IC5Notice: We have upgraded developerWorks Community to the latest version of IBM Connections. For more information, read our upgrade FAQ.
  • No replies
1 Post

Pinned topic Auto detect disk problem(I/O error) then takeover

‏2011-03-28T11:04:25Z |
Dear all :

AIX v6.1
HACMP v6.1 fixpak4
- two node (node A , node B)
- one resource group ( boot+service IP ) , two adapter each node
- one shared VG (normal VG) on SAN environment , two adapter each node

Base on my HACMP v5 experience , to test ip network takeover with unplug all cable on node A,
-> HACMP detect network down event
-> stop resource on node A
-> start resoure on node B
In other words , *resource group will failover from node A to node B cause by network down*(ether network).

But if do same test with shared VG environment (SAN). The HACMP will NOT trigger any event. that mean HACMP did NOT protect all I/O failed on one node.

my question is if I do same test with unplug all SAN fibre cable within node A :
1. if HACMP support auto detect shared VG (I/O error) problem then takeover resource to standby node ?
2. if yes , How to configure HACMP for above requirement ?
by the way : there are NO any error notification now. and shared VG is normal VG (not enhance concurrent)