DS 2300 - full bus mode for two rp3440 faceing problem

[Walker_3]

Hi all,

I am configuring ds2300 in full bus mode as a shared storage with 4 disks. I need to see all disks from both nodes. I start storage first and then node1. It works fine but the problem starts when I start the second node. The second node gives a red system alert and it started dumping and rebooting. Please help.

P.S. This thread has been moved from HP-UX Technical Documentation to HP-UX > sysadmin. -HP Forum Moderator

[Slawomir Gora]

Hi,

did you change on node2 ID of SCSI controler ?
In default configuration each controler has ID=7, and you have two devices with the same SCSI ID.

[Walker_3]

I did not change controller ID of any node. But I can see all disks from both node with the same device names.
Could you please help how can I change controller ID?

[Slawomir Gora]

Hi,

if you use SCSI you have to change ID on one of conntroler.

steps are:
1. enter to BCH (stop during booting)
2. go to service menu
3. use scsi command to set new id

for details look at chapter BCH Commands/Service Menu in doc:
http://docs.hp.com/en/A7137-96003/A7137-96003.pdf

[Walker_3]

Hi,
I am not clear on it. Could you please explain a bit details?

Why do I need to change the SCSI ID of one node? since I am still seeing all disks with the same device names from both nodes.
Please let me clear on it.

[Slawomir Gora]

Hi,
because on scsi bus there can't be two or more devices with the same scsi id.

[Torsten.]

Every SCSI ID on a bus has to be unique. If you have 2 initiators (controllers) on the same bus, you have to switch one of them from default 7 to 6.

Locate the hardware path, boot to the boot menu, go to the service menu and change the ID, e.g.:

Service Menu: Enter command > scsi 1/0/12/0/0 init 6

[Walker_3]

Hi,
I understand now, but still have confusion on SCSI id. Look, I have two rp3440 with one HBA on each node. DS2300 has two controllers. One controller is connected to node1 and another one is connected to node2. The other port of the controller is terminated.

Now please tell me what should I do now and on which node?

[A. Clay Stephenson]

It doesn't matter which node you change to 6. For simplicity, I always leave the controllers on one node at 7 (the default) and then change all the controller SCSI ID's on the shared bus(es) on the other node to 6 (the next highest priority. If you actuall had 3 hosts per bus (which would require a "vee" connector) the controllers on that host would be set to 5 -- next highest priority after 6.

[A. Clay Stephenson]

By the way, unless you are running raw devices don't even think about mounting shared disks as read-write. The only way to share filesystems is read-only or read-write connected to one node and NFS mounted by all other hosts. There is also a shared filesystem available from Veritas but the standard HP-UX only handles shared read-only mounts.

[Walker_3]

Hi Stephenson,

Thanks a lot. Should I keep storage ds2300 powered on during changing the controller ID or I can change it keeping storage down?

[A. Clay Stephenson]

I would power-down the other host (that you are leaving at 7) and power off the array but that's not really necessary.

[Walker_3]

well done.