Mount faults on failover, works when done manually - Veritas Cluster Server
This is a discussion on Mount faults on failover, works when done manually - Veritas Cluster Server ; I've got VCS 2.0 running on a pair of Sun 280R's. I have two service groups,
one for each server, configured so that each server is exporting an NFS share,
and acts as a backup for the other server.
What ...
-
Mount faults on failover, works when done manually
I've got VCS 2.0 running on a pair of Sun 280R's. I have two service groups,
one for each server, configured so that each server is exporting an NFS share,
and acts as a backup for the other server.
What I'm experiencing is easy to describe: If I start the service group
manually, resource by resource (specifically, I do DiskGroup, then Volume,
them Mount, and then the rest), everything works fine. But if I just tell
the service group to go online, OR if it fails over to the other server,
it hangs for awhile and then faults on the Mount resource, with a timeout
error.
My gut feeling is that this could be solved by making sure that the Mount
resource doesn't try to start until after DiskGroup and Volume have come
online, but I just can't figure out how to make that work. I've got a big
flat spot on my forehead from working on this all evening -- any good ideas?
Thanks!!
David Stults
-
Re: Mount faults on failover, works when done manually
Hi,
You have to make dependencies between the different resources in the service
group. It will look like a tree structure but built from the bottom up, i.e. the
top resources will depend on the lower level ones to start before it they be
brought online. The dependency is shown as a line between the resources.
Mount should depend on Volume, which in turn should depend on DiskGroup. This is
always the case when you design a Service Group: Always make sure that you get the
resource dependencies right. That will make sure that everything is started in the
correct order.
BR,
Peter
David Stults wrote:
> I've got VCS 2.0 running on a pair of Sun 280R's. I have two service groups,
> one for each server, configured so that each server is exporting an NFS share,
> and acts as a backup for the other server.
>
> What I'm experiencing is easy to describe: If I start the service group
> manually, resource by resource (specifically, I do DiskGroup, then Volume,
> them Mount, and then the rest), everything works fine. But if I just tell
> the service group to go online, OR if it fails over to the other server,
> it hangs for awhile and then faults on the Mount resource, with a timeout
> error.
>
> My gut feeling is that this could be solved by making sure that the Mount
> resource doesn't try to start until after DiskGroup and Volume have come
> online, but I just can't figure out how to make that work. I've got a big
> flat spot on my forehead from working on this all evening -- any good ideas?
>
> Thanks!!
> David Stults