FAS3050 - cannot set up dual-path configuration

[michal_psota]

Hi,

I am trying to set up dual-path configuration on FAS3050 filer + DS14Mk2 shelves.

When I create sigle loop from 0a to shelves, everything is OK and OnTap boots normally.

But when I connect second loop from port 0c (image is attached), OnTap does not boot.

During boot it writes:

Loading: 0x200000/42260756 0x2a4d914/14350236 0x37fd0b0/1732666 0x39a40ea/6 Entry at 0x00200000
Starting program at 0x00200000
Press CTRL-C for special boot menu
Fri Jul 16 12:14:07 GMT [nvram.battery.state:info]: The NVRAM battery is currently ON.
Fri Jul 16 12:14:41 GMT [fci.initialization.failed:error]: Initialization failed on Fibre Channel adapter 2b.
Fri Jul 16 12:14:41 GMT [fci.initialization.failed:error]: Initialization failed on Fibre Channel adapter 2a.
Fri Jul 16 12:14:41 GMT [fci.initialization.failed:error]: Initialization failed on Fibre Channel adapter 0d.

NetApp Release 7.3.3: Thu Mar 11 22:43:52 PST 2010
Copyright (c) 1992-2009 NetApp.
Starting boot on Fri Jul 16 12:14:02 GMT 2010
Fri Jul 16 12:14:41 GMT [fci.initialization.failed:error]: Initialization failed on Fibre Channel adapter 0b.
add net 127.0.0.0: gateway 127.0.0.1
Fri Jul 16 12:14:53 GMT [fmmb.instStat.change:info]: no mailbox instance on local side.
Fri Jul 16 12:14:53 GMT [monitor.chassisPower.degraded:notice]: Chassis power is degraded:
Fri Jul 16 12:14:55 GMT [fmmb.current.lock.disk:info]: Disk ?.? is a partner HA mailbox disk.
Fri Jul 16 12:14:55 GMT [fmmb.current.lock.disk:info]: Disk 0a.16 is a partner HA mailbox disk.
Fri Jul 16 12:14:55 GMT [fmmb.instStat.change:info]: missing lock disks, possibly stale mailbox instance on partner side.
Fri Jul 16 12:14:56 GMT [cf.fm.noMBDisksOrIc:warning]: Could not find the local mailbox disks. Could not determine the firmware state of the partner through the cluster interconnect.
Fri Jul 16 12:14:56 GMT [fmmb.current.lock.disk:info]: Disk ?.? is a local HA mailbox disk.
Fri Jul 16 12:14:56 GMT [fmmb.current.lock.disk:info]: Disk 0a.16 is a local HA mailbox disk.
Fri Jul 16 12:14:56 GMT [fmmb.instStat.change:info]: missing lock disks, possibly stale mailbox instance on local side.
Shelf where disk 0a.18 resides is connected to this system by both Fibre
Channel port A and Fibre Channel port B. This is illegal in a Clustered
Failover configuration.

PANIC: Illegal Disk Configuration in process config_check on release NetApp Release 7.3.3 on Fri Jul 16 12:14:58 GMT 2010

and it falls into CFE.

Can anyone help me with it ? Please note that I am novice in NetApp technology, I am self-learning last 3 weeks.

Thanks

Michal

[aborzenkov]

Is port 0c configured as target or initiator? What is output of “fcadmin config”?

[michal_psota]

Aborzenkov, for it, I had to remove second loop, otherwise my OnTap did not boot and thus fcadmin was not available in CFE> environment.

fcadmin config shows now:

fcadmin config

                         Local
Adapter Type      State                  Status
---------------------------------------------------
  0a   initiator  CONFIGURED.            online
  0b   initiator  CONFIGURED.            offline
  0c   initiator  CONFIGURED             offline
  0d   initiator  CONFIGURED.            offline

It seems that all 4 interfaces are configured as initiators.

[mitchells · Accepted solution]

It looks like you have hardware disk ownership enabled and not software disk ownership.  That's what you are getting an illegal configuration that the shelf can be seen on both module A and module B.

If you run disk show -v it will tell you if you have software disk ownership enabled.

Thanks,

Mitchell

[michal_psota]

Yes, I have hardware ownership enabled, disk show -v command does not work and it shows me only usage of "disk" command.

When I enter "storage show" command, the last line is "SANOWN not enabled", it is proof that disks are owned by HW

as per Storage Management Guide.

It is needed to enable SW ownership, when I want use second loop ? I thought that it is needed only for Active/Active

configuration, which is solution with 2 filers connected to cluster.

[aborzenkov]

Yes, you have to enable software based ownership for multipath to work. Without it second connection is interpreted as partner (because it goes to channel B).

[michal_psota]

In this case I will have to learn about disk ownerships and how to set it up.

Isn't there any possibility to have HW ownership and set second loop, only for cases, that the first loop

should fail (because of 0a port fails) ? I think that you will say "NO", I am just asking.

Thanks

[mitchells]

MPHA is going to do what you need.  The software disk ownership controls what head actually owns the disks.  Inside of Data ONTAP the disks are going to appear down both adapters and down one adapter should one adapter be disabled.

https://now.netapp.com/Knowledgebase/solutionarea.asp?id=kb28560

Thanks,

Mitchell

[michal_psota]

Thanks Mitchell, it was helpful !