MC/ServiceGuard NFS Toolkit A.11.11.03 Release Notes

Known Problems and Workarounds

Problem

If you run AutoFSon NFS servers that participate in a Highly Available NFS cluster (through

MC/ServiceGuard), the behavior of building a local IP interface list can, in certain cases,

render MC/ServiceGuard unable to migrate an NFS package during a fail-over event.

NOTE This problem impacts only HA/NFS server systems running AutoFS. You can

safely use AutoFS on client systems.

Cause

HA/NFS uses relocatable IP addresses that migrate from a primary NFS server to an

adoptive node during a fail-over event. If you start AutoFS on an HA/NFS server after

MC/ServiceGuard is running, these relocatable IP addresses are included in AutoFS’ list of

local IP interfaces. AutoFS map entries referencing these relocatable addresses cause AutoFS

to create LOFS mounts to the exported local ﬁlesystems. These LOFS mounts can create

problem with MC/ServiceGuard, because HA/NFS does not unmount LOFS ﬁlesystems before

a package migration. AutoFS-managed LOFS mounts that are holding resources in the

exported ﬁlesystems, when ServiceGuard tries to initiate a package failover, may render

MC/ServiceGuard unable to successfully unmount the local ﬁlesystems and migrate them to

the adoptive node.

Workarounds

The potential workarounds for this problem are as follows:

• Ensure that AutoFS is running on the NFS server before HA/NFS is started.

• Run the legacy automounter on HA/NFS server systems.

• Conﬁgure the HA/NFS server as a dedicated NFS server. This may involve migrating all

user applications off from the HA/NFS servers so that there is no need to run either

automount or AutoFS on these systems.

• Halt the NFS packages running on the server node, perform the necessary action, and

then run the NFS packages again.

• Move the NFS packages manually from primary server node to adoptive server node.