>Alex Kompel wrote:
<mailto:ben@xxxxxxxxxxxxxxx>> wrote:
*** Can I setup my two node HA cluster with iSCSI quorum drives and
cluster service storage volumes? (anyone do this before?)
I am going to play devils advocate here: is there a specific reason you
want to use GFS in this setup? NetApp has excellent NFS and CIFS support
and it looks like you already paid for both and HA option for NetApp
(270c is a clustered filer).
I wasn't going to use GFS.
I am going to setup an HA cluster with Oracle services.
So it is not simultaneous cluster node access to the disk storage.
(Kind of like HP Service Guard or Veritas Cluster Services).
*** I was thinking about getting 10Gbit/sec uplinks for the NetApp and
two ethernet switches that have 10Gbit uplink ports and that support
802.3ad. The two cluster nodes would use 802.3ad NIC channel bonding
> for the storage access bandwidth. (anyone do this before?)
Do not get 10gbE. First, I don't think you can get 10gbE interfaces in
FAS270.
Second, FAS270 won't be able to saturate even 1gb link. The bottleneck
is usually the filer CPU.
It does support link aggregation but you won't see much of the
difference vs active/passive bonding.
-Alex
Yeah, I should ask NetApp if I can get FOUR nic's into the NetApp.
Then I can do two uplinks (aggregated) to Switch1--Node1,
and two uplinks to Switch2--Node2, then have the Nodes crossconnected
with heartbeat......
Hmmm, THANKS!
--
Linux-cluster mailing list
Linux-cluster@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/linux-cluster
