This file contains entries for all of the local and remote file systems that are to be used as resources in the configuration. RMS is responsible for mounting and unmounting each of these file systems in order to bring them online or offline, respectively, according to the requirements of the running configuration.
For each file system to be managed by RMS, create a line in /etc/fstab.pcl with the standard fstab fields, and then insert the string #RMS# at the beginning of the line. For more information, see the fstab manual page.
Note the following restrictions when you create /etc/fstab.pcl:
Do not specify the same file system in both a standard /etc/fstab entry and an RMS /etc/fstab.pcl entry. The standard entry will mount the file system at system startup, and this will create a conflict when RMS starts up and attempts to mount the same file system.
If a remote file system is specified in the form <server_name>:<server_path>, then <server_name> must be a host name that appears in the /etc/hosts file. It cannot be an IP address, and you should not rely on DNS to resolve the name.
Example
#RMS#/dev/sdb2 /fs2 ext2 defaults 1 2 #RMS#/dev/sda1 /mnt/data1 auto noauto,user 0 0 #RMS#/dev/sda2 /mnt/data2 auto noauto,user 0 0
If the RMS comment is of the form #RMS:<appname>#, the file system entry applies only to the specified application. From an RMS perspective, file systems assigned to a given application are independent of those assigned to other applications. A file system can be assigned to two or more applications, provided only one of the applications is online at any time.
Example
#RMS:app1#/dev/sdb2 /data3 auto noauto,user 0 0 #RMS:app2#/dev/sdb6 /data4 auto noauto,user 0 0
In general, if you create an /etc/fstab.pcl control entry for a remote file system or a shared filer on one node, then you should duplicate that entry on every other node in the cluster, even if some nodes will not mount that file system. This helps to ensure that the configuration behaves consistently throughout the cluster.
Use a similar procedure for entries that specify local file systems and mount points. If all the nodes have the same architecture, you may be able to simply copy the entire /etc/fstab.pcl control file. However, if the local physical disk device differs from node to node, you must individually adjust the entries for the same mount point. For example, the respective entries for /mnt1 on node1 and node2 might be as follows:
#RMS#/dev/sda3 /mnt1 ...
#RMS#/dev/sdb5 /mnt1 ...
In all cases, for each mount point that appears in /etc/fstab.pcl, be sure to create the directory on every node in the cluster.