Nfs Guidelines (Nas 8000); Use Of Nis; Setting The Number Of Nfsd Daemons - HP StorageWorks 8000 - NAS User Manual

single volume group. For example, it is possible to create two LUNs, one on each redundancy group and
then create one volume group using both LUNs. If a hard drive was added to the array enclosure after
storage had been defined, the space of the hard drive could be made available by creating a new LUN,
and then adding that LUN to an existing volume group. Volume groups combine the storage of the LUNs.
One possible reason for multiple volume groups would be to utilize the NAS 8000 in a heterogeneous
environment in which CIFS/SMB shares are being protected with HP Virus Guard Real Time Protection
while there are also nfs clients accessing the NAS 8000 storage. See the constraints section below for
more details. For the NAS 8000 system that has a VA 7400, multiple volume groups on separate
redundancy groups can be used to increase the fault tolerance level. The operating system of the NAS
8000 will always use the most efficient data pathway available to access the VA 7xxx array.
After creating volume groups, the next step is to create file volumes. A file volume is somewhat equivalent
to a file system. There are many possibilities at this level, and each choice should be weighed carefully
before committing to an overall storage design. One or more file volumes are entirely contained in a single
volume group. A file volume cannot be a member of more than one volume group. Placing an entire
database into one single file volume may facilitate easier backup through the snapshot features of the NAS
8000. Snapshots can still be utilized for backing up the database even if the database data files exist in
multiple file volumes and multiple volume groups. If the backup or usage plan calls for the use of NAS
8000 snapshots, space must be reserved in the volume group(s) for the snapshots.
Finally, once the file volumes have been created, the DBA/NAS 8000 administrator can create the desired
directory structure and nfs exports. A directory structure can be created from the Command View NAS GUI
and then individual directories can be exported via nfs. Another method is to share the root of the storage
file volume(s) for read/write access by the DBA. The DBA can then create directories under the file volume
root as desired and set the permissions. The directories can then be exported through the use of the
Command View NAS GUI. The nfs export name does not have to be the same as the directory name. For
example, the root directory (/) of a file volume 'rnd' could be exported as 'RndHome'. Please take note of
the export path displayed by the Command View NAS GUI in order to correctly define the mount points on
the clients. It is permissible to name an nfs export and an smb share the same, but it is not permissible to
define multiple nfs exports (or multiple smb shares) with the same name (even if they are in different file
volumes / volume groups). For example, it is possible to have an nfs export and an smb share both named
Oracle_Data, but it would not be possible to have another nfs export or smb share in a different volume
group or file volume named Oracle_Data. Once the nfs exports have been defined, the Oracle server(s)
may mount to those exports, and the database may be moved, started or created.

nfs guidelines (NAS 8000)

use of nis

The NAS 8000 administrator has the option of enabling or disabling the use of NIS. The use of NIS will
not affect Oracle functionality with the NAS 8000 if the NIS server is properly configured for the
environment and the Oracle servers have been correctly configured for NIS. However, if NIS is enabled,
the NIS environment (user ID's etc.) should be setup prior to the use or installation of Oracle. NOTE: The
use of NIS does not eliminate the need to configure an Oracle server as a "trusted host" if the Oracle
executables are to be stored on the NAS 8000.

setting the number of nfsd daemons

The NAS 8000 administrator has the ability to set the number of nfsd daemons. The default number of nfsd
daemons is 10. This is sufficient for many environments, but for serving data to database(s) that have many
simultaneous users, the administrator may need to increase this number. The use of multiple nfsd daemons
allows multiple nfs requests to be handled concurrently. However, each nfsd daemon is able to process
only one nfs request at a time. When the number of nfs requests is greater than the number of nfsd
daemons, it is possible that the communications "channel" will "fill up" and further requests will be refused
until outstanding nfs requests have completed. There is a tradeoff in managing the number of nfsd
daemons; too many nfsd daemons can cause unnecessary system overhead and scheduling contention on
the NAS 8000 while too few nfsd daemons will cause a slower response to nfs requests and may cause the
following errors to appear on the client:
nfs: Server NASName not responding, still trying.
6