|
|
JF Mezei <jfmezei.spamnot@xxxxxxxxxxxx> wrote:
<snip>
>
> To take a node out of a cluster, you can:
>
> MC SYSGEN SYSGEN> USE CURRENT SYSGEN> SET VAXCLUSTER = 0 SYSGEN> WRITE
> CURRENT
>
>
> Then reboot, and the machine will be out of the cluster. If you do
> that to all machines in the cluster, the cluster will no longer be
> active. But setting the parameters back on will re-enable it
But *DON'T* do this without reading the rest of JFs note. Nodes in a
cluster share the system disk in many/nost cases, and just following the
above advice will trash your system disks within a few seconds.
So read on:
> There are big caveats to this: If node B relies on access to drives
> served by node A, then node B will fail because it will no longer have
> access to A's drives.
>
> If there are drives that are shared at the hardware level (shared SCSI
> for instance), then there will not be any lock coordination between
> drives and you will get corruption in files if both nodes try to
> access the file at the same time.
>
> Also note that a node may rely on another node's system disk to boot.
> (a satellite node), and when you take that node out of the cluster. If
> so, you need to change the nodes's default boot device to its own
> local disk, as well as ensuring that its onw local disk has its own
> copy of VMS (and this entails copying all configuration files from
> that node's SYS$SPECIFIC tree from the server's disk to the node's own
> disk so when it boots "alone", it can see all its own files.
>
Hans.
|
|