPs. A datamover does not failover its work to an active datamover, only a whole datamover can failover to a passive (standby) datamover.
Network Configuration: FSN is designed to connect one datamover to two network switches. One of these connections is the active one, the other is a standby one which will be activated, if the first one will fail (broken cable, network port, switch ...). Of course the clients need to be able to be served using both switches, else FSN wouldn´t make sense.
If there is only Network Switch, I would recommend using trunking instead of FSN, because then all network connectivity will be used, (no passive standby). The standby datamover does need the same network connectivity than the active datamover. A FSN is not mapping two datamovers, only one datamover but two network switches.
Peter_EMC
674 Posts
194
0
Posted May 20th, 2009 05:00
The Controlstation will use proxy arp and forward the IPs of the storage processors using its mgmt port into your customer network. If your network is configured for VLAN, then the SP IPs should also be part of the Controlstation mgmt VLAN.