Novell is now a part of Micro Focus

A Closer Look at Novell Standby Server: Enhanced Features and Some Servicing Senarios

Articles and Tips: tip

01 Dec 1998


In our October issue, we presented an overview of the Novell Stand-byServer which allows server mirroring on NetWare servers. This month we take a closer look at Stand-byServer, highlighting some extra features that enhance the abilities of mirrored servers.

Enhanced Features

Utility Server. Normally, a standby machine is dedicated to mirroring the disk devices of the primary server and monitoring the status of the primary to ensure that it is still functioning correctly. However, using the Utility Server feature, the standby machine can be made to function as an independent server while still maintaining a real-time mirror of the primary server.

To accomplish this, the standby machine is given its own SYS volume on a disk device that is not mirrored over the dedicated link to the primary. Thus, with its own SYS volume, the standby machine can have network clients of its own, or run processes that are independent of the primary server, see Figure 1.

Figure 1: StandbyServer allows the standby to have its own non-mirrored disk.

If the primary server fails, the standby machine's clients will be disconnected from the standby when it switches over and becomes the primary server. This is useful if the primary server is "dominant" and the standby machine can help share the load of network services by functioning as a backup server, print server, CD-ROM server, or some other process that is necessary.

Novell's SnapShotServer product makes use of this feature to actually freeze data as it appeared at specific moments in time and keep these "frozen" images on the standby machine. This allows backup engines to make use of the standby machine to perform backups at any time without encumbering the primary server with the additional overhead normally associated with performing backups.

Remote Disk Read Blocking. When used in WAN or long-distance configurations, Standby Server has the ability to disable disk device read operations from the standby machine, enhancing total system throughput and performance. When disk devices are mirrored, the operating system will spread read requests among the different devices that make up a mirrored set. This normally increases system performance by spreading the requests among several devices. But when the disk devices are separated by implementing StandbyServer in a WAN configuration, the disk reads may take longer than is acceptable since the data may have to travel over several routers or hops. To avoid any system performance degradation, these remote reads can be disabled.

Dedicated Link Throttling. When a dedicated link cannot be used, or when using a shared data backbone for such link, Stand-byServer can implement a throttling mechanism so as not to monopolize the bandwidth available on the link. This allows client network traffic to have acceptable performance. In most cases, a throttle is not necessary. But in high network usage scenarios where a dedicated link is not available, the StandbyServer throttle mechanism can be used to allow acceptable network performance.

Common Device Sharing. StandbyServer allows the primary and standby machines to be connected to shared devices. Since the primary server is usually the active machine, it may need to have access to external devices such as RAID, tape changers, and CD-ROM jukeboxes. In this configuration, if the primary server fails, the standby machine also will need to have access to these devices to fully function as the new primary server. Stand-byServer's special command files can be executed at switch-over time. Devices that have a dual-connect capability can be used automatically with StandbyServer.

Performance Tuning. The IPX protocol that StandbyServer makes use of is configurable. A default packet size of 512 bytes transfers data. In order to increase performance, this packet size is automatically increased to make the best use of high performance adapters and media types. For example, Ethernet can make use of 1500-byte packets; FDDI can make use of much higher packet sizes—4096 bytes or larger. The StandbyServer IPX protocol is adaptable to these different media topologies and is automatically modified to achieve maximum performance. Several NetWare SET parameters can be modified to enhance individual disk/system performance as well.

Some Servicing Scenarios

After a switchover occurs, the failed server should be repaired or replaced in order to restore the system to normal operation. With StandbyServer, the system administrator can schedule service or debug the failed server while the standby machine is operating as the active or primary server. Though maintenance on the failed server can be performed at any time, you should restore the failed server so that system high availability is restored. When only one machine is functioning, if it fails network clients will not have access to data or services.

The StandbyServer system allows customers to have equal or dissimilar machines working as a high data accessibility pair. After the failed server has been serviced or replaced, the following steps need to be taken to restore StandbyServer operation.

Scenario 1—Failed primary server; standby machine running as new primary; switching server roles back to original configuration not necessary:

  1. Reconnect all network cards, disk devices, and dedicated link (if used), of the repaired server.

  2. If a disk device failed that contained the NetWare SYS volume and boot data, it must be rebuilt using standard DOS and NetWare installation tools. StandbyServer software needs to be reloaded as well.

  3. Power up the failed server and execute the Standby batch file. This will start the failed server operating as a standby machine to the new primary server.

  4. If a disk device failed and was replaced, use NetWare's installation program to re-mirror data from the primary server to the new standby machine. Otherwise mirroring will automatically take place.

  5. Check the mirror status of the disk devices and the Stand-byServer user interface to ensure proper functionality.

Scenario 2—Failed primary server; standby machine running as new primary; switching server roles back to original configuration is desired:

  1. After systems have synchronized, down and exit both machines.

  2. Start the original standby machine as a standby server by executing the standby batch file.

  3. Start the original primary server as a NetWare server by executing the server program.

  4. Check the mirror status of the disk devices and the Stand-byServer user interface to ensure proper functionality.

Scenario 3—Failed standby machine; no switchover takes place:

  1. Reconnect all network cards, disk devices, and dedicated link (if used), etc., on the repaired machine.

  2. If a disk device failed that contained the DOS partition, it will have to be rebuilt with standard DOS and NetWare installation tools. StandbyServer software has to be reloaded as well.

  3. Power up the failed machine and execute the Standby batch file. This will start the failed machine operating as a standby machine to the original primary server.

  4. Check the mirror status of the disk devices and the Stand-byServer user interface to ensure proper functionality.

You can learn more about Novell StandbyServer from the following document:

http://www.novell.com/whitepapers/ha/sbswp.html

* Originally published in Novell AppNotes


Disclaimer

The origin of this information may be internal or external to Novell. While Novell makes all reasonable efforts to verify this information, Novell does not make explicit or implied claims to its validity.

© Micro Focus