How to tell which resource(s) in a service group cause the PARTIAL state of the group.
Article ID: 100007596
Updated On:
Description
Problem
The VCS service group is in a partial state. A service group may include many resources. Large configurations may make determination of resource(s) that are causing the PARTIAL state difficult.
This can be especially true where the resources are marked "Not Critical" therefore, the resource(s) may not show a FAULTED state.
Solution
- To determine the resource(s) which cause the PARTIAL state, first it must be determined which resource(s) are not online.
For this example, a 2 node cluster exists with 2 Oracle service groups in an Active - Passive configuration.
The status of the cluster is shown:
# hastatus -sum
-- SYSTEM STATE
-- System State Frozen
A node1 RUNNING 0
A node2 RUNNING 0
-- GROUP STATE
-- Group System Probed AutoDisabled State
B ORA_DEV-SG node1 Y N PARTIAL
B ORA_DEV-SG node2 Y N OFFLINE
B ORA_APP-SG node1 Y N OFFLINE
B ORA_APP-SG node2 Y N ONLINE
The ORA_APP-SG is fully online on node2, which is expected. In the above example, the ORA_DEV-SG service group on system "node1" is shown in PARTIAL state. Since the two groups are both Active - Passive (failover) groups, the state of OFFLINE on one cluster node is expected.
Note that no resources are listed as faulted in the output above.
- To determine the resource that is OFFLINE and causing the PARTIAL state of group ORA_DEV-SG, further investigation is required.
The following command example may be used to determine the OFFLINE resource(s).
Syntax:
hares -display -attribute State -group
Using the example above, to find the resource(s) causing the PARTIAL in the ORA_DEV-SG group:
# hares -display -attribute State -group ORA_DEV-SG -sys node1
#Resource Attribute System Value
diskgroup State node1 ONLINE
mount State node1 ONLINE
listener State node1 ONLINE
oracle State node1 OFFLINE
It is shown that the "oracle" resource is the particular resource that is in OFFLINE state on both the listed nodes, causing the PARTIAL state. Once resources causing the PARTIAL state are confirmed, further actions to resolve the OFFLINE resource may be taken to bring the service group fully online.
- The same steps above may be used for determining PARTIAL status for Active - Active (non-failover, parallel) service groups.
In those scenarios, a PARTIAL status will be shown where all the resources in a group are expected to show ONLINE on all nodes, as in Oracle RAC / CFS configurations.
A simple example of PARALLEL setup might be like this:
# hastatus -sum
-- SYSTEM STATE
-- System State Frozen
A node1 RUNNING 0
A node2 RUNNING 0
-- GROUP STATE
-- Group System Probed AutoDisabled State
B ORA_DEV-SG node1 Y N ONLINE
B ORA_DEV-SG node2 Y N ONLINE
B ORA_APP-SG node1 Y N ONLINE
B ORA_APP-SG node2 Y N ONLINE
B cvm node1 Y N ONLINE
B cvm node2 Y N ONLINE
and a PARTIAL status example where all the service groups are parallel might look like this:
# hastatus -sum
-- SYSTEM STATE
-- System State Frozen
A node1 RUNNING 0
A node2 RUNNING 0
-- GROUP STATE
-- Group System Probed AutoDisabled State
B ORA_DEV-SG node1 Y N PARTIAL
B ORA_DEV-SG node2 Y N ONLINE
B ORA_APP-SG node1 Y N ONLINE
B ORA_APP-SG node2 Y N ONLINE
B cvm node1 Y N ONLINE
B cvm node2 Y N ONLINE
Using the information above, to find the resource(s) causing the PARTIAL in the ORA_DEV-SG group:
# hares -display -attribute State -group ORA_DEV-SG
#Resource Attribute System Value
diskgroup State node1 ONLINE
diskgroup State node2 ONLINE
mount State node1 ONLINE
mount State node2 ONLINE
listener State node1 ONLINE
listener State node2 ONLINE
oracle State node1 OFFLINE
oracle State node2 ONLINE
Issue/Introduction
