H.3 July 25, 2006 (Updates)

Updates were made to the following sections. The changes are explained below.

H.3.1 Management Tools for NSS

The following change was made to this section.

Location

Change

Protocols

To check the WBEM status, as root in a console shell, enter

rcowcimomd status

To start WBEM, as root in a console shell, enter

rcowcimomd start

H.3.2 Coexistence and Migration Issues

The following changes were made to this section.

Location

Change

NSS Kernel Module (km_nss) for SUSE Linux Enterprise Server 9 and Later

The kernel modules are in SLES 9 and later.

Access Control Issues for NSS on OES Linux

Links to the following resources were added:

WBEM

To check the WBEM status, as root in a console shell, enter

rcowcimomd status

To start WBEM, as root in a console shell, enter

rcowcimomd start

H.3.3 Managing Partitions

The following change was made to this section.

Location

Change

Creating Partitions

This section was revised for clarity.

H.3.4 Managing Multiple Connection Paths to Devices (NetWare)

The following changes were made to this section.

Location

Change

Managing Multiple Connection Paths to Devices (NetWare)

This section was revised for clarity.

Planning Your Multiple Connection Path Solution

IMPORTANT:The NSS Media Manager supports fault tolerance with the NSS multipath solution; it does not support load balancing across the multiple paths. Some third-party multipathing systems (such as EMC PowerPath) support load balancing across the multiple connection paths. When you use third-party multipath solutions, make sure you do not turn on NSS MPIO.

H.3.5 Managing Pools

The following changes were made to this section.

Location

Change

Creating a Pool

IMPORTANT:The IP address of the virtual server must be in the same IP subnet as the server nodes in the cluster where you plan to use it.

Verifying and Rebuilding NSS Pools and Volumes

This section was revised for clarity.

H.3.6 Managing Volumes

The following changes were made to this section.

Location

Change

Read Ahead Blocks in Section 13.2, Understanding Volume Attributes

Information about Read Ahead Blocks was moved to Section 23.10, Read Ahead Blocks Commands because the readaheadblocks parameter is set as a command switch, not as an attribute of the volume.

Section 13.5, Updating eDirectory Volume Objects

IMPORTANT:If you delete and replace a Volume object for a volume where user home directories are stored, the User object loses its home directory attribute. This is because part of the home directory attributes points to a particular Volume object. When that Volume object is deleted, eDirectory needs to cleanup all references to the object being deleted.

Section 13.10, Renaming (Modifying) the Mount Point (Linux)

This section is new.

Section 13.13, Activating and Deactivating CD and DVD Image Files as NSS Volumes (NetWare)

While the CD or DVD image file exists on the server as an NSS volume, deactivation of the NSS volume where the image source file resides causes the image file to be removed as an NSS volume. You cannot modify the ISO file when it is in use.

H.3.7 Using Volume Attributes that Help Ensure Data Integrity

The following change was made to this section.

Location

Change

Guidelines for Using File-Level Snapshots

This section is new.

H.3.8 Managing Compression on NSS Volumes

The following changes were made to this section.

Location

Change

Some Files Do Not Compress Well

IMPORTANT:If you use Sybase database files in a volume, such as for ZENworks databases, do not enable compression on the volume, or mark each database file with the Don’t Compress ( dc) attribute so that it is never compressed even if compression is enabled for the volume. For details, see Technical Information Document 10075966 .

Understanding Server-Level Compression Parameters

The Enable File Compression parameter controls compression behavior only for volumes on the server where the Compression attribute is enabled. It does not control the Compression attribute itself.

The Days Untouched Before Compression parameter uses the date the file was last accessed for reading or writing to determine if a file is inactive, and therefore, eligible for compression. When background compression starts, it first evaluates which files meet this inactivity requirement to determine which files are to be compressed during the compression period.

Guidelines for Compression

This section was previously titled “Planning Compression for Your Server.” It was moved to the Understanding Compression section in hopes that users might actually read it.

H.3.9 Managing Space Quotas for Volumes, Directories, and Users

The following changes were made to this section.

Location

Change

Managing User Space Quotas

This section was revised for clarity.

Configuring User Space Quotas for Users Who Were Not Initially Enabled for Linux Access (Linux)

Some methods to disconnect NSS volume users while you set user quotas are:

  • Request that the users log out until you are done, then notify them when it is safe to reconnect.

  • Remove file system access rights to the volume for the affected users to prevent access while you are working, then add the rights back when you are done. This is easier, of course, if all of the users involved have file system access rights granted through a group assignment so that you remove and add rights for the group instead of each individual user at a time.

  • Allow the users to remain logged in until you are done, then reboot the server to break their connections. When users log in again, the user quotas are enforced.

H.3.10 Salvaging and Purging Deleted Volumes, Directories, and Files

The following changes were made to this section.

Location

Change

Configuring Salvage for the NSS Volume

Background information about the difference between the salvage area for NSS volumes and Traditional NetWare volumes was provided for clarity.

Setting the Low and High Watermarks for Automatically Purging Deleted Files

At the minimum setting of 0%, the low watermark activates the autopurge only when the system is totally out of free space. If it is set this low, the users are likely to get out-of-space errors when they try to save files to the full disk. Setting the low watermark to a percentage a little higher than 0% guarantees that autopurging begins before free space is completely used up and users are less likely to get out-of-space errors.

H.3.11 Managing Distributed File Services (NetWare)

The following change was made to this section.

Location

Change

Creating a DFS Management Context

IMPORTANT:Whenever you rebuild the VLDB, it will search all services in the context you specify as the DFS Management Context. If you use DFS for only a small subset of your total servers, the VLDB rebuild is faster if you place only the servers that use DFS in a separate context, then specify the DFS Management Context at that same level.

H.3.12 Using NSS Commands and Utilities

The following changes were made to this section.

Location

Change

Help Command

Initially, the value reported is the default setting for the parameter. If you modify the value, it reports the actual value.

Extended Attributes Commands (Linux)

This section is new.

O_NOATIME Option Support for the Linux open Command (Linux)

This section is new.

nss /ZLSSPendingWriteIOs

Currently, there is no way to see the current value. Initially, you can assume that the value is the default setting. If you set ZLSSPendingWriteIOs to a valid value within the minimum and maximum range, you can assume that NSS correctly applied the new value if no error is reported when you issue the command. You must keep track manually of whatever valid value you set.

H.3.13 Configuring the System Cache to Fine-Tune NSS Performance (NetWare)

The following changes were made to this section.

Location

Change

Before You Adjust Cache Parameters

IMPORTANT:Do not use the Monitor utility to check memory configuration and usage when running NSS volumes. The Cache Buffers statistics in Monitor apply only to NetWare Traditional file systems. Instead, use Novell Remote Manager for NetWare.

Restoring Settings After a Failed Tuning Attempt

Replace parameter with the parameter you want to reset to its previous value, such as 

server -z CacheBalance

H.3.14 Troubleshooting

The following change was made to this section.

Location

Change

NSS Volume Disappears from NCPCON (Linux)

This section is new.

H.3.15 Upgrading Legacy NSS and NetWare Traditional Volumes

The following change was made to this section.

Location

Change

Copying and Upgrading Legacy Volumes with Volume Copy Upgrade

IMPORTANT:For Traditional NetWare volumes, VCU does not migrate the deleted.sav directory and the deleted files it contains to the new NSS volume. The deleted files are essentially purged. If you need to salvage deleted files, do it before using VCU to migrate the data.

H.3.16 Using EVMS to Manage Devices with NSS Volumes (Linux)

The following changes were made to this section.

Location

Change

Viewing Free Space Available to EVMS-Managed Devices

To exclude multiple devices (such as /dev/hda and /dev/hdb), use 

exclude = [hda hdb]

Edit the Boot Loader File

This section is new. It explains how to modify the boot loader file so that devices load from /dev/evms/ instead of /dev/.

H.3.17 Comparison of NSS for Linux and Linux Traditional File Systems

The following change was made to this section.

Location

Change

User space quotas (user space restrictions)

Modified the entry to indicate that user quotas are not supported for traditional Linux volumes with NCP. This enhancement is expected in a future release.