Hey, I'm currently experiencing a problem with FAS ports.
Do you have the command to modify the port of a cluster and not of a vserver?
#Problem Port FAS
simple comet
Can you be more specific about the problem that you're experiencing?
sullen token
Can you be more specific about the problem that you're experiencing?
Hello, I finally made it:
The problem was physical: the port cable no longer worked, I just had to change the cable and everything came back.
However, I'd be interested to know if it's possible to change the cluster port.
Let me explain:
So i have :
x_clus 1 --> node1 port e0b
x_clus2--> node1 port e0b
x_clus 2 -->node 2 port e0b
x_clus1 --> node 2 port e0b
What would be the order to change the port?
dull cedar
what do you mean "change the port"? configure a different port as cluster port? configure options (like MTU, flowcontrol, ...) on the port?
dull cedar
So i have :
x_clus 1 --> node1 port e0b
x_clus2--> node1 port e0b
x_clus 2 -->no...
what do you mean by this? what is "x_clus 1"? A cable? a LIF name? and why is everything on e0b everywhere, that sounds wrong...
sullen token
what do you mean "change the port"? configure a different port as cluster port? ...
yeah for exemple change the port e0b to e0c
sullen token
what do you mean by this? what is "x_clus 1"? A cable? a LIF name? and why is ev...
Basically, I didn't install the FAS, but port e0b is connected from the right controller e0b to the left controller e0b.
So For example, the first two x_clus are the left controller and the second two are the x_clus of the right controller.
x_clus = Cluster name
dull cedar
when you say "the first two x_clus", are you talking about LIFs? You should put the LIFs onto different ports (e.g. e0b and e0c). Your cluster needs two cluster network connections for redundancy
please post network port show -ipspace Cluster and network interface show -role cluster so that we can debug your setup
sullen token
This
and this
@dull cedar
But to make it simple, I have my e0b port which was down, so my question is to change the e0b port to e0c for example.
e0c-d-e-f is connected to the switch
e0b is connecter into the other e0b controler netapp
dull cedar
in the current config you need to connect the e0a ports, because currently your cluster network is degraded, and if e0b goes down, so goes your cluster 😉
you can configure other ports for cluster networking but you need to make sure that they are connected correctly (either directly, aka. "switchless cluster", or to the two netapp backend cluster switches)
the ports that are in the "Cluster" ipspace are for the cluster-internal communication and you should use only those two because that's generally the only ports that are supported for the cluster network. in your case e0a and e0b
sullen token
Yes, the e0a is being rewired, and with this morning's breakdown, we're replacing all the cables with new ones, just to be on the safe side.
dull cedar
you cannot use e0c for the cluster network because it looks like that's being used for the frontend network right now, and also it's only 1Gbit which is too slow for the cluster network
sullen token
the ports that are in the "Cluster" ipspace are for the cluster-internal communi...
Oh ok, So if the e0a and e0b fall, there's no more server?
So if I have :
e0a --> e0a
e0b --> e0b
The rest of the ports --> switch
Is good ?