3PAR network connectivity

[nmatejak]

HI!

I have a problem with 3PAR 7200; after restarting Cisco switch there is no more connectivity to management address.
I manage to connect to MFG port, do the "Re-enter network configuration" on one Node, the management address is pingable again, but I'm not able to connect with Management Console or SPOCC.
Today I do the "Re-enter network configuration" on both nodes, now the management address is not pingable at all.

Please help!

Nedeljko

[profp62]

Hi

Maybe i have a similar problem. After 7200 installation the network people from customer did connectivity tests - pinging 3par, they plugged/unplugged one network - still pinging. Then they unplugged second network and connectivity was lost and remains disconnected even after returning the interface. As usual their network guys have everything correct ; no security, no mac spoofing... New 3par, no data and hosts, so i restarted both controllers and after some time (not immediatelly) the connection was back.
I have no control over their network but must suggest a solution.
I'm not network specialist, so is there someone who knows the reason ?

Thanks

[Darking]

Its a known issue caused by link flapping.

See more in this forum post:
viewtopic.php?f=18&t=1160&hilit=Link+flap

[nmatejak]

I have session with HP support yesterday, the problem was in netc process.
The proces was terminated and restarted from root account.

Thank you all for replaying!

[profp62]

Hi

Aditional question - is there any way how to switch active-passive network mgmt connection from command line ? (without unplugging the cable)

Thanks

[profp62]

Hi

The customer still has network connection problem.
They asked me, what are the requirements / recommendations from HP to set the parameters of physical Cisco ports (autonegotiation, port speed, duplex, flow control, jumbo frame, ...)
I cannot find it nowhere.

everything i found is

>>
Each controller node supports one Ethernet connection to a switch or hub. Separate
connections from the Ethernet switch or hub to at least two controller nodes are required to support
redundancy. With redundancy, one IP address is shared between the two connections, and only
one network connection is active at a time. If the active network connection fails, the IP address is
automatically moved to the surviving network connection.
<<

I think there is no that recomendation, but maybe someone knows that info?

Thanks.

[cali]

Try "setnet changenode"
or disable the Port on the CISCO.

[profp62]

To Cali :

Thats it. Thank you.

[T16]

Is it normal for that switch over of mgmt nics to take up to 20 seconds?

Rebooting a node and watching the pings, we lose about 12 pings before the management kicks in on the other node.

I am assuming its only a management thing, as if that happened with the live data ports, there would be chaos.

[MammaGutt]

Is it normal for that switch over of mgmt nics to take up to 20 seconds?

Rebooting a node and watching the pings, we lose about 12 pings before the management kicks in on the other node.

I am assuming its only a management thing, as if that happened with the live data ports, there would be chaos.

Spanning tree on the switches may take up to 30 seconds to bring the port online.