HACMP - cluster verification warnings

[ogniemi]

I have cascading cluster built of 2 nodes. Each node has one boot and one stdby interface. The cluster has 2 resource groups one running on one node and second one on the other.

service2:/ # cllsif
Adapter Type Network Net Type Attribute Node IP Address Hardware Address

cln1_boot boot clether ether public cln1 10.2.2.65
service1 service clether ether public cln1 10.2.2.64
cln1_stby standby clether ether public cln1 10.1.144.65
cln1_tty1 service cl_serial rs232 serial cln1 /dev/tty1
cln1_tmssa2 service cl_tmssa tmssa serial cln1 /dev/tmssa2
cln2_boot boot clether ether public cln2 10.2.2.66
service2 service clether ether public cln2 10.2.2.63
cln2_stby standby clether ether public cln2 10.1.144.66
cln2_tty1 service cl_serial rs232 serial cln2 /dev/tty1
cln2_tmssa1 service cl_tmssa tmssa serial cln2 /dev/tmssa1
service2:/ #


service2:/ # clshowres

Resource Group Name service1_rg
Node Relationship cascading
Participating Node Name(s) cln1 cln2
Service IP Label service1
Filesystems ALL
Filesystems Consistency Check logredo
Filesystems Recovery Method sequential
Filesystems/Directories to be exported
Filesystems to be NFS mounted
Network For NFS Mount clether
Volume Groups datavg
Concurrent Volume Groups
Disks
Shared Tape Resources
Connections Services
Fast Connect Services
Application Servers service1_as
Highly Available Communication Links
Miscellaneous Data
Automatically Import Volume Groups false
Inactive Takeover true
Cascading Without Fallback false
9333 Disk Fencing false
SSA Disk Fencing false
Filesystems mounted before IP configured false
Run Time Parameters:

Node Name cln1
Debug Level high
Host uses NIS or Name Server false

Node Name cln2
Debug Level high
Host uses NIS or Name Server false


Resource Group Name service2_rg
Node Relationship cascading
Participating Node Name(s) cln2 cln1
Service IP Label service2
Filesystems
Filesystems Consistency Check logredo
Filesystems Recovery Method sequential
Filesystems/Directories to be exported
Filesystems to be NFS mounted
Network For NFS Mount
Volume Groups
Concurrent Volume Groups
Disks
Shared Tape Resources
Connections Services
Fast Connect Services
Application Servers service2_as
Highly Available Communication Links
Miscellaneous Data
Automatically Import Volume Groups false
Inactive Takeover true
Cascading Without Fallback false
9333 Disk Fencing false
SSA Disk Fencing false
Filesystems mounted before IP configured false
Run Time Parameters:

Node Name cln2
Debug Level high
Host uses NIS or Name Server false

Node Name cln1
Debug Level high
Host uses NIS or Name Server false


During the cluster veriffication I get below warnings - does anyone know what is wrong in above cluster config?:


No custom disk methods to verify.

Verifying export files for resource group: service1_rg.
------------------------------------------

Verifying export files for resource group: service2_rg.
------------------------------------------
WARNING: Node cln1 is a member of more cascading groups with IPAT than it has
standby adapters
WARNING: Node cln2 is a member of more cascading groups with IPAT than it has
standby adapters

Verification has completed normally.

Remember to redo automatic error notification if configuration has changed.
Contacting node cln1 ...
Contacting node cln1 ...
Succeeded in synchronizing HACMPnode ODM to the remote node cln1.

Contacting node cln1 ...
Succeeded in synchronizing HACMPnetwork ODM to the remote node cln1.

Contacting node cln1 ...
Succeeded in synchronizing HACMPcluster ODM to the remote node cln1.

Contacting node cln1 ...
Succeeded in synchronizing HACMPnim ODM to the remote node cln1.

Contacting node cln1 ...
Succeeded in synchronizing HACMPadapter ODM to the remote node cln1.

Contacting node cln1 ...
Succeeded in synchronizing HACMPsubnet ODM to the remote node cln1.


---- end ----

[Jan 11 2004, 17:22:47]

 

[sbix]

It's really strange, 'cause IPATs are basically the service IP address which can be moved between boot/standby adapters in a node and on the other adapters of the others nodes partecipating to the cluster.
You have 2 Resource Group, 2 IPATs and 4 adapters ..... looks fine to me

 

[ogniemi]

Thx for reply. I think similiar.

(This is only warning, verification completes successfuly and cluster works without any problems.)

regards,m.

 

[haexpert]

What the warning is saying is that if cln1 had a network failure AND cln2 then had a critical failure node than HA/CMP takeover from cln2 to cln1 would not complete as the standby adapter would already have been used.

As you state this is only a warning and would only cause you problems if you had two failures within the cluster