Netserver LPR 2 Node Cluster Isssue - Shared Array

[Paul T Foster - EL]

Hi,
I have 2 HP LPRs, each with a NetRAID 3si card, connected to a RS/12 (cluster-enabled) as a cluster. Both servers are running Windows NT Entrprise Advanced Server.
NODE2 had a hardware failure - and after manual switch over of their oracle application NODE3 succfessfully took control over the shared strorage array. Hardware was replaced in the failed NODE2 server (I/O RISER CARD)

Issue we have now is now the system has been repaired NODE2 was re-introduced and this seemed to work ok, However when NODE3 is powered up the entire Drive cabinet goes offline red lights appear on all drives. It also knocks NODE2 access out. The only way the system will run is with NODE3 Off.

Any advice wouuld be appreciated thanks

[Oscar Gil]

Paul,

Looks like a netraid configuration mismatch, try to delete this configuration in the NETRAID of NODE3 when disconnected from the cabinet, then switch off both servers, connect everything, switch on only the NODE3 and try to see if in the configuration of the NETRAID you can load the configuration from the disks, then restart this NODE3 server and stop it on OS Selection, switch on NODE2 and if everything is OK it won't complain about the disks, and you will then be able to start both systems.

Hope this helps

Oscar

[kris rombauts]

Hi Paul,

Oscar gave a good explanation to fix this if the issue is a config mismatch.

Before doing anyhting else, first check if both 3Si controllers are still on a different SCSI id i.e. node 2 has id=7 and node 3 has id=6. If not, then this was your issue and for some reason during the repair this was changed (should not).


Another method that has proven to fix this (again assuming the issue is the config mismatch between the 3Si in the node3 and the raid information stored on the disks in the RS/12 shared storage), is this :


- power both nodes down
- power up node 3 and stop it at the
Express tools menu (CTRL M)
- power up node 2 and let it boot in Windows
- once succesfully booted, open Netraid
Assistant and change some non importent
parameter, such that you can re-save the
configuration in to NVRAM.
Because both Netraid controllers are
present on the SCSI bus, both will get
their NVRAM re-written with the current
RAID configuration, rebooting node3 would
then fix this issue.

- In Netraid assistant in between your
disks, you should see a icon for the 3Si
controller (at ID=6 or 7, depends
which node has which ID). This is in fact
the controller from the remote node (node
3) because Netraid Assistant does never
show the local controller.

- The reason why a change of some parameter
is needed, is because if nothing is
changed, the save function doesn't do
anyhting. One parameter that can be
safely changed is i.e. read policy

- check first how the Adapter Properties
are set (select Adapter, then Properties
in the pull down menu) . This shows you
how the Read policy is currently set.

- select a LD and from the menu choose
Logical Drive, Change Config, Read then
select one of the choices and confirm the
change and save that config to NVRAM by
choosing Configuration, Save from the
Netraid Assistant menu.


- once the Config is saved , you can set
the Read policy back to what it was
originally and resave again.


HTH

Kris