SET


Purpose

Use at the server console to view and configure operating system parameters.

HINT:  You can also modify server parameter values from MONITOR's Available Options, and then Server Parameters or using NetWare Remote Manager's, Set Parameters link.

The default SET parameter values give maximum performance for most systems. Server parameter values should seldom need to be modified.


Syntax

SET [parameter] = [value]


Parameter categories:

Common File System Parameters
Communications Parameters
Directory Services Parameters
Disk Parameters
Error Handling Parameters
Licensing Services Parameters
Memory Parameters
Miscellaneous Parameters
Multiprocessor Parameters
NCP Parameters
Novell Storage Service Parameters
Service Location Protocol Parameters
Time Parameters
Traditional File Parameters


Using SET

Although most default values of the server parameters don't need to be modified, you might increase the performance of your system by adjusting the values of certain parameters. Suggestions for improving server performance can be found in Optimizing the NetWare Server and Setting Server Parameter Values in the Server Operating System Administration Guide.

Server parameter settings are persistent in NetWare®. If the server goes down, any settings that you have made to tune your server for performance will not be lost.


Displaying and Changing Current Settings

You can also use the MONITOR utility to modify server parameter values. See MONITOR.


Entering Parameters


Parameters that Control the Allocation of Services

Some SET parameters control how the system dynamically allocates services. Three types of parameters interact to control the allocation of a service:


Common File System Parameters

Common File System Parameters contain parameters that apply to both the NSS file system and the traditional file system. For a description of file system parameters, see the following table.


Table 4. Common File System Parameters

Parameter Use to

Maximum Transactions = number

Specify how many transactions can occur at the same time.

Supported values: 100 to 10000
Default: 10000

Maximum Concurrent Directory Cache Writes = number

Specify how many write requests from directory cache buffers are put in the elevator before the disk head begins a sweep across the disk.

Supported values: 5 to 500
Default: 75

A high number creates more efficient write requests. A low number creates more efficient read requests.

Minimum File Delete Wait Time = time

Specify how long a deleted file remains salvageable on the volume.

Supported values: 0 seconds to 7 days
Default: 1 minute 5.9 seconds

Files deleted for less than this minimum aren't automatically purged even if the volume is full and users can't create new files.

Immediate Purge of Deleted Files = value

Supported values: ON, OFF
Default: OFF

If this parameter is set to ON, all files are purged immediately when they are deleted.

Compression Daily Check Stop Hour = number

Specify the hour when you want the file compressor to stop scanning enabled volumes for files that need to be compressed.

Supported values: 0 to 23
Default: 6

Hours are specified by a 24-hour clock: 0 = midnight; 23 = 11 p.m.

This parameter can be set in the STARTUP.NCF file.

Compression Daily Check Starting Hour = number

Specify the hour when you want the file compressor to start scanning enabled volumes for files that need to be compressed.

Supported values: 0 to 23
Default: 0

Hours are specified by a 24-hour clock: 0 = midnight; 23 = 11 p.m.

Note: If the Compression Daily Check Stop Hour parameter is the same as the Compression Daily Check Starting Hour, then the file compressor starts checking every day at the Compression Daily Starting Hour time and runs as long as necessary to finish all files that meet the compressible criteria.

This parameter can be set in the STARTUP.NCF file.

Minimum Compression Percentage Gain = number

Set the minimum percentage a file must compress to remain in a compressed state.

Supported values: 0 to 50
Default: 20

This parameter can be set in the STARTUP.NCF file.

Enable File Compression = value

Specify whether file compression is suspended.

Supported values: ON, OFF
Default: ON

ON allows file compression on compression-enabled volumes.

OFF suspends compression; immediate compress requests are queued until value is reset to ON, when the files meeting criteria will be compressed.

This parameter can be set in the STARTUP.NCF.

Maximum Concurrent Compressions = number

Specify the maximum concurrent or simultaneous compressions allowed.

Supported values: 1 to 8
Default: 2

Concurrent compressions can occur only if there are multiple volumes.

This parameter can be set in the STARTUP.NCF file.

Convert Compressed to Uncompressed Option = value

Specify what the file system does with an decompressed version of a file after the server has decompressed it.

Supported values:

0 = Always leave the file compressed.
1 = Leave the file compressed until second access if it is read only once during the time specified by the Days Untouched Before Compression parameter.
2 = Always leave the file decompressed.

Default: 1

This parameter can be set in the STARTUP.NCF file.

Decompress Percent Disk Space Free to Allow Commit = number

Specify the percentage of free disk space required on a volume for file decompression to permanently change compressed files to decompressed.

Supported values: 0 to 75
Default: 10

This parameter prevents newly decompressed files from filling up the volume.

This parameter can be set in the STARTUP.NCF file.

Decompress Free Space Warning Interval = number

Specify the time between alerts when the file system is not changing compressed files to decompressed because of insufficient disk space. Setting the interval to 0 turns off the alert.

Supported values: 0 seconds to 29 days 15 hours 50 minutes 3.8 seconds
Default: 30 minutes 57.2 seconds

This parameter can be set in the STARTUP.NCF file.

Deleted Files Compression Option = number

Specify whether and when deleted files are compressed.

Supported values:

0 = Don't Compress deleted files
1 = Compress deleted files the next day
2 = Compress deleted files immediately

Default: 1

This parameter can be set in the STARTUP.NCF file.

Days Untouched Before Compression = number

Specify the number of days the system waits after a file was last accessed before it is compressed.

Supported values: 0 to 100000
Default: 14

This parameter can be set in the STARTUP.NCF file.


Communications Parameters

Communication parameters control settings for communication buffers. Four parameters configure packet receive buffers; four control the watchdog.

For a description of communications parameters, see Table 5.


Table 5. Communications Parameters

Parameter Use to

Maximum Packet Receive Buffers = number

Specify the maximum number of packet receive buffers the operating system can allocate.

Supported values: 50 to 25000
Default: 5000

Before increasing this parameter, use MONITOR to view the server's use of packet receive buffers and service processes.

If the number of packet receive buffers is at maximum, increase this parameter in increments of 10 until you have one packet receive buffer per workstation.

If you have EISA or microchannel bus master boards in your server, increase this parameter to provide at least five buffers per board.

If the board is producing No ECB available count errors, provide 10 buffers per board. Use MONITOR (LAN/WAN Drivers) to determine if the board is producing errors.

If the number of allocated service processes is at maximum, you can increase the Maximum Service Processes parameter to decrease the need for more packet receive buffers.

The value of this parameter should be greater than the value of the Minimum Packet Receive Buffers parameter. If it is less, the system increases the value to match that of the Minimum Packet Receive Buffers parameter.

This parameter can be set in the appropriate startup file.

Minimum Packet Receive Buffers = number

Specify the minimum number of packet receive buffers the operating system can allocate. The operating system allocates this number of buffers as soon as the server boots. You must add this command to the startup.ncf file. You cannot change the setting at the console prompt.

Supported values: 10 to 20000
Default: 500

Before increasing this parameter, use NetWare Remote Manager or MONITOR to view the server's use of packet receive buffers.

If you have EISA or microchannel bus master boards in your server and are receiving No ECB available count errors (see LAN/WAN Drivers in MONITOR) right after the server boots, increase this parameter so that each board can have at least five packet receive buffers.

If the allocated number is higher than 10 and the server doesn't respond immediately after booting, increase this parameter.

The value of this parameter must be less than the value of the Maximum Packet Receive Buffers parameter. If it is greater, the system increases the value of the Maximum Packet Receive Buffers parameter to match that of the Minimum Packet Receive Buffers parameter.

Maximum Physical Receive Packet Size = number

Specify the maximum size of packets that can be transmitted on the network. You must add this command to the STARTUP.NCF file. You cannot change the setting at the console prompt or using any other utilties.

Supported values: 618 to 24682
Default: 4202

The default allows 2 KB (data with the packet header). If you use token ring or Ethernet boards, the default is acceptable.

If some of your network boards transmit more than 512 bytes of data per packet, set this parameter for the largest packet size.

IPX NetBIOS Replication Option = number

Specify how the IPXTM router handles replicated NetBIOS broadcasts.

Supported values:

0 = Do not replicate NetBIOS broadcasts
1 = Duplicate broadcasts when there are redundant routes
2 = Suppress duplicate broadcasts

Default: 2

Maximum Interrupt Events = number

Specify the maximum number of interrupt time events (such as IPX routing) allowed before a thread switch is guaranteed to have occurred.

Supported values: 1 to 1000000
Default: 10

Reply to Get Nearest Server = value

Specify whether the server responds to Get Nearest Server requests from stations trying to locate directory and file servers.

This parameter can be set in the startup.ncf file.

Supported values: ON, OFF
Default: ON

Number of Watchdog Packets = number

Specify the number of unanswered watchdog packets that the server sends to a workstation before closing its connection.

Supported values: 5 to 100
Default: 10

Delay Between Watchdog Packets = time

Specify the amount of time between watchdog packets.

Supported values: 9.9 seconds to 10 minutes 26.2 seconds Default: 59.3 seconds

After a server sends out the first watchdog packet, it waits the specified time before sending out succeeding packets if it receives no reply.

Delay Before First Watchdog Packet = time

Specify the amount of time the server waits without receiving a request from a workstation before sending out the first watchdog packet to that station.

Supported values: 15.7 seconds to 14 days
Default: 4 minutes 56.6 seconds

New Packet Receive Buffer Wait Time = time

Specify how long the operating system waits after receiving a request for a packet receive buffer before granting a new buffer.

Supported values: 0.1 second to 20 seconds
Default: 0.1 second

This parameter prevents the system from granting too many buffers during a sudden peak in usage.

If you have an EISA bus master board in your server, don't change this parameter.

Console Display Watchdog Logouts = value

Specify whether a console message is displayed when a connection is cleared.

Supported values: ON, OFF
Default: OFF

If your network is running smoothly, you don't need to display watchdog logouts.

If your workstations are having connection problems, the watchdog logout messages can help you isolate which stations aren't receiving or sending watchdog packets.


Directory Services Parameters

Directory services parameters allow you to do the following:

For descriptions of directory services parameters, see Table 6.


Table 6. Directory Services Parameters

Parameter Use to

eDirectory Trace to Screen = value

Enable the eDirectory trace screen; this displays information about eDirectory events on the monitor.

Supported values: ON, OFF
Default: OFF

eDirectory Trace to File = value

Send messages about eDirectory events to the eDirectory trace file on volume SYS. The default file is SYSTEM:\DSTRACE.DBG.

Supported values: ON, OFF
Default: OFF

The file path and name can be changed with the eDirectory Trace Filename parameter.

The file is circular; it grows to a maximum length of approximately 500 KB and then starts to overwrite itself at the beginning of the file.

If this parameter is set to ON, the trace information is also scrolled on the screen.

eDirectory Trace Filename = path\name

Specify the path and name of the eDirectory trace file on volume Sys.

Maximum length: 254

Default: SYSTEM:\DSTRACE.DBG

eDirectory External Reference Life Span = number_in_hours

Specify the number of hours unused external references are allowed to exist before being removed.

Supported values: 1 to 384 hours
Default: 192

External references are local IDs assigned to users when they access other servers. When users no longer have access, the external references should be removed.

eDirectory Inactivity Synchronization Interval = number_in_minutes

Specify the maximum elapsed time between exhaustive synchronization checks. As soon as you change this value, the system executes the synchronization check. Synchronization checks then recur at the specified interval.

Supported values: 2 to 1440 minutes
Default: 30

If the system has replicas across a WAN link, this value should be set as high as 240 minutes (4 hours) to reduce WAN traffic.

eDirectory Synchronization Restrictions = value, version_number_list

Specify which versions of eDirectory the server can synchronize with.

Supported values: OFF, ON, list of version numbers
(Maximum length of version number list: 131 characters)

Default: OFF

To determine what version is currently loaded on a server, type MODULES at the server prompt. The eDirectory version number is displayed under the heading DS.NLM.

If this parameter is set to OFF, the server synchronizes with all versions available.

If this parameter is set to ON, the server synchronizes only with those versions specified as parameters to the ON value.

Example: ON,420,421

eDirectory Servers Status = value

Mark the status of all server objects in the local namebase as UP or DOWN.

Supported values: UP, DOWN

Use this parameter to reset the status of all the servers if the status of one server isn't accurately recognized by the system.

For example, if a server is up but the system recognizes it as down, set this parameter to mark all servers as up.

Subsequently, the system would reassess the status of all servers and change the status to down for those servers that were truly down.

eDirectory Janitor Interval = number_in_minutes

Specify the interval in minutes at which the janitor process is executed. The janitor process is executed as soon as you change this value and then recurs at the specified interval.

Supported values: 1 to 10080 minutes
Default: 60

The janitor process cleans up unused records, reclaims disk space, and purges objects flagged for deletion.

eDirectory Backlink Interval = number_in_minutes

Specify the interval in minutes at which backlink consistency checking is performed. Backlink consistency checking is executed as soon as you change this value. It then recurs at the specified interval.

Supported values: 2 to 10080 minutes
Default: 780

A backlink indicates that an object in a replica has an ID on a server where the replica doesn't exist.

This process creates needed backlinks and deletes unnecessary ones.

eDirectory Distributed Reference Link Interval = number_in_minutes

Specify the interval in minutes at which distributed reference link consistency checking is performed. Distributed reference link consistency checking is executed as soon as you change this value. It then recurs at the specified interval.

Supported values: 2 to 10080 minutes
Default: 780

A distributed reference link indicates that an object in a partition has an ID in that partition where the actual object doesn't exist.

This process creates needed distributed reference links and deletes unnecessary ones.

eDirectory Trace File Length to Zero = value

Delete the contents of the trace file. This parameter does not delete the file itself. As soon as the file is cleared, the value of the parameter resets to OFF.

Supported values: ON, OFF
Default: OFF

To use this parameter, you must also set the eDirectory trace to file parameter to ON, because the trace file must be open for the system to delete its contents.

Check Equivalent to Me = value

Enforce checking of the Equivalent To Me attribute on eDirectory authentication.

Supported values: ON, OFF
Default: OFF

If this parameter is set to ON, DSREPAIR must be used to synchronize the Equivalence attribute and the Equivalent To Me attribute.

Setting this parameter to ON might adversely affect communication performance.

Bindery Context = context;context

Specify one or more containers to be used by eDirectory when it provides bindery services.

Maximum: 2047 characters, 16 contexts

Multiple contexts are separated by semicolons. Whatever string it specified is set. To make that string effective for all valid contexts the container you specify in the context must be present on that server.

Example:
SET BINDERY CONTEXT = OU=SALES_LA.OU=SALES. O=NOVELL_US;OU=ACCOUNTING.O=NOVELL

This parameter can be set in the startup.ncf file.

eDirectory Bootstrap Address = address

Allow eDirectory to operate properly in the absence of SLP. When SLP is not available to advertise servers and partitions, the local server uses this value to set the bootstrap address that the server would use to find its tree and authenticate to it.

Format for IP Address: Use the standard IP format of four decimal values delimited by periods.

Example: 123.45.67.89 <:524> or 137.65.62.144 (port number is optional) This number is used for both TCP and UDP connections.

Format for IPX Address: Use the standard IPX format of hexadecimal digits representing Network, Node, and Socket.

Example: 12345678:23456789ABCD:0451 or 01010480:00001B1E983A:0451
Defaults apply to any field not present, where default is current values.


Disk Parameters

Disk parameters control Hot FixTM redirection, as well as other aspects of disk reads and writes.

For a description of disk parameters, see the following table.


Table 7. Disk Parameters

Parameter Use to

Sequential Elevator Depth = number

Set the maximum elevator depth for sequential requests.

Media Manager sends the number of sequential requests up to this value to the same device. When the device contains this number of requests and another device in the mirror group is empty, Media Manager begins sending requests to the idle device.

Supported values: 0 to 4294967295
Default: 8

You can set this parameter in the STARTUP.NCF file.

Enable IO Handicap Attribute = value

Enable drivers and applications to inhibit read requests from one or more devices.

Supported values: ON, OFF.
Default: OFF.

Setting this parameter to ON enables the inhibit attribute to function. Setting this parameter to OFF prevents the attribute from functioning.

Do not set this attribute to ON unless instructed to do so by a device manufacturer.

You can set this parameter in the STARTUP.NCF file.

Mirrored Devices Are Out of Sync Message Frequency = time

Set the frequency (in minutes) for checking out-of-sync devices.

Supported values: 5 to 9999 minutes
Default: 28 minutes

You can set this parameter in the STARTUP.NCF file.

Remirror Block Size = value

Set the remirror block size in 4 KB increments. (1=4 KB, 2=8 KB, 8=32 KB, etc.)

Supported values: 1 to 8
Default: 1

Concurrent Remirror Requests = value

Set the number of remirror requests per mirror object.

Supported values: 2 to 32
Default: 32

Ignore Disk Geometry = value

Create nonstandard and unsupported partitions.

Supported values: ON, OFF
Default: OFF

If you set this parameter to ON before modifying or creating a partition, the software ignores disk geometry when creating the partition. This allows you to create nonstandard partitions.

Caution: Setting this parameter to ON may harm other file systems contained on the disk.

You can set this parameter in the STARTUP.NCF file.

Enable Hardware Write Back = value

Enable hardware write back, if supported.

Hardware write back means that I/O write requests may be cached at the device and succeeded before data is committed to the media. Hardware write back usually improves write performance.

Supported values: ON, OFF
Default: OFF

You can set this parameter in the STARTUP.NCF file.

Enable Disk Read After Write Verify = value

Control whether information written to disk is read back and compared with the original data.

Supported values: ON, OFF
Default: OFF

Setting this parameter to ON may decrease performance significantly.

To set this value for currently loaded disks, use the Storage Devices option of MONITOR.

You can set this parameter in the STARTUP.NCF file.


Error Handling Parameters

Error handling parameters control the size of error logs and specify what happens when logs exceed the specified size. They also control how the server responds to an abend or to an NLMTM that does not unload from a protected address space.

For descriptions of error handling parameters, see the following table.


Table 8. Error Handling Parameters

Parameter Use to

Server Log File State = number

Control what happens when the SYS$LOG.ERR file is larger than the size specified by the Server Log File Overflow Size parameter.

Supported values:

0 = Leave SYS$LOG.ERR as is
1 = Delete SYS$LOG.ERR
2 = Rename SYS$LOG.ERR

Default: 1

You can set this parameter in the STARTUP.NCF file.

Server Log File Overflow Size = number

Specify the maximum size of the SYS$LOG.ERR file before the action specified by the Server Log File State parameter occurs.

Supported values: 65536 to 4294967295
Default: 4194304

You can set this parameter in the STARTUP.NCF file.

Boot Error Log File State = number

Control what happens when the BOOT$LOG.ERR file is larger than the size specified by the Boot Error Log Overflow Size parameter.

Supported values:

0 = Leave BOOT$LOG.ERR as is
1 = Delete BOOT$LOG.ERR
2 = Rename BOOT$LOG.ERR
3 = Start a new log file whenever the server is restarted

Default: 3

You can set this parameter in the STARTUP.NCF file.

Boot Error Log File Overflow Size = number

Specify the maximum size to which the BOOT$LOG.ERR file can grow before the action specified by the Boot Error Log File State parameter occurs.

Supported values: 65536 to 4294967295
Default: 4194304

You can set this parameter in the STARTUP.NCF file.

Boot Error Log = value

Specify which error messages from the console will be saved in the BOOT$LOG.ERR file.

Supported values: ON, OFF
Default: ON

If you set this parameter to ON, all error messages displayed on the console will be saved in BOOT$LOG.ERR.

If you set this parameter to OFF, only error messages displayed during the boot procedure will be saved in BOOT$LOG.ERR.

You can set this parameter in the STARTUP.NCF file.

Hung Unload Wait Delay = number

Specify the amount of time the server waits for an NLM to be unloaded from a protected address space after the UNLOAD ADDRESS SPACE command is executed. If the NLM is not successfully unloaded within this interval, the server displays a prompt to kill the address space.

Supported values: 0 seconds to 1 minute 58.3 seconds
Default: 30 seconds

You can set this parameter in the STARTUP.NCF file.

Auto Restart Down Timeout = number

When the server tries to go down after an abend, it sets a timeout just in case there is a problem going down. Specify the amount of time (in seconds) that the server will wait before automatically restarting.

Supported values: 0 to 600 seconds
Default: 180 seconds

Auto Restart After Abend Delay Time = number

Specify the amount of time (in minutes) before the server is brought down after an abend, if the Auto Restart After Abend parameter is in effect.

Supported values: 2 to 60 minutes
Default: 2 minutes

Regardless of what time is set, the server sends a message every two minutes to warn users that the server will be brought down.

Auto Restart After Abend = value

Specify the system's automatic response to an abend.

Supported values:

0 = The system does not respond to the abend.

1 = After an abend, the system determines the source of the abend. Based on findings, the system either keeps the computer running or shuts down the computer and attempts to restart it.

2 = After an abend, the system attempts to recover from the problem and restart the computer after the time specified by the Auto Restart After Abend Delay Time parameter.

3 = After an abend, the system immediately restarts the computer.

Default: 1

For values 1, 2, or 3 to take effect, the Developer Option parameter described in Miscellaneous Parameters must be set to OFF.

If the server is to be shut down, the system sends a message every 2 minutes to warn all connections.

To configure the amount of time the system waits before shutting down the server, use the Auto Restart After Abend Delay Time parameter.

After an abend, the screen displays information about the abend. This information is sent to the ABEND.LOG file on the C: drive. After the server is shut down and restarted, the ABEND.LOG file is transferred to SYS:SYSTEM.

Important: Because the server can abend and be restarted automatically, you should determine whether any abends have occurred. Periodically check either the ABEND.LOG file or the Server Up Time field in MONITOR's General Information screen.


Licensing Services Parameters

Licensing service parameters control Novell Licensing Service diagnosis features.

For a description of licensing parameters, see the following table.


Table 9. Licensing Service Parameters

Parameter Use to

NLSDIAG

Supported values: 19 (string length)
Default: 19

NLS Search Type

Specify the scope of a license certificate search.

Supported values: 0, 1
Default: 0

0 = Search to the root of the tree

1 = Search to the root of the partition

You can set this parameter in either the STARTUP.NCF or AUTOEXEC.NCF files.

Store NetWare 5 Conn SCL MLA usage in NDS

Supported Values: ON, OFF
Default: ON


Memory Parameters

Memory parameters control garbage collection, corruption checking, the amount of memory below 16 MB available to device drivers, and aspects of protected and virtual memory.

For a description of memory parameters, see Table 10.


Table 10. Memory Parameters

Parameter Use to

Average Page In Alert Threshold = value

Specify the point at which the server sends an alert to the console because excessive memory swapping indicates the server might be low on memory.

Supported values: 0 to 4294967295
Default: 2000

The value of the parameter is the average number of pages swapped from disk to memory per second, calculated over the last five seconds.

If the average page-in rate is high, it means a large amount of data is being swapped to and from disk, indicating that the server might be running low on memory.

You can set this parameter in the STARTUP.NCF file.

Memory Protection No Restart Interval = value

Prevent a server from restarting a protected address space that continues to fault.

Supported Values: 0 to 60 minutes
Default: 1 minute

If the address space faults, but has already been restarted within the time period specified by this parameter, the server will not restart the address space. Restart functionality is disabled for the address space.

A restartable address space that continues to fault consumes server resources because the server is forced to continually shut down and restart the space.

Setting the value to 0 disables the parameter so that the address space will always be restarted if it faults.

You can set this parameter in the STARTUP.NCF file.

Memory Protection Fault Cleanup = value

Enable the server to clean up a protected address space that has faulted.

Supported Values: ON, OFF
Default: ON

If this parameter is ON and a protected address space faults, the server removes the address space and its NLMTM programs and returns the resources to the system.

If this parameter is OFF and a protected address space faults, the server does not remove the address space or return resources to the system. The situation is left to the abend recovery mechanism.

You can set this parameter in the STARTUP.NCF file.

Garbage Collection Interval = number

Specify the maximum time between garbage collections.

Supported values: 1 minute to 1 hour
Default: 5 minutes

You can set this parameter in the STARTUP.NCF file.

Alloc Memory Check Flag = value

Specify whether the server will check for corruption in the alloc memory nodes.

Supported values: ON, OFF
Default: OFF

You can set this parameter in the STARTUP.NCF file.

Reserved Buffers Below 16 Meg = number

Specify the number of file cache buffers reserved for device drivers that can't access memory above 16 MB.

Supported values: 8 to 2000
Default: 300

You must set this parameter in the startup.ncf file. You cannot set the value at the command line.


Miscellaneous Parameters

Miscellaneous parameters do not fit aptly or neatly into other categories. The functionality is listed below and described in Table 11.

Default response to optional commands
Adding the server's name to the console prompt
Configuring alerts
Controlling alerts for lost or spurious interrupts
Specifying response when invalid parameters are detected
Enabling developer options
Controlling allocation of service processes
Password security
Enabling execution of secure.ncf at startup
Controlling aspects of the kernel, scheduler, and threads


Table 11. Miscellaneous Parameters

Parameter Use to

Allow Unencrypted Passwords = value

Control the use of unencrypted passwords. This parameter provides for backward compatibility with networks services that do not support encrypted passwords in the NetWare authentication protocol.

Supported values: ON, OFF
Default: OFF

We strongly recommend that you retain the default OFF setting.

NetWare 3.1x and later versions support encrypted passwords. If servers on your network run earlier versions of NetWare, set the value of this parameter to ON.

Warning: Because the use of unencrypted (plaintext) passwords represents a significant security risk, you should update your servers, utilities, print servers and NetWare clients to versions that support encrypted passwords. If you cannot update or upgrade to a newer version and are willing to assume the security risk, use the ON setting.

New Service Process Wait Time = number

Specify how long the system should wait to make an allocation when it receives a request for another service process.

Supported values: 0.3 second to 20 seconds
Default: 2.2 seconds

Maximum Service Processes = number

Specify the maximum number of service processes that the operating system can create. (View the number of service processes in MONITOR.)

Supported values: 5 to 1000
Default: 40

Decrease this parameter temporarily if the server is low on memory. If the server is always low on memory, add more RAM for memory.

Increase this parameter if the number of service processes is at the maximum.

Increasing this number helps only if more than 20 requests are being delayed simultaneously for a disk I/O to be completed.

Minimum Service Processes = number

Specify the minimum number of service processes the operating system can create without having to wait for the time that is specified by the New Service Process Wait Time parameter to elapse.

Supported values: 10 to 500
Default: 10

Global Pseudo Preemption = value

Specify whether or not all threads on the server that use the Traditional File System will also use pseudo preemption.

Supported values: ON, OFF
Default: ON

Set this parameter to ON if you rely on pseudo preemption to enforce blocking for threads.

Pseudo Preemption Count = number

Specify the number of times threads are allowed to make file read or write system calls before a relinquish is forced.

Supported values: 1 to 4294967295
Default: 40

If you increase the value too much, the thread could run for its lifetime.

If you decrease the value too much, every read or write call that normally would not block might, in fact, end up blocking.

Display Lost Interrupt Alerts = value

Control alert messages about lost interrupts.

Supported values: ON, OFF
Default: ON

A lost interrupt occurs when a driver or board requests a service with an interrupt call and then drops the request before the processor can respond.

Lost interrupts generate the following message: Interrupt controller detected a lost hardware interrupt.

This message indicates a hardware or driver problem that could degrade performance.

Unload all drivers and then reload them one at a time to determine which driver has a problem. Then contact the vender of the driver.

Set the parameter to OFF while you are waiting for a resolution.

You can set this parameter in the STARTUP.NCF file.

Display Spurious Interrupt Alerts = value

Control alert messages about spurious interrupts.

Supported values: ON, OFF
Default: ON

A spurious interrupt occurs when hardware in the server creates an interrupt that is defined and reserved for another device.

Spurious interrupts generate the following message: Spurious hardware interrupt <number> detected.

This message indicates a serious error in the hardware.

If your server console displays this message, remove all add-on boards and run SERVER.

If the message doesn't appear, add the boards one at a time to determine which hardware is creating the spurious interrupt. Then contact the vendor.

Set the parameter to OFF while you are waiting for a resolution.

You can set this parameter in the STARTUP.NCF file.

Display Lost Interrupts Threshold = value

Specify the minimum number of lost interrupts per second that must be detected before a lost interrupt alert message will be displayed on the system console.

Supported values: ON, OFF
Default: OFF

Note: Set Display Spurious Interrupt Alerts must also be set to ON.

You can set this parameter in the STARTUP.NCF file.

Display Spurious Interrupts Threshold = value

Specify the minimum number of spurious interrupts per second that must be detected before a spurious interrupt alert message will be displayed on the system console.

Supported values: ON, OFF
Default: OFF

Note: Set Display Spurious Interrupt Alerts must also be set to ON.

You can set this parameter in the STARTUP.NCF file.

Developer Option = value

Control whether options associated with a developer environment are enabled.

Supported values: ON, OFF
Default: OFF

You can set this parameter in the STARTUP.NCF file.

CPU Hog Timeout Amount = value

Specify the amount of time in seconds to wait before terminating a thread that has not relinquished control of the processor.

Supported values: 0 to 1 hour
Default: 1 minute

A value of 0 (zero) disables this option.

You can set this parameter in the STARTUP.NCF file.

Display Old API Names = value

Control messages about obsolescent API functions from earlier versions of NetWare. This parameter can be used as a debugging tool.

Supported values:

ON = Use if you write your own modules and you are upgrading your NetWare modules to use newer APIs.

OFF = Use if you aren't upgrading modules.

Default: OFF

The following types of messages appear when a module is loaded that uses the obsolescent APIs:

Module is using old API: SetInterruptVector
Module is using old API: ReturnPermanentMemory
Module is using old API: AllocateReturnablePermMemory

If you receive messages such as these, contact the vender of the module.

You can set this parameter in the STARTUP.NCF file.

Note: Old API functions are those that have been replaced with newer, more efficient functions. The old functions work---they are not obsolete---but they work more slowly.

Old API functions are not limited to NetWare 3 versions. Nor have all older API functions been replaced---many of them are still in use.

Display Relinquish Control Alerts = value

Control whether messages about processor control are sent to the server console. This parameter can be used as a debugging tool.

Supported values:

ON = Use if you're writing your own loadable modules.
OFF = Use if you're not writing your own loadable modules.

Default: OFF

If a module uses the processor for more than 0.4 second without relinquishing control to other processes, the following types of messages appear:

process_name Process did not relinquish control frequently.
Module: module_name
Code offset in module: memory_address

You can set this parameter in the STARTUP.NCF file.

Halt System on Invalid Parameters = value

Specify whether to stop the system when invalid parameters are detected.

Supported values: ON, OFF
Default: OFF

ON = You want the system to halt when an invalid parameter or condition is detected.

OFF = You want the system to display an alert and continue running when an invalid parameter or condition is detected.

You can set this parameter in the STARTUP.NCF file.

Worker Thread Execute In a Row Count = number

Specify the number of times the scheduler consecutively dispatches new work before allowing other threads to run.

Supported values: 1 to 20
Default: 10

Worker threads are created by the kernel to perform work for the kernel itself.

Alert Message Nodes = number

Specify the number of alert message nodes that have been previously allocated.

Supported values: 10 to 256
Default: 20

You can set this parameter in the startup.ncf file.

Classic Work to do Pre-check Flag

When a classic work to do is scheduled, this flag causes the scheduler to verify if the work to do is scheduled. If it's on the scheduled list, the operating system will trap.

Replace Console Prompt with Server Name = value

Control whether the console prompt is replaced with the NetWare server name.

Supported values: ON, OFF
Default: ON

You can set this parameter in the STARTUP.NCF file.

Sound Bell for Alerts = value

Control whether a bell sounds when an alert message appears on the console.

Supported values: ON, OFF
Default: ON

Command Line Prompt Time Out = time

Specify how long an .ncf file waits before executing the default response to an optional command.

An optional command in an. ncf file is preceded by a question mark (?). The question mark causes the file to prompt you to execute the command. For information about optional commands in .ncf files, see Using Server Batch Files.

Supported Values: 0 to 4294967295 seconds
Default: 10 seconds

Command Line Prompt Default Choice = value

Specify a default response to an optional command in an .ncf file.

An optional command in an .ncf file is preceded by a question mark (?). The question mark causes the file to prompt you to execute the command. For information about optional commands in .ncf files, see Using Server Batch Files.

Supported Values: ON, OFF
Default: ON

ON means the default response to the command prompt is Y (Yes). If the user does not respond to the prompt within a specified time period, the command is executed by default.

OFF means the default response to the command prompt is N (No). If the user does not respond to the prompt within a specified time period, the command is not executed.

To specify the time period before the response is executed, use the Command Line Prompt Timeout parameter.

Allow Audit Passwords = value

Specify whether passwords can be used to identify auditors.

Supported values: ON, OFF
Default: OFF

Enable SECURE.NCF=value

Execute the secure.ncf file at server startup.

Supported values: ON, OFF
Default: OFF

You can set this parameter in the AUTOEXEC.NCF or STARTUP.NCF file.

Display Incomplete IPX Packet Alerts = value

Specify whether alert messages are displayed when IPX receives incomplete packets. This parameter can be used as a debugging tool.

Supported values: ON, OFF
Default: ON

You can set this parameter in the STARTUP.NCF file.


Multiprocessor Parameters

Multiprocessor parameters allow you to

Multiprocessor parameters are described in the following table.


Table 12. Multiprocessor Parameters

Parameter Use to

System Threshold = number

Adjust the load balancing threshold for multiprocessor servers.

This number controls the main value used in calculating thread shedding for load balancing across multiple processors. The optimum value has been preset by Novell.

Supported values: 0 to 102400
Default: 1536

Important: The default value is the optimum value. We recommend that you retain it. If after careful consideration you decide to change the setting, experiment only in a nonproduction environment.

You can set this parameter in the STARTUP.NCF file.

Auto Start Processors = value

Start secondary processors on startup.

Supported values: ON, OFF
Default: ON

ON = Secondary processors are automatically started when the Platform Support Module (PSM) is loaded.

OFF = Secondary processors are not automatically loaded on bootup. When this is the case, the command START PROCESSORS must be used at the console prompt to activate secondary processors.

This parameter is preferably set in the STARTUP.NCF file.

Auto Clear Interrupt Statistics = value

Specify whether interrupt statistics are removed from memory at the time a processor is taken offline or an interrupt handler (ISR) is removed.

Supported values: ON, OFF
Default: ON

To retain per-processor interrupt handler statistics for offline processors or to retain the total interrupt contribution from a previously loaded handler (ISR), set the parameter to OFF.

This parameter is preferably set in the STARTUP.NCF file.


NCP Parameters

With NetWare Core ProtocolTM (NCPTM) parameters, you can

For a description of NCP parameters, see Table 13.


Table 13. NCP Parameters

Parameter Use to

NCP Packet Signature Option = number

Control the NCP packet signature level on the server.

Supported values:

0 = Server doesn't sign packets (regardless of the client level)

1 = Server signs packets only if the client requests it (client level is 2 or higher)

2 = Server signs packets if the client is capable of signing (client level is 1 or higher)

3 = Server signs packets and requires all clients to sign packets (or logging in will fail)

Default: 1

NCP Packet Signature prevents packet forgery on servers and clients using NCP by requiring server and client to sign each NCP packet.

Note: Because Packet Signature consumes processor resources and slows performance on both client and server, NCP Packet Signature is optional.

After starting the server, you can only increase the level of packet signature. To decrease the level, you must add the SET command to the startup.ncf file and restart the server.

You can set this parameter in the STARTUP.NCF file.

Enable IPX Checksums = number

Enable IPXTM checksums.

Supported values:

0 = No checksums
1 = Checksum if enabled at the client
2 = Require checksums

Default: 1

You can set this parameter in the STARTUP.NCF file.

Enable UDP Checksums on NCP Packets

Enable checksumming of NCP UDP packets.

Supported values:

0 = No checksums
1 = Checksum if enabled at client
2 = Require checksums

Default: 1

You can set this parameter in the STARTUP.NCF file.

Client File Caching Enabled

This parameter allows or disallows the client side ecaching of opened files. This parameter is also settable in the STARTUP.NCF file.

NCP Protocol Preferences = value

When multiple protocols are supported, specify the order in which eDirectory selects a protocol to use when communicating with other servers in the replica.

Supported values: IPX, TCP, UDP, none
Default: None

For example, the following command instructs eDirectoryTM to try TCP first and IPX second when it communicates with another server:

SET NCP PROTOCOL PREFERENCES = TCP IPX

If only one protocol is supported, that protocol is used by eDirectory, no matter what values are specified for this parameter.

If multiple protocols are supported, but no value is specified for this parameter, eDirectory uses the protocols in the order in which they are loaded.

NCP File Commit = value

Control whether applications can flush pending file writes to disk.

Supported values: ON, OFF
Default: ON

If the value is set to ON then when a File Commit NCP is issued, a file is sent from cache to disk immediately, instead of waiting for the cache manager to send it to disk later.

Display NCP Bad Component Warnings = value

Control whether NCP bad component alert messages are displayed.

Supported values: ON, OFF
Default: OFF

You can set this parameter in the STARTUP.NCF file.

Reject NCP Packets with Bad Components = value

Specify whether NCP packets that fail component checking are rejected.

Supported values: ON, OFF
Default: OFF

You can set this parameter in the STARTUP.NCF file.

Display NCP Bad Length Warnings = value

Control whether NCP bad length alert messages are displayed.

Supported values: ON, OFF
Default: OFF

You can set this parameter in the STARTUP.NCF file.

Reject NCP Packets with Bad Lengths = value

Specify whether NCP packets that fail boundary checking are rejected. This parameter can be used as a debugging tool.

Supported values: ON, OFF
Default: OFF

You can set this parameter in the STARTUP.NCF file.

Maximum Outstanding NCP Searches = number

Specify the maximum number of NCP directory searches that can be processed simultaneously.

Supported values: 10 to 1000
Default: 51

Normally, only one NCP directory search occurs at a time.

Increase the default only if you use applications that support multiple outstanding directory search operations and you have problems with corrupted or invalid directory information.

Allow Change to Client Rights = value

Control whether a job server can assume the rights of a client for NCP packet signatures.

Supported values: ON, OFF
Default: ON

Note: Some job servers and third-party applications can't function without changing to client rights.

Using OFF might prevent some job servers from getting access to the files they need, but it prevents the forging of a packet through the job or print server.

You can set this parameter in the STARTUP.NCF file.

Allow LIP = value

Set Large Internet Packet (LIP) support.

Supported values: ON, OFF
Default: ON

You can set this parameter in the STARTUP.NCF file.


Novell Storage Service Parameters

For Novell Storage Service set parameters, see NSS Load Commands, NSS DOS FAT Commands, PURGE and SALVAGE Commands, NSS Buffer Cache Commands, and Other NSS Commands.


Service Location Protocol Parameters

The service location protocol parameters define how SLP locates and distributes information about services that are available on the network.

For descriptions of service location protocol parameters, see Table 14.


Table 14. Service Location Protocol Parameters

Parameter Use to

SLP TCP = value

Use TCP packets instead of UDP packets when possible.

Supported values: OFF, ON
Default: OFF

This parameter can be set in the STARTUP.NCF file.

SLP Debug = value

Enable SLP debug mode.

  • Bit 0x01 = COMM
  • Bit 0x02 = TRAN
  • Bit 0x04 = API
  • Bit 0x08 = DA
  • Bit 0x10 = ERR
  • Bit 0x20 = SA

Supported values: 0 to 4294967255
Default: 0

This parameter can be set in the STARTUP.NCF file.

SLP DA Discovery Options = value

Use multicast DA advertisements.

  • Bit 0x01 = Use multicast Directory Agent advertisements
  • Bit 0x02 = Use DHCP discovery
  • Bit 0x04 = Use static file SYS:ETC\\SLP.CFG
  • Bit 0x08 = Scopes Required.

These bits can be ordered together for multiple values.

Supported values: 0 to 8
Default: 3

This parameter can be set in the STARTUP.NCF file.

SLP Multicast Radius = value

Specify an integer describing the multicast radius.

Supported values: 0 to 32
Default: 32

This parameter can be set in the STARTUP.NCF file.

SLP Broadcast = value

Use broadcast packets instead of multicast packets.

Supported values: OFF, ON
Default: OFF

This parameter can be set in the STARTUP.NCF file.

SLP MTU Size = value

Specify an integer describing the maximum transfer unit size.

Supported values: 0 to 4294967255
Default: 1472

This parameter can be set in the STARTUP.NCF file.

SLP Rediscover Inactive Directory Agents = value

Specify the minimum time period in seconds that SLP will wait to issue service requests to rediscover inactive Directory Agents.

Supported values: 0 to 4294967255
Default: 60

This parameter can be set in the STARTUP.NCF file.

SLP Retry Count = value

Specify an integer value describing the maximum number of retries.

Supported values: 0 to 128
Default: 3

This parameter can be set in the STARTUP.NCF file.

SLP Scope List = value

Specify a comma-delimited scope policy list.

Max Length: 1023
Default: 1023

This parameter can be set in the STARTUP.NCF file.

SLP SA Default Lifetime = value

Specify an integer value describing the default lifetime in seconds of service registers.

Supported values: 0 to 4294967255
Default: 900

This parameter can be set in the STARTUP.NCF file.

SLP Event Timeout = value

Specify an integer value describing the number of seconds to wait before timing out multicast packet requests.

Supported values: 0 to 4294967255
Default: 53

This parameter can be set in the STARTUP.NCF file.

SLP DA Heart Beat Time = value

Specify an integer value describing the number of seconds before sending the next Directory Agent heartbeat packet.

Supported values: 0 to 4294967255
Default: 10800

This parameter can be set in the STARTUP.NCF file.

SLP Close Idle TCP Connections Time = value

Specify an integer value describing the number of seconds before idle TCP connections should be terminated.

Supported values: 0 to 4294967255
Default: 300

This parameter can be set in the STARTUP.NCF file.

SLP DA Event Timeout = value

Specify an integer value describing the number of seconds to wait before timing out Directory Agent packet requests.

Supported values: 0 to 429
Default: 5

This parameter can be set in the STARTUP.NCF file.


Time Parameters

Time parameters control time synchronization, the TIMESYNC.CFG file, and time zone settings to ensure that the time reported by all servers is consistent, or synchronized. For more information, see the Network Time Management Administration Guide.

For a description of time parameters, see Table 15.


Table 15. Time Parameters

Parameter Use to

TIMESYNC ADD Time Source = server name

Specify a server as a time source.

Use EDIT, rather than this parameter, to add a server to the time source list in the TIMESYNC.CFG file.

Maximum: 48 characters

TIMESYNC Configuration File = path

Specify the path where the TIMESYNC.CFG configuration file is located.

Maximum: 255 characters

Example: SET TIMESYNC CONFIGURATION FILE = SYS:SYSTEM\TIMESYNC.CFG

TIMESYNC Configured Sources = value

Specify what time sources the server listens to.

Supported values:

ON = The server ignores SAP time sources and relies on time sources custom-configured with the TIMESYNC Time Source parameter

OFF = The server listens to any advertising time source

Default: OFF

TIMESYNC Directory Tree Mode = value

Control the use of SAP packets in the eDirectoryTM tree.

Supported values:

ON = Time synchronization ignores SAP packets that don't originate from within the eDirectory tree the server is on

OFF = The server can receive SAP packets from any time source on the network

Default: ON

Don't set this parameter to OFF if SAP is set to ON. Using OFF could corrupt the time synchronization for this server's eDirectory tree.

TIMESYNC Hardware Clock = value

Controls hardware clock synchronization.

Supported values:

ON = The primary and secondary time servers set the hardware clock, and the single reference and reference time servers set their time from the hardware clock at the beginning of each polling interval

OFF = Use only if this server uses an external time source (such as a radio clock)

Default: ON

TIMESYNC Polling Count = number

Specify how many time packets to exchange while polling. Increasing the number of packets adds unnecessary traffic to the network.

Supported values: 1 to 1000
Default: 3

TIMESYNC Polling Interval = number

Specify the long polling interval, in seconds.

Supported values: 10 to 2678400 seconds (31 days)
Default: 600 seconds (10 minutes)

All servers in the same tree must use the same setting.

TIMESYNC REMOVE Time Source = server_name

Delete a server as a time source.

Use EDIT, rather than this parameter, to delete a server from the time source list in the TIMESYNC.CFG file.

Maximum: 48 characters

TIMESYNC RESET = value

Reset time synchronization and clear the time source list.

Use EDIT, rather than this parameter, to reset values in the TIMESYNC.CFG file and to remove the time source list from the file.

Supported values:

ON = Selected internal values are reset and the configured server list is cleared. Flag automatically resets to OFF.

Default: OFF

TIMESYNC Restart Flag = value

Control restarts of time synchronization.

Supported values: ON, OFF
Default: OFF

Set this parameter to ON only if you want to reload TIMESYNC without rebooting the server.

TIMESYNC Service Advertising = value

Control time source advertising.

Supported values:

ON = The single reference, reference, and primary time source servers advertise using SAP

OFF = Use only if you are using a custom-configured list of time sources

Default: ON

Note: Secondary time services do not advertise.

TIMESYNC Synchronization Radius = number

Control the maximum time adjustment (in milliseconds) a server is allowed while still being considered synchronized.

Supported values: 0 to 2147483647 milliseconds
Default: 2000

Increase this parameter to allow a wider margin of error for time synchronization between servers.

Important: Lowering the synchronization radius increases the chance of servers losing synchronization due to randomness between clocks.

Setting the synchronization radius for under 2000 milliseconds (2 seconds) is not recommended.

Set the synchronization radius for under two seconds only if you are using an application that uses synchronized time stamps which do not tolerate a two-second deviation between time sources.

TIMESYNC Time Adjustment = [+ or -] hour:minute:second [at month/day/year hour:minute:second AM or PM]

Specify when a time adjustment will take place.

Note: You cannot use this parameter on a secondary time server.

Maximum: 99 characters
Default: None scheduled

Use this parameter sparingly to correct network-wide time errors.

Default date and time is six polling intervals or 1 hour (whichever is longer) from the current time.

Misuse of this parameter can corrupt time synchronization and the order of events on your network.

TIMESYNC Time Source: server_name

Specify a server as time source. If no server name is entered, the parameter displays the list of configured servers.

Use EDIT, rather than this parameter, to add a server to the configuration list in the TIMESYNC.CFG file.

Maximum: 48 characters

TIMESYNC Type = type_of_time_source

Specify the default time source type.

Use EDIT, rather than this parameter, to specify the default time source type in the TIMESYNC.CFG file.

Supported types: reference, primary, secondary, single reference
Default: Single reference

TIMESYNC Write Parameters = value

Specify whether parameters specified by the TIMESYNC Write Value parameter are written to the configuration file.

Supported values: ON, OFF
Default: OFF

TIMESYNC Write Value = number

Control which parameters are written by TIMESYNC Write parameters.

Supported values:

1 = Write internal parameters only
2 = Write configured time sources only
3 = Write both parameters and configured time sources

Default: 3

Time Zone = time_zone_string

Specifies the time zone string, which indicates:

  • The abbreviated time zone name
  • The offset from Coordinated Universal Time (UTC)
  • The alternate abbreviated time zone name to be used when daylight saving time is in effect

Maximum: 80 characters
Default: No Time Zone

This parameter causes UTC time to be recalculated from local time.

You can set this parameter in the STARTUPCF file.

Default Time Server Type = type_of_time_source

Specify the default time synchronization server type. This parameter can be overridden by other time synchronization parameters.

Supported types: Reference, primary, secondary, single reference.

Supported types: Reference, primary, secondary, single reference
Default: Secondary

You can set this parameter in the STARTUPCF file.

Start of Daylight Savings Time = date_and_time

Specify the local date and time when the change to daylight saving time should occur.

Maximum: 79 characters

IMPORTANT:  

You must set both the start and end of daylight saving time before either date is actually scheduled.

End of Daylight Savings Time = date_and_time

Specify the local date and time when the change from daylight saving time should occur.

Maximum: 79 characters.

IMPORTANT:  

You must set both the start and the end of daylight saving time before either date is actually scheduled.

Daylight Savings Time Offset = [+ or -] hour:minute:second

Control the offset applied to time calculations when daylight saving time is in effect.

Default: +1:00:00

This parameter causes UTC time to be recalculated from local time.

You can set this parameter in the STARTUP.NCF file.

Daylight Savings Time Status = value

Indicate whether daylight saving time is in effect.

Supported values: ON, OFF
Default: OFF

If this parameter is set to ON, you should also use the Daylight Savings Time Offset parameter.

Changing the daylight saving time status does not change the local time.

You can set this parameter in the STARTUP.NCF file.

New Time With Daylight Savings Time Status = value

Control the adjustment of local time when daylight saving time is in effect.

Supported values: ON, OFF
Default: OFF

ON = Adjust the local time by adding or subtracting the Daylight Savings Time Offset parameter.


Traditional File Parameters

Traditional file parameters now contain parameters that were previously designated as Directory Caching, File Caching, Error Handling, File System, Locks, and Transistion Tracking


File Caching Parameters for the Traditional File System

File caching allows faster access to frequently used files by holding a file (or a portion of it) in disk cache memory. Files being read from or written to are kept in file cache buffers. When the requested data is already in cache, disk reads are unnecessary.

The number of files kept in memory depends on the number of file cache buffers allowed. This is determined by the amount of disk cache memory available and the value of the Minimum File Cache Buffers parameter.

Disk cache memory not only speeds up access to file data, it is used to cache portions of the eDirectory database. If you want to tune your NetWare server in general, or eDirectory in particular, tune the file caching parameters. The key to optimizing cache utilization is to monitor both Long Term Cache Hits and LRU Sitting Time statistics. Both parameters are found in MONITOR's Available Options > Disk Cache Utilization. (File caching parameters can also be set from MONITOR's Available Options > Server Parameters.)

File caching and directory caching work together, however. In tuning the server, directory caching and file caching need to be balanced for maximum performance. See How Directory Caching and File Caching Work Together.

Remember also that the file and directory caching parameters in SET apply only to NetWare servers using the Traditional File System. In contrast, the Novell® Storage ServicesTM (NSSTM) file system provides improved resource use, requiring less memory to mount multiple volumes.

For information about NSS directory caching parameters, see NSS Buffer Cache Commands and NSS Features and Benefits.

For a description of file caching parameters, see Table 16.


Table 16. File Caching Parameters for the Traditional File System

Parameter Use to

Read Ahead Enabled = value

Specify whether the system conducts background reads to cache in advance of blocks soon to be requested.

Supported values: ON, OFF
Default: ON

Read-aheads can happen only when sequential file accesses are occurring.

Read Ahead LRU Sitting Time Threshold = number

Specify the minimum cache LRU (least recently used) sitting time for read-aheads to take place.

Supported values: 0 seconds to 1 hour
Default: 10 seconds

Minimum File Cache Buffers = number

Set the minimum number of cache buffers the operating system can allow for file caching.

Supported values: 20 to 1000
Default: 20

All memory not allocated for other processes is given to disk cache memory to use for file and directory caching. As memory is requested for other processes, the server gives up cache buffers.

This limit specifies when the server must stop giving file cache buffers to other processes.

If you set the minimum too high, other processes might not be able to allocate necessary memory resources.

For example, a module could fail to load because the server is out of available memory.

Maximum Concurrent Disk Cache Writes = number

Specify the number of write requests for changed file data that can be put in the elevator before the disk head begins a sweep across the disk.

Supported values: 10 to 4000
Default: 50

A high number creates more efficient write requests. A low number creates more efficient read requests.

Monitor the number of Dirty Cache Buffers displayed on the Cache Utilization Statistics screen in MONITOR.

If this number is above 70 percent of total cache buffers, optimize the write speed by increasing this parameter.

Dirty Disk Cache Delay Time = time

Specify how long the system keeps a write request (that does not fill a cache buffer) in memory before writing the request to disk.

Supported values: 0.1 second to 10 seconds
Default: 3.3 seconds

Increasing the time makes disk writing more efficient if your users make many small write requests.

Decreasing the time can reduce performance drastically, and reduces the chances of losing data only slightly.

Minimum File Cache Report Threshold = number

Specify how few cache buffers can be available before the operating system warns you that the number of buffers is getting low.

Supported values: 0 to 1000
Default: 2

For example, if the value of the Minimum File Cache Buffers parameter is set to 20 and this parameter is set to 25, you are warned when all but 45 cache buffers are allocated for other processes.

Regardless of how this parameter value is set, the operating system will issue a warning when it reaches the minimum number of cache buffers. You receive the following message: Cache memory allocator exceeded minimum cache buffer left limit.


Directory Caching Parameters for the Traditional File System

Directory caching allows fast access to frequently used directories. A directory cache buffer is a portion of NetWare® server memory (disk cache memory) that holds entries from the directory table.

A directory entry stays in a cache buffer as long as it is accessed frequently (default=33 seconds). The system can overwrite the directory entry if the allocated directory cache buffers are accessed more frequently.

Directory caching and file caching work together, however. In tuning the server, directory caching and file caching need to be balanced for maximum performance.

Remember also that the directory and file caching parameters in SET apply only to NetWare servers using the Traditional File System. In contrast, the Novell Storage Services (NSS) file system provides improved resource use, requiring less memory to mount multiple volumes.

For information about NSS caching parameters, see NSS Features and Benefits and NSS Buffer Cache Commands.

For a description of directory caching parameters, see Table 17.


Allocation of Directory Cache Buffers

When the server boots, the system allocates a minimum number of directory cache buffers (default=20). It creates these buffers immediately when it receives a request for a new buffer.

The server obtains directory cache buffers by removing a minimal set of file cache buffers from the file cache during server startup. Additional directory cache buffers are obtained from the file cache later on if they are needed.

When the minimum number of buffers is allocated and another is needed, the system must wait a specified amount of time before allocating another buffer.

If necessary, the system allocates buffers until it reaches the maximum. If enough directory cache buffers are allocated and enough memory is available for directory caching, all directory tables can be cached in memory.


How Directory Caching and File Caching Work Together

As directory cache buffers increase, file cache buffers decrease. Thus, a tradeoff exists between directory caching and file caching, and .directory caching and file caching must be balanced for maximum performance.

If you adjust the system to use too much memory for directory caching, you can leave too little memory for file caching, and vice versa.

For a description of file caching parameters, see Table 16. For a description of directory caching parameters, see Table 17.


Table 17. Directory Caching Parameters for the Traditional File System

Parameter Use to

Dirty Directory Cache Delay Time = time

Specify how long the system keeps a directory table write request in memory before writing it to disk.

Supported values: 0 to 10 seconds
Default: 0.5 second

Increasing the delay time gives slightly quicker performance but increases the probability of the directory tables becoming corrupted.

Decreasing the delay time slightly reduces the chance of directory tables becoming corrupted, but can reduce performance slightly.

A zero setting reduces performance dramatically.

Directory Cache Allocation Wait Time = number

Specify how long the system must wait after allocating a new directory cache buffer before it can allocate another buffer.

Supported values: 0.5 second to 2 minutes
Default: 2.2 seconds

During this time, all requests for a new directory cache buffer are ignored.

If the wait time is too low, peak usage requests cause more resources than necessary to be allocated to directory caching.

If the wait time is too high, the system is very slow in allocating the directory cache buffers necessary to service the usual number of directory requests.

If directory searches seem slow even after the server has been running for 15 minutes, you should decrease this parameter.

Directory Cache Buffer NonReferenced Delay = time

Specify how long a directory entry must be cached before it can be overwritten by another directory entry.

Supported values: 1 second to 5 minutes
Default: 5.5 seconds

Increasing this parameter speeds up directory access. The system allocates more directory cache buffers, and a directory is more likely to be cached in memory.

Decreasing this parameter slows directory access but also reduces the need for directory cache buffers.

Maximum Directory Cache Buffers = number

Specify the maximum number of cache buffers that the system can allocate for directory caching.

Supported values: 20 to 4000
Default: 500

This parameter keeps the system from allocating too many directory cache buffers so that memory is available for other server processes.

Increase this limit if the server responds slowly to directory searches.

Decrease this limit if too much memory is being allocated for directory caching.

If users are warned that the server is low on memory, this parameter should be one of the first to be reduced. (Reboot the server to return the memory to the cache buffer memory pool.)

When a directory cache buffer is allocated, the allocation is permanent until the server reboots; the buffers don't return to file caching when the need for directory cache buffers decreases.

This parameter can be set in the STARTUP.NCF file.

Minimum Directory Cache Buffers = number

Specify the minimum number of cache buffers that the system can allocate for directory caching.

Supported values: 10 to 2000
Default: 20

The number needs to be high enough that directory searches can be done quickly, but no higher than necessary.

If the system doesn't need the minimum number of directory cache buffers, the buffers can't be reallocated to file caching. The unneeded portion remains unused.

If the server responds slowly to directory searches after it is booted, monitor the number of directory cache buffers usually allocated for directory caching.

If the number is significantly higher than this limit, consider increasing the limit to remove the delay time that normally occurs while the server is self-configuring. Use MONITOR to view the current statistics.

This parameter can be set in the STARTUP.NCF file.

Maximum Number of Internal Directory Handles = number

Specify the maximum number of directory handles available for internal NLM programs that use connection 0.

Supported values: 40 to 1000
Default: 100

A directory handle is a version of the directory access rights that is held in cache memory. Caching the access rights speeds mapping to the rights.

Connection 0 (zero) is the connection number reserved for use by the server itself and by NLMTM programs operating within the server.

Each time an NLM using connection 0 accesses a file or directory, a directory handle is allocated (up to the total specified by this parameter). This value represents the total number of handles available to be shared by all NLM programs using connection 0.

Maximum Number of Directory Handles = number

Specify the maximum number of directory handles available for each connection.

Supported values: 20 to 1000
Default: 20

A directory handle is a version of the directory access rights that is held in cache memory. Caching the access rights speeds mapping to the rights.

Each time a connection accesses a file or directory, a directory handle is allocated (up to the total specified by this parameter).


File Parameters for the Traditional File System

For a description of file system parameters, see the following table.


Table 18. File Parameters for the Traditional File System

Parameter Use to

Automatically Repair Bad Volumes = value

Specify whether VREPAIR runs automatically on a volume that fails to mount.

Supported values: ON, OFF
Default: ON

You can set this parameter in the STARTUP.NCF file.

File Delete Wait Time = time

Specify when a salvageable file can be purged to create free space on a volume.

Supported values: 0 seconds to 7 days
Default: 5 minutes 29.6 seconds

Set this parameter as high as is useful for your users, but this parameter doesn't guarantee that a file remains salvageable.

The system tries to keep at least 1/32 of available space on the volume free for new files.

This parameter guarantees only that files aren't purged to maintain this free disk space.

Files that haven't met this time limit are purged if the volume is full and the system needs space for a user to create a new file.

When a deleted file remains on the server longer than the File Delete Wait Time parameter setting, the system marks the file as purgeable.

When the volume is full of purgeable files and needs free space, the oldest purgeable files are purged.

Allow Deletion of Active Directories = value

Specify whether a directory can be deleted when another connection has a drive mapped to it.

Supported values: ON, OFF
Default: ON

Maximum Percent of Volume Space Allowed for Extended Attributes = percentage

Limit the portion of volume space used for extended attribute storage. The setting takes effect only when the volume is being mounted.

Supported values: 5 to 50
Default: 10

Maximum Extended Attributes per File or Path = number

Limit the number of extended attributes that can be assigned to a file or path (subdirectory). This limit applies to all volumes on the server.

Supported values: 4 to 512
Default: 16

Fast Volume Mounts

Increase the speed at which volumes are mounted.

Supported values: ON, OFF
Default: ON

The operating system checks only the most important fields during the mounting process. Use this parameter only if volumes were dismounted normally.

This parameter can be set in the STARTUP.NCF file.

Maximum Percent of Volume Used by Directory = percentage

Limit the portion of a volume that can be used as directory space.

Supported values: 5 to 85
Default: 13

Maximum Subdirectory Tree Depth = number

Specify how many levels of subdirectories the system supports.

Supported values: 10 to 100
Default: 25

This parameter can be set in the STARTUP.NCF or in the AUTOEXEC.NCF file, or at the server console.

Increase the number if your applications support trees deeper than 25.

Decrease the number if your applications support only shallow tree structures.

(Some DOS applications can't support more than 10 levels if the subdirectories have 11-character names.)

Volume Low Warn All Users = value

Have the system notify users when a volume is almost full.

Supported values: ON, OFF
Default: ON

If you choose not to have your server warn users, monitor volume statistics at least daily with MONITOR.

Volume Low Warning Reset Threshold = number

Control how much disk space must be freed up before a second warning is issued that the volume is almost full. (The first warning is controlled by the Volume Low Warn All Users parameter.)

Supported values: 0 to 100000 blocks
Default: 256 blocks

When a volume is almost full, it can hover at its warning threshold as users create and delete files.

This parameter controls the minimum amount of space that must be made available above the threshold before the warning message disappears.

Example: Assume the volume block size is 4 KB, the Volume Low Warning Reset Threshold is set to 256, the Volume Low Warning Threshold is set to 256, and the volume has less than 1 MB of free space.

Given these conditions, the volume must gain at least 1 MB of free space (for a total of 2 MB of free space) and then dip below the 1 MB threshold before the system sends another warning that the volume is almost full.

Volume Low Warning Threshold = number

Specify the number of blocks of free disk space that can remain on a volume before the system issues a warning.

Supported values: 0 to 1000000 blocks
Default: 256 blocks

Consider the following: A block is the minimum space allocated to a file.; a file can grow only in multiples of the block size.

The block's physical size is determined when the volume is created. A volume can be assigned these block sizes: 4 KB, 8 KB, 16 KB, 32 KB, or 64 KB.

For example, if you enter 256 blocks for this parameter and the volume's block size is 4 KB, the system warns you that the volume is full when about 1 MB of space is left.

If your volumes are assigned different block sizes, each volume has a different amount of free space when the warning is issued.

Purge Files On Dismount = value

Have the system purge deleted files on a volume when it is dismounted.

Supported values: ON, OFF
Default: OFF

Turbo FAT Re-Use Wait Time = time

Specify how long a turbo FAT buffer remains in memory after an indexed file is closed.

Supported values: 0.3 second to 1 hour 5 minutes 54.6 seconds
Default: 5 minutes 29.6 seconds

Once the wait-time value has passed, the system can allocate the buffer to another indexed file.

Increase the wait time if

  • You want the turbo FAT index to remain in memory for long periods of time, even when the file is closed.
  • You frequently reopen the same file after a specific delay and know that another file opened during that delay will reuse the index.

Decrease the wait time if you want the memory released immediately to service the next file that needs to be indexed.

When a program randomly accesses a file that contains more than 64 FAT entries, the system builds a turbo FAT index for the file so that information is accessed quickly. (NetWare indexes any randomly accessed file with 64 FAT entries.)

The system doesn't immediately delete the index from its buffer when the file is closed.

Having the turbo FAT index in memory makes reopening the file and accessing information faster.

Compression Daily Check Stop Hour = number

Specify the hour when you want the file compressor to stop scanning enabled volumes for files that need to be compressed.

Supported values: 0 to 23
Default: 6

Hours are specified by a 24-hour clock: 0 = midnight; 23 = 11 p.m.

This parameter can be set in the STARTUP.NCF file.

Allow Unowned Files To Be Extended = value

Specify whether files can be modified when the owner has been lost or deleted.

Supported values: ON, OFF
Default: ON

This parameter can be set in the STARTUP.NCF file.


Locks Parameters for the Traditional File System

Locks parameters control

There are three types of locks:

For a description of locks parameters, see the following table.


Table 19. Locks Parameters for the Traditional File System

Parameter Use to

Maximum Record Locks Per Connection = number

Control how many record locks a client connection can use at one time. (Use MONITOR to view how many record locks each connection is using.)

Supported values: 10 to 100000
Default: 500

Increase the value of this parameter when an application can't lock enough records and fails.

Decrease the value of this parameter if one or more client connections are using an excessive amount of server resources.

Maximum File Locks Per Connection = number

Control how many opened and locked files a client connection can use at one time. (Use MONITOR to view how many opened and locked files a connection is using.)

Supported values: 10 to 1000
Default: 250

Increase the value of this parameter when an application can't open enough files and fails.

Decrease the value of this parameter if client connections are using an excessive amount of server resources.

Maximum Record Locks = number

Control how many record locks the operating system can handle. (Use MONITOR to view how many record locks each client connection is using on that server.)

Supported values: 100 to 400000
Default: 20000

Increase the value of this parameter if users have problems running applications and receive messages indicating that not enough record locks are available.

Decrease the value of this parameter if client connections are using an excessive amount of server resources.

Maximum File Locks = number

Control how many opened and locked files the operating system can handle. (Use MONITOR to view the number of files that are open during peak usage.)

Supported values: 100 to 100000
Default: 10000

Increase the value of this parameter if the number of open files is near or equal to the default.

Decrease the value of this parameter to restrict the amount of available server resources.


Error Handling Parameters for the Traditional File System

Error handling parameters control the size of error logs and specify what happens when logs exceed the specified size. They also control how the server responds to an abend or to an NLMTM that does not unload from a protected address space.

For descriptions of error handling parameters, see the following table.


Table 20. Error Handling Parameters for the Traditional File System

Parameter Use to

Volume Log File State = number

Control what happens when the VOL$LOG.ERR file is larger than the size specified by the Volume Log File Overflow Size parameter.

Supported values:

0 = Leave VOL$LOG.ERR as is
1 = Delete VOL$LOG.ERR
2 = Rename VOL$LOG.ERR

Default: 1

You can set this parameter in the STARTUP.NCF file.

Volume TTS Log File State = number

Control what happens when the TTS$LOG.ERR file is larger than the size specified by the Volume TTS Log File Overflow Size parameter.

Supported values:

0 = Leave TTS$LOG.ERR as is
1 = Delete TTS$LOG.ERR
2 = Rename TTS$LOG.ERR

Default: 1

You can set this parameter in the STARTUP.NCF file.

Volume Log File Overflow Size = number

Specify the maximum size of the VOL$LOG.ERR file before the action specified by the Volume Log File State parameter occurs.

Supported values: 65536 to 4294967295
Default: 4194304

You can set this parameter in the STARTUP.NCF file.

Volume TTS Log File Overflow Size = number

Specify the maximum size of the TTS$LOG.ERR file before the action specified by the Volume TTS Log File State parameter occurs.

Supported values: 65536 to 4294967295
Default: 4194304

You can set this parameter in the STARTUP.NCF file.


Transaction Tracking Parameters for the Traditional File System

For a description of transaction tracking parameters for the Traditional File System, see the following table.


Table 21. Transaction Tracking Parameters for The Traditional File System

Parameter Use to

Auto TTS Backout Flag = value

Specify whether a server with transactional files will automatically back out of incomplete transactions when the server is restarted.

Supported values:

ON = The server backs out incomplete transactions when booting
OFF = The server waits for you to answer a prompt before it backs out of transactions

Default: ON

You must set this parameter in the STARTUP.NCF file. You cannot set the parameter at the command line.

TTS Abort Dump Flag = value

Specify whether a file is created to log transactional backout data.

Supported values:

ON = The information is saved in TTS$LOG.ERR
OFF = The information backed out is not saved

Default: OFF

If the parameter is set to ON, the backed-out information is written to file TTS$LOG.ERR on volume SYS:. You can print the file or view it with a text editor.

TTS UnWritten Cache Wait Time = time

Specify how long a block of transactional data can be held in memory.

Supported values: 11 seconds to 10 minutes 59.1 seconds Default: 1 minute 5.9 seconds

Some blocks of transactional data wait for other transactional blocks to be written first.

If one of these blocks reaches its maximum time limit, other write requests are held up and this block is written to disk.

TTS Backout File Truncation Wait Time = time

Specify how long allocated blocks remain available for the TTS backout file when these blocks are not in use.

Supported values: 1 minute 5.9 seconds to 1 day 2 hours 21 minutes 51.3 seconds
Default: 59 minutes 19.2 seconds


Additional Information

Topic See

Using NetWare Remote Manager to set server parameters

Server Parameters option in the NetWare Remote Manager Administration Guide

Using MONITOR to set server parameters

Server Parameters option in MONITOR

Improving server performance

Server Operating System Administration Guide > Optimizing the NetWare Server

Editing .NCF files

Server Operating System Administration Guide

Time synchronization and the TIMESYNC.CFG file

Network Time Management Administration Guide



Previous | Next