Architecture, Administration, Operations, and Migration
NetWare 3.12 through 6.5 SP8 | Bindery, NDS/eDirectory, TFS, NSS, IPX, and native IP
Version 1.0 | 8 August 2026
For legacy operations, recovery, technical archaeology, and controlled migration
Contents
Select an entry in Word or a compatible viewer to jump to that section. Major sections begin on a new page for field use.
Compatibility questions to ask first
Architecture and operating model
Bindery versus NDS and eDirectory
Boot, configuration, and runtime
Safe boot and recovery switches
NLMs and protected address spaces
Networking and service discovery
Native IP and SLP environments
NDS and eDirectory administration
Partitions, replicas, and replica rings
Time synchronization and directory health
Directory administration tools
eDirectory object and property rights
Effective rights and inherited rights filters
Storage, volumes, and file systems
Traditional file system versus NSS
Namespaces and path compatibility
Salvage, purge, quotas, and capacity
Clients, drive mappings, and login scripts
Administration quick reference
Console command quick reference
Workstation utility quick reference
Daily, weekly, and monthly checks
Controlled maintenance shutdown
Server will not start or SYS will not mount
Authentication or login script failure
Access denied or files are invisible
Slow response or high utilization
eDirectory synchronization errors
What a usable backup must preserve
Printing and ancillary services
Containment, preservation, and migration
Virtualization and historical preservation
Appendix A – Common paths and files
Appendix C – Official source set
How to use this reference Back to contents
This is a practical reference for engineers who must understand, recover, operate, or retire a Novell NetWare environment. It is not a replacement for the manual matching the exact server version, support pack, hardware driver set, eDirectory build, and installed applications.
WORKING ASSUMPTION: The operational detail is centered on NetWare 4.x through 6.5, while NetWare 3.12 and Bindery behavior are called out where they differ. Commands marked as examples must be validated on the target server before use.
Scope and conventions Back to contents
- Server-console commands appear in uppercase for readability; NetWare commands are generally not case-sensitive.
- A path such as SYS:SYSTEM identifies a volume and directory. A path such as SERVER/SYS:PUBLIC also identifies the server.
- NDS refers to Novell Directory Services; later documentation uses eDirectory. In this guide, NDS/eDirectory means the directory service family.
- TFS means the NetWare Traditional File System. NSS means Novell Storage Services.
- Source markers such as [S2] refer to the official source set in Appendix C.
Support and security status Back to contents
NetWare 6.5 SP8 is the terminal NetWare release line. It entered extended support in 2010, and the vendor’s later Premium Lifeline offering ended on 31 December 2016. It must therefore be treated as unsupported legacy infrastructure in 2026. [S1, S13]
SECURITY BOUNDARY: Do not expose NCP, SLP, IPX routing, Telnet, RConsoleJ, legacy web administration, LDAP, or old TLS endpoints directly to the Internet or to an untrusted enterprise segment. Place the server behind an allow-list firewall on an isolated VLAN and administer it through a controlled jump host or modern encrypted tunnel.
- Use unique legacy credentials; do not reuse current privileged passwords.
- Disable services and protocols that are not required, particularly Telnet, IPX, anonymous LDAP, and legacy web components.
- Keep ALLOW UNENCRYPTED PASSWORDS set to OFF unless a documented, temporary compatibility exception exists. [S2]
- Assume that old cryptographic implementations and browser-based interfaces do not meet modern security baselines.
- Capture configuration and recovery media before every change because replacement drivers, patches, and vendor support are scarce.
Version families at a glance Back to contents
Version family summary
| Family | Directory model | Network emphasis | Operational significance |
| 2.x | Per-server Bindery | IPX/SPX | Dedicated 286-era file server; highly version- and hardware-specific. |
| 3.x / 3.12 | Per-server Bindery | IPX/SPX with SAP/RIP | 32-bit 386 line; NLM model; mature departmental file and print platform. |
| 4.x / intraNetWare | NDS tree plus Bindery emulation | IPX/SPX; IP add-ons | Introduced global directory, partitions, replicas, and directory-based administration. |
| 5.0 / 5.1 | NDS | Native IP plus optional IPX | NCP became transport-independent; SLP and NSS became central; multiprocessor and memory model advanced. [S14] |
| 6.0 | eDirectory | IP preferred; IPX optional | Expanded web access, iPrint/iFolder era services, and user-oriented licensing. |
| 6.5 / SP8 | eDirectory 8.7.3 or 8.8.x | IP preferred; IPX retained | Final mature NetWare line. New SP8 installs used eDirectory 8.8.4; updated systems could retain 8.7.3. [S1] |
| OES / Enterprise Server | eDirectory on Linux | IP | Successor platform providing NCP, NSS, trustee semantics, CIFS, iPrint, and migration paths without the NetWare kernel. [S12] |
Compatibility questions to ask first Back to contents
- What exact NetWare version, support pack, eDirectory version, JVM, and application build are installed?
- Is the server Bindery-only, NDS/eDirectory-native, or serving legacy clients through Bindery emulation?
- Are clients using IPX, native IP, or both? Which Ethernet frame types and SLP scopes are in use?
- Are volumes Traditional or NSS? Which namespaces, trustee assignments, quotas, compression, encryption, and salvage policies exist?
- Does the server hold directory partitions or replicas, and is it a Master replica, time source, SLP Directory Agent, Organizational CA host, licensing host, or cluster node?
- Which third-party NLMs, backup agents, database engines, and hardware-specific .HAM, .CDM, .LAN, and .PSM drivers are required?
- Are licenses, installation media, overlay media, support packs, driver disks, and keys preserved and legally usable?
Architecture and operating model Back to contents
The NetWare mental model Back to contents
NetWare is a network services operating system, not a general-purpose desktop Unix or Windows server. The kernel is optimized around file, print, directory, protocol, and application services. Administrators interact with a server console and loadable modules; users interact through NCP clients, mappings, login scripts, and directory objects.
Logical layers
| Layer | Examples | Role |
| Clients | DOS requester, VLM, Client32, Novell Client, NetStorage | Authenticate, discover services, map paths, consume file/print services. |
| Directory | Bindery or NDS/eDirectory | Stores identities, groups, servers, volumes, policies, schema, and service objects. |
| Application/file service | NCP, queue print, NDPS/iPrint, GroupWise, Btrieve | Presents network resources and application services. |
| Discovery and transport | SAP/RIP over IPX; SLP over IP; TCP/UDP | Locates services and carries NCP or application traffic. |
| File system | Traditional volumes or NSS pools and volumes | Stores data, trustees, attributes, quotas, namespaces, and salvage metadata. |
| Runtime | SERVER.EXE, NLMs, protected address spaces | Executes kernel services, drivers, protocol stacks, and server applications. |
| Hardware interface | PSM, HAM, CDM, LAN drivers, NWPA | Connects processors, storage, and network adapters to the runtime. |
KEY DISTINCTION: An eDirectory Volume object represents a volume in the directory, but file access is governed by trustee metadata stored in the file system. Directory rights and file-system rights are related administration domains, not interchangeable ACLs. [S8]
Bindery versus NDS and eDirectory Back to contents
Directory model comparison
| Characteristic | Bindery | NDS/eDirectory |
| Scope | One database per server | Distributed tree spanning servers and sites |
| Naming | Flat object names on a selected server | Hierarchical distinguished names in containers |
| Administration | Repeat users/groups on each server | Create identities and policies once in the tree |
| Resilience | Server-local backup and recovery | Partitions and replicas provide distributed availability |
| Legacy support | Native to 2.x/3.x | 4.x+ can expose selected containers as a Bindery context |
| Authentication target | Server | Tree and context, with a server used to reach a replica |
Core services stack Back to contents
- NCP provides file-service semantics, connection management, locking, trustee enforcement, and related client services.
- NDS/eDirectory provides identities, objects, schema, authentication, partitions, and replication.
- NSS provides a journaling file system, storage pools, volumes, trustee metadata, salvage, quotas, compression, and optional encryption.
- IPX/SPX with SAP/RIP supplies legacy transport and discovery; TCP/IP with SLP supplies the later native-IP equivalent.
- NLMs extend the kernel with drivers, protocol stacks, management tools, backup agents, and server applications.
Boot, configuration, and runtime Back to contents
Boot sequence Back to contents
- The machine firmware starts the boot device and the small DOS boot environment used by classic NetWare installations.
- AUTOEXEC.BAT normally changes to C:\NWSERVER and invokes SERVER.EXE.
- SERVER.EXE reads STARTUP.NCF from the boot directory, applies pre-mount SET parameters, and loads platform and storage drivers.
- The server discovers storage and mounts SYS. If SYS cannot mount, SYS:SYSTEM modules and AUTOEXEC.NCF are unavailable.
- SYS:SYSTEM\AUTOEXEC.NCF executes, setting the server identity and loading LAN drivers, protocols, directory services, logging, and installed applications.
- Additional service-specific NCF files are called in their configured order. Users and clients can then discover and connect to the server.
RECOVERY PRINCIPLE: Separate pre-SYS failures from post-SYS failures. STARTUP.NCF, platform support, and storage drivers dominate the first class. AUTOEXEC.NCF, network bindings, directory services, and application NLMs dominate the second.
NCF configuration files Back to contents
Important NCF files
| File | Normal location | Purpose |
| STARTUP.NCF | C:\NWSERVER | Pre-SYS parameters plus platform and storage driver load order. |
| AUTOEXEC.NCF | SYS:SYSTEM | Server identity, network drivers/bindings, services, and application start order. |
| SHUTDOWN.NCF | SYS:SYSTEM | Optional orderly unload or stop commands run by DOWN or restart. [S2] |
| SECURE.NCF | Configured location | Optional commands executed through the secure-start mechanism. |
| Application .NCF | Usually SYS:SYSTEM or application path | Starts or stops a product-specific set of NLMs. |
- Use EDIT or NWCONFIG to change NCF files, and retain a dated known-good copy before editing.
- Place CONLOG near the beginning of AUTOEXEC.NCF if early console messages are needed; the default log is SYS:ETC\CONSOLE.LOG. [S2]
- Only persist a SET parameter after confirming whether it belongs in STARTUP.NCF or AUTOEXEC.NCF. The SET display identifies valid locations. [S2]
- Do not reorder storage, directory, or application modules without documenting dependencies.
Safe boot and recovery switches Back to contents
SERVER and restart switches
| Invocation | Effect | Use |
| SERVER -NS | Skips STARTUP.NCF | Diagnose a bad pre-mount parameter or driver line; storage may need loading manually. |
| SERVER -NA | Skips AUTOEXEC.NCF | Mount SYS but prevent post-mount services and applications from starting. |
| SERVER -S filename.NCF | Uses an alternate startup file | Boot a controlled known-good driver set. [S1, S2] |
| RESTART SERVER -NS | Restarts without STARTUP.NCF | Repeat controlled pre-mount diagnosis. |
| RESTART SERVER -NA | Restarts without AUTOEXEC.NCF | Repeat controlled post-mount diagnosis. |
BEFORE REPAIR: Photograph or capture the console, preserve BOOT$LOG.ERR, CONSOLE.LOG, ABEND.LOG, STARTUP.NCF, AUTOEXEC.NCF, driver versions, and disk layout. Do not begin with VREPAIR, REBUILD, or DSREPAIR repair operations merely because the server failed to boot.
NLMs and protected address spaces Back to contents
LOAD links an NLM or driver into the operating system; UNLOAD releases it and returns resources. Many server utilities can be loaded when needed, while LAN, storage, directory, and protocol modules form persistent dependencies. MODULES lists loaded modules and their address spaces. [S2, S3]
- Kernel address space provides maximum integration but a faulty NLM can abend the server.
- Protected address spaces run suitable applications in ring 3. PROTECT filename.NCF loads the modules from an NCF into a named protected space.
- PROTECTION lists protected spaces and can enable restart behavior. Drivers, SERVER.EXE, and some core modules cannot run protected.
- Unload dependent modules in reverse order. Never force-kill an address space until the data-integrity and vendor implications are understood.
Networking and service discovery Back to contents
IPX and SPX environments Back to contents
Legacy IPX/SPX components
| Component | Function | Diagnostic focus |
| IPX | Connectionless routed network protocol | Network numbers, frame types, bindings, routes |
| SPX | Connection-oriented transport over IPX | Sessions, sequence/retry behavior, compatible stack |
| SAP | Advertises server and service names | DISPLAY SERVERS; hop count; filtering |
| RIP/NLSP | Routes IPX networks | DISPLAY NETWORKS; duplicate network numbers; convergence |
| NCP/IPX | Carries NetWare file and service requests | NCPIPX.NLM, connection state, packet loss |
| ODI | Client LAN driver and protocol interface | LSL, NIC driver, frame type, IPXODI/VLM order |
Common Ethernet frame types include ETHERNET_802.2, ETHERNET_II, ETHERNET_802.3, and ETHERNET_SNAP. A client and server can share the physical Ethernet while remaining logically invisible if frame type or external network numbers do not match.
IPX DISPLAY CAVEAT: DISPLAY SERVERS and DISPLAY NETWORKS show SAP/RIP information. They are not native-IP service-discovery commands; use SLP and TCP/IP tools for IP-only systems. [S2]
Native IP and SLP environments Back to contents
NetWare 5 made NCP transport-independent and introduced a practical pure-IP deployment model. NCP over TCP/UDP uses native IP, while Service Location Protocol (SLP) replaces much of the name-to-address discovery previously supplied by SAP. [S3, S14]
- SLP User Agents issue queries, Server Agents register services, and Directory Agents provide a repository for registrations.
- Named SLP scopes partition discovery information. A server or client that queries the wrong scope can appear unable to find an otherwise healthy service.
- SLP uses TCP and UDP port 427. NCP over IP uses port 524. [S14]
- SYS:ETC\SLP.CFG can define static Directory Agents with DA IPV4 entries; DHCP options 78 and 79 can also supply agents and scopes.
- Directory replication can be affected when NDAP and Bindery service entries are absent from the scopes used by replica servers.
Core ports and flows Back to contents
Common TCP/UDP ports – verify against the installed service configuration
| Port | Protocol/service | Operational note |
| 524 TCP/UDP | NCP over IP / eDirectory service access | Primary Novell client and server service path. |
| 427 TCP/UDP | SLP | Service queries, registrations, and Directory Agent traffic. |
| 389 TCP | LDAP | Directory access; clear-text unless protected externally or upgraded to TLS. |
| 636 TCP | LDAPS | Directory access over legacy TLS; validate certificate and cipher compatibility. |
| 123 UDP | NTP | Time synchronization when XNTPD/NTP is selected. |
| 53 TCP/UDP | DNS | Name service when DNS is hosted or consumed. |
| 80/443 TCP | Apache, iManager, NetStorage, iPrint or application web services | Actual bindings vary by installed pattern and reverse proxy design. |
| 8008/8009 TCP | Novell Remote Manager, commonly | Version/configuration dependent; never expose to an untrusted segment. |
| 413 TCP | SMDR, commonly | Storage Management Services remote backup communication. [S3] |
| 2034-2036 TCP | RConsoleJ agent/proxy variants | Historical remote console ports; firewall-only and version dependent. [S2] |
FIREWALL RULE METHOD: Inventory listening modules and configured bindings on the actual server, capture a known-good traffic trace, then allow only required source/destination pairs. Do not use a generic ‘NetWare ports’ rule set as an exposure baseline.
Client-to-file-service flow Back to contents
- The client obtains a server or tree target from a preferred server, preferred tree, explicit name, SLP, SAP, DNS, or cached configuration.
- The client resolves the service to an IP or IPX address and opens an NCP connection.
- The user authenticates to the Bindery server or to NDS/eDirectory through a server holding or locating the required replica.
- Container, profile, and user login scripts execute and create drive/search mappings.
- NCP evaluates trustee rights, IRFs, security equivalence, file attributes, locks, quotas, and namespace rules for each operation.
NDS and eDirectory administration Back to contents
Tree structure and naming Back to contents
Common eDirectory objects
| Object | Purpose | Typical relationship |
| [Root] | Top of one directory tree | Contains top-level organizations and holds the root partition. |
| O / Organization | Top-level administrative container | Often represents the enterprise. |
| OU / Organizational Unit | Delegation and policy container | Often represents geography, function, or service domain. |
| User | Identity and login properties | Member of groups; may have home directory and login script. |
| Group | Security equivalence and shared assignment | Used for file trustees and application roles. |
| Server | Represents a server | Associated with volumes, addresses, services, and directory replicas. |
| Volume | Directory representation of a volume | Points users and tools to file storage; data rights remain in the file system. |
| Profile | Reusable login script | Assigned to users between container and user scripts. |
| Alias / Directory Map | Alternate object name or path abstraction | Reduces path coupling and supports user-friendly mappings. |
Typed name: CN=PJONES.OU=ARCHITECTURE.O=ACME
Typeless name: PJONES.ARCHITECTURE.ACME
Absolute name: .PJONES.ARCHITECTURE.ACME
Relative name: PJONES (when the current context is ARCHITECTURE.ACME)
Dot notation is written from the leaf toward [Root]. LDAP notation normally reverses the order and separates components with commas, for example CN=PJONES,OU=ARCHITECTURE,O=ACME.
Partitions, replicas, and replica rings Back to contents
- A partition is a contiguous subtree stored and replicated as a unit.
- The Master replica coordinates partition operations. Read/Write replicas accept updates; Read-Only replicas serve reads; Subordinate Reference replicas preserve connectivity across partition boundaries.
- All servers holding a replica of a partition form its replica ring. A healthy ring exchanges changes and agrees on partition and replica metadata.
- Partitions improve scale and locality; replicas improve availability. Excessive partitioning or poorly placed replicas increase synchronization and WAN complexity.
- The first servers in a new tree normally receive root-partition replicas; later placement should be planned around site availability and directory dependencies. [S7]
MASTER IS NOT PRIMARY: Ordinary object writes can occur on writable replicas. The Master is special for partition and replica operations; it is not a single writable directory server in the Active Directory PDC sense.
Bindery emulation Back to contents
NetWare 4.x and later can present selected NDS/eDirectory containers to Bindery-aware clients and applications. The BINDERY CONTEXT SET parameter identifies up to 16 containers, separated by semicolons, whose objects are exposed through Bindery services. [S2]
SET BINDERY CONTEXT = OU=SALES.O=ACME;OU=ACCOUNTING.O=ACME
- The specified containers must be available on the server through local directory replicas or references.
- Bindery-aware applications see a flat view and can encounter duplicate short names across contexts.
- Changing the Bindery context is a compatibility change; test authentication, print, backup, and application dependencies.
Time synchronization and directory health Back to contents
Directory operations depend on coherent timestamps. Official health procedures call for time checks, replica synchronization checks, schema checks, and review of obituaries and directory versions. A dynamic tree should be checked about weekly; a static tree about monthly, and every tree before a major directory operation. [S10]
LOAD DSREPAIR
Time Synchronization
Report Synchronization Status
SET DSTRACE=ON
SET DSTRACE=NODEBUG
SET DSTRACE=+S
SET DSTRACE=*H
Expected healthy indicator: All Processed = Yes
SET DSTRACE=NODEBUG
SET DSTRACE=OFF
DO NOT REPAIR BY REFLEX: DSREPAIR is both a diagnostic and a repair tool. Start with time, version, replica, and synchronization reports. Preserve a supported directory backup and understand the replica ring before initiating destructive or topology-changing repair options.
Directory administration tools Back to contents
Directory tool map
| Tool | Best use | Caution |
| NWAdmin | Classic Windows NDS object administration | Snap-ins and behavior are version specific. |
| ConsoleOne | Cross-platform objects, schema, rights, and product snap-ins | Java/runtime dependencies can be fragile. |
| iManager | Browser-based role and task administration | Old TLS and plug-ins require isolation and compatible browser/runtime. |
| iMonitor | Directory health, partitions, replicas, agents, traces | Diagnostic visibility can expose sensitive directory data. |
| DSREPAIR | Time, replica, database, and synchronization diagnostics/repair | Use health reports first; repairs can alter directory state. |
| DSTRACE | Live directory process and synchronization trace | Filters can be noisy; capture to file and disable when finished. |
Rights and security model Back to contents
File system trustee rights Back to contents
File and directory trustee rights
| Code | Right | Meaning |
| S | Supervisor | All file-system rights; cannot be blocked by an IRF. |
| R | Read | Open and read files. |
| W | Write | Modify file contents. |
| C | Create | Create files/subdirectories; supports salvage semantics where applicable. |
| E | Erase | Delete files and directories. |
| M | Modify | Rename items and change file or directory attributes. |
| F | File Scan | See and search names in the file-system structure. |
| A | Access Control | Add/remove trustees and change trustee rights and IRFs. |
The workstation RIGHTS utility displays or changes assignments. ALL grants all rights except Supervisor. A plus adds rights, a minus removes rights, and a rights list without plus/minus replaces the assignment. [S2]
RIGHTS SYS:DATA R W C E M F /NAME=.TEAM_ARCHITECTURE.ACME
RIGHTS SYS:DATA /NAME=.PJONES.ARCHITECTURE.ACME /I
RIGHTS SYS:DATA /T
RIGHTS SYS:DATA REM /NAME=.OLDGROUP.ACME
eDirectory object and property rights Back to contents
eDirectory rights
| Class | Right | Meaning |
| Object | Supervisor | All rights to the object and its properties. |
| Object | Browse | See the object; does not reveal its property values. |
| Object | Create | Create objects beneath a container; includes Browse. |
| Object | Delete | Delete the target object. |
| Object | Rename | Change the target object’s name. |
| Property | Supervisor | Complete control over the selected property. |
| Property | Compare | Test a value without reading it. |
| Property | Read | Read property values; includes Compare. |
| Property | Write | Create, change, or delete property values. |
| Property | Add Self | Add/remove the trustee itself in object-valued properties such as group membership. |
SENSITIVE DELEGATION: Broad Read access to all User attributes can expose password-management attributes in some eDirectory configurations. Delegate only the object and property rights required for the task. [S8]
Effective rights and inherited rights filters Back to contents
Effective rights are the rights available at the moment of access after NetWare/eDirectory combines explicit trustee assignments, group membership, security equivalence, inherited assignments, and the applicable IRFs. [S8]
- An IRF removes selected rights as they flow down a directory or file-system hierarchy.
- A lower explicit trustee assignment can add required rights back at the target level.
- File-system Supervisor cannot be filtered by a file-system IRF; assign it sparingly.
- An apparent rights failure can instead be a visibility failure: File Scan is required to see names, and Read is required to open content.
- Check the user’s direct assignment, group assignments, security equivalences, IRFs along the path, the target’s explicit assignment, and file attributes.
Practical rights patterns Back to contents
Common assignment patterns
| Use case | Suggested trustee pattern | Notes |
| Home directory | User: R W C E M F; administrator group: S | Set at each home root or through automated provisioning; protect parent visibility. |
| Shared read-only data | Reader group: R F | Read alone is insufficient for useful browsing; File Scan exposes names. |
| Shared working area | Contributor group: R W C E M F | Exclude Access Control unless users must delegate rights. |
| Drop box | Purpose-built rights and visibility design | Test create, read-back, overwrite, rename, delete, and listing behavior separately. |
| Delegated folder owner | R W C E M F A | Access Control permits trustee/IRF changes; does not confer Supervisor. |
| Service account | Dedicated group with minimum path-specific rights | Avoid security equivalence to Admin or broad container-level Supervisor. |
File and directory attributes Back to contents
Common attributes
| Attribute | Effect / use |
| Read Only | Prevents file modification; older implementations may also imply rename/delete protection. |
| Archive | Marks a file as changed for archive-aware backup workflows. |
| Hidden / System | Controls client visibility and marks operating-system content. |
| Shareable | Permits compatible shared access semantics. |
| Transactional | Enables transaction tracking where supported. |
| Purge Immediate | Deletes without retaining a salvageable copy. |
| Rename Inhibit | Prevents renaming. |
| Delete Inhibit | Prevents deletion. |
| Copy Inhibit | Restricts copying where supported by the client/protocol path. |
Attribute support varies between Traditional and NSS volumes and between access protocols. A change is effective only when the underlying file system and NCP path can enforce it. [S6]
Storage, volumes, and file systems Back to contents
Traditional file system versus NSS Back to contents
File-system comparison
| Characteristic | Traditional file system | NSS |
| Structure | NetWare partition containing one or more volumes | Devices/partitions feed pools; pools contain logical volumes |
| Recovery | VREPAIR on an unmounted volume | VERIFY/REBUILD and NSS-specific tools; VREPAIR is not used |
| Mount behavior | Directory/FAT structures can make large-volume recovery slow | Journaling and modern metadata support faster activation |
| Capacity model | Fixed allocation within the NetWare partition | Volumes allocate pool space dynamically and can be overbooked |
| Namespaces | DOS plus added LONG/MAC/NFS namespaces | Multiple namespaces integrated into NSS semantics |
| Advanced features | Trustees, salvage, compression on supported versions | Trustees, quotas, salvage, compression, snapshots/DFS options, encryption |
| Migration | Requires metadata-aware copy and namespace planning | Designed for compatibility with OES NSS and NCP services |
NSS storage hierarchy Back to contents
- Physical disks, SAN LUNs, RAID devices, virtual disks, or multipath devices are presented as storage devices.
- NSS partitions or segments allocate device space to one or more pools.
- A pool aggregates space and can span devices. Its failure domain therefore includes every device contributing segments.
- One or more NSS volumes allocate space from a pool only as needed. A volume belongs to one pool; a pool can contain multiple volumes.
- NCP, CIFS, AFP, NetStorage, and application services expose the volume through their configured identity and trustee model.
NetWare 6.5 NSS supports dynamic volume growth within a pool and overbooking. The official guide describes volumes up to 8 TB and very large file counts, while individual devices presented to NetWare NSS are limited to 2 TB. Verify the exact build, device carving, and storage vendor limits before expansion. [S5]
POOL FAILURE DOMAIN: Spanning a pool across devices increases capacity but also couples availability. Hardware RAID, NSS mirroring, SAN protection, multipathing, and backup are different controls; document which layer actually protects each segment.
Namespaces and path compatibility Back to contents
A namespace records names and metadata for a client environment. Traditional volumes begin with DOS semantics and can add LONG for Windows/OS2, MAC for Macintosh, and NFS for Unix-style names. [S2]
LOAD LONG.NAM
ADD NAME SPACE LONG TO DATA
REM Equivalent choices exist for MAC and NFS where the modules are installed.
- Do not remove a namespace until every file name and metadata dependency has been assessed.
- Migration tools may require the NFS namespace on a Traditional source volume to preserve names and metadata correctly.
- Case, forbidden characters, alternate data, Macintosh metadata, and long-name collisions must be tested against the destination protocol.
Salvage, purge, quotas, and capacity Back to contents
- Salvage retains deleted files until they are restored, purged, or reclaimed according to policy. It is not a backup because it shares the same volume and failure domain.
- Purge permanently removes salvageable files. PURGE IMMEDIATE bypasses recovery for selected files/directories.
- NSS can enable salvage per volume; NSS console commands include NSS /SALVAGE=volume and NSS /NOSALVAGE=volume. [S2, S5]
- User, directory, and volume quotas control consumption. Always check the pool as well as the logical volume: an overbooked pool can exhaust physical space before volume quotas appear full.
- Keep SYS for operating-system and extension content where practical; place user data and applications on separate pools/volumes. [S5]
Repair boundaries Back to contents
Repair decision table
| Target | Diagnostic / repair family | Boundary |
| Traditional volume | VOLUME/VOLUMES, MONITOR, VREPAIR | Dismount before VREPAIR; preserve logs and backup first. |
| NSS volume/pool | NSS /STATUS, NSSMU, VERIFY/REBUILD, iManager/NRM | Do not use VREPAIR; confirm pool/device state before metadata repair. |
| eDirectory | iMonitor, DSREPAIR, DSTRACE | Directory repair is separate from volume repair; start with health checks. |
| Hardware/storage path | NWPA/driver tools, vendor array, mirror/RAID status | File-system repair cannot correct a failing controller, path, or LUN. |
| Application data | Vendor consistency and recovery tools | A mounted volume does not prove a database or message store is consistent. |
Clients, drive mappings, and login scripts Back to contents
Client families Back to contents
Client generations
| Client | Typical environment | Key components / notes |
| NETX requester | Early DOS / Bindery | Small conventional-memory requester; server-centric login. |
| VLM client | DOS / NDS | LSL + ODI LAN driver + IPXODI + VLM modules; NDS-aware. |
| Client32 | Windows 3.x/9x | 32-bit client stack, NDS login, improved cache and transport support. |
| Novell Client | Windows NT through supported later Windows releases | NCP, eDirectory authentication, SLP, mappings, trustee extensions. |
| Novell Client for Linux | Linux workstations | NCP and eDirectory integration; some login-script commands differ. |
| NetStorage | Browser/WebDAV-style access | Interprets selected MAP and conditional login-script commands. |
| CIFS/AFP/NFS | Native OS access | Protocol-specific identity mapping and metadata semantics; not identical to NCP. |
Path syntax and mappings Back to contents
Path examples
| Form | Example | Meaning |
| Current server | SYS:PUBLIC | PUBLIC directory on SYS of the current/default server. |
| Server qualified | NW65LAB/SYS:PUBLIC | PUBLIC on SYS of server NW65LAB. |
| Object qualified | .DATA.NW65LAB.SERVERS.ACME:PROJECTS | Directory under a Volume object identified from [Root]. |
| Drive mapping | MAP G:=NW65LAB/DATA:PROJECTS | Map G to the explicit server/volume path. |
| Home mapping | MAP H:=%HOME_DIRECTORY | Map from the user’s eDirectory Home Directory property. |
| Search drive | MAP INS S1:=SYS:PUBLIC | Insert a program search mapping without replacing existing search drives. |
| Fake root | MAP ROOT F:=SERVER/VOL:APP | Present APP as the apparent root for a legacy application. |
Login script execution order Back to contents
- The container login script runs first and establishes defaults for users in that O or OU.
- The assigned Profile object’s login script runs next and adds role- or team-specific mappings.
- The User object’s login script runs last and can override earlier mappings.
- If the user has no user login script, the built-in default login script runs unless NO_DEFAULT was issued by a container or profile script. [S9]
- Prefer reusable container and profile scripts; reserve user scripts for true exceptions.
- Use IF MEMBER OF to drive group-based mappings, INCLUDE for shared text scripts, and MAP DISPLAY OFF/ON for clean output.
- The last conflicting MAP wins. Diagnose the full chain rather than only the user script.
Login script example Back to contents
REM Container/Profile example – validate names and client behavior
MAP DISPLAY OFF
MAP ERRORS OFF
MAP INS S1:=NW65LAB/SYS:PUBLIC
MAP H:=%HOME_DIRECTORY
IF MEMBER OF “.TEAM_ARCHITECTURE.ACME” THEN
MAP G:=NW65LAB/DATA:ARCHITECTURE
END
IF MEMBER OF “.NETWARE_ADMINS.ACME” THEN
MAP M:=NW65LAB/SYS:SYSTEM
END
MAP ERRORS ON
MAP DISPLAY ON
MAP
TEST MATRIX: Test login scripts with each supported client family, transport, context, roaming site, and group combination. A script that works in the Windows Novell Client may be only partially implemented by the Linux client or NetStorage. [S9]
Administration quick reference Back to contents
Primary administration tools Back to contents
Administration tool map
| Tool | Runs at | Primary job |
| System Console | Server | Core commands, screen switching, NLM control, boot and emergency operation. |
| MONITOR | Server console | Connections, CPU, memory, service processes, storage, LAN statistics, SET parameters. |
| NWCONFIG | Server console | Drivers, products, NCF editing, installation options, traditional volume tasks. |
| INETCFG / TCPCON | Server console | Protocol configuration and live TCP/IP status/statistics. |
| NSSMU | Server console | NSS devices, partitions, pools, volumes, RAID, and attributes. |
| Novell Remote Manager | Web browser | Health, console, modules, connections, volumes, parameters, logs, diagnostics. [S11] |
| iManager | Web browser | Role-based eDirectory, NSS, files, rights, certificates, and service administration. |
| ConsoleOne / NWAdmin | Workstation or server GUI | Directory objects, schema, rights, login scripts, and product snap-ins. |
Console command quick reference Back to contents
Common server-console commands
| Command | Purpose | Notes |
| HELP [command] / HELP ALL | Show console command help | Prefer local help because loaded modules register additional commands. |
| VERSION | Show NetWare, support pack, license, and eDirectory versions | Record before any change. |
| CONFIG | Show server, LAN, IPX, tree, and Bindery context information | Useful hardware/network baseline. |
| TIME | Show server time | Compare with directory time sources. |
| MODULES [prefix*] | List loaded modules and address spaces | Use before unload or abend analysis. |
| SEARCH | Show or modify NLM search paths | Unexpected paths can load the wrong module version. |
| MEMORY | Show installed/addressable memory | Use NRM for deeper attribution. |
| DISPLAY PROCESSORS | Show processor online/offline state | NetWare 5/6 multiprocessor environments. |
| DISPLAY ENVIRONMENT | Show search paths and SET parameters | DISPLAY MODIFIED ENVIRONMENT shows deviations only. |
| SET | Browse or change server parameters | Confirm valid range and persistence file. |
| MONITOR | Open live system monitor | Connections, resources, parameters, storage, LAN. |
| LOAD / UNLOAD | Link or unlink an NLM or driver | Respect dependencies and application shutdown procedure. |
| PROTECT file.NCF | Load an NCF into a protected address space | Only for compatible modules. |
| PROTECTION | List/configure protected spaces | Can enable restart behavior. |
| MOUNT volume / MOUNT ALL | Mount volumes | Use NSS tools for NSS-specific activation issues. |
| DISMOUNT volume | Make a volume unavailable | Close files and stop dependent applications first. |
| VOLUME / VOLUMES | List mounted volumes | Spelling varies by release/module registration. |
| NSS /STATUS / NSS /HELP | Show NSS state and help | NSS commands are version specific. |
| NSSMU | Open NSS management utility | Destructive functions can erase device metadata. |
| NWCONFIG | Open server configuration | Driver/product/NCF and traditional storage tasks. |
| INETCFG / TCPCON | Configure or monitor networking | Save before restart; distinguish configuration from live state. |
| PING / TPING | Test IP reachability | TPING syntax and implementation vary. |
| DISPLAY SERVERS | List SAP-advertised IPX services | Not an IP/SLP discovery test. |
| DISPLAY SLP … | Show SLP agents, services, addresses, or types | Exact subcommands depend on SLP.NLM build. |
| DSREPAIR | Directory diagnostics and repair | Start with time and sync reports. |
| SET DSTRACE=… | Control directory trace | Disable filters/logging when complete. |
| CONLOG | Capture console messages | Default SYS:ETC\CONSOLE.LOG; load early. |
| DISABLE LOGIN / ENABLE LOGIN | Control new logins | Existing connections remain until cleared or logged out. |
| SECURE CONSOLE | Restrict console operations | Load required nonstandard-path modules first. [S2] |
| DOWN | Orderly shutdown | Flushes caches, closes files, executes SHUTDOWN.NCF. |
| RESTART SERVER [-NA|-NS] | Orderly NetWare restart | Use diagnostic switches deliberately. |
Workstation utility quick reference Back to contents
Common client/workstation utilities
| Utility | Purpose | Example |
| LOGIN | Authenticate and execute login scripts | LOGIN TREE/USER or LOGIN SERVER/USER |
| LOGOUT | Close authenticated connections | LOGOUT or client GUI equivalent |
| MAP | View/create/delete drive and search mappings | MAP G:=SERVER/VOL:PATH |
| CX | View/change eDirectory context | CX /T /A |
| RIGHTS | View/change file trustees, rights, IRF, and sources | RIGHTS path /NAME=user /I |
| FLAG | View/change file or directory attributes | Syntax varies; prefer client property page for safety |
| SALVAGE / PURGE | Restore or permanently remove deleted files | Use client GUI or matching release utility |
| CAPTURE / NPRINT | Legacy queue-based print redirection/submit | Version and client dependent |
Core NLM quick reference Back to contents
Common modules – not exhaustive
| Module | Role | Operational warning |
| DS.NLM | NDS/eDirectory engine | Directory-dependent services and authentication rely on it. |
| NCP.NLM / CONNMGR.NLM | Core NCP and connection services | Foundation for client file service. |
| NCPIP.NLM | NCP over TCP/UDP | Unloading removes IP NCP access. [S3] |
| NCPIPX.NLM | NCP over IPX | Legacy transport; not intended for casual unload after activation. |
| TCPIP.NLM | TCP/IP stack | Large dependency tree; use INETCFG/TCPCON. |
| IPXSPX.NLM | IPX/SPX stack | Required by legacy IPX clients/services. |
| SLP.NLM / SLPDA.NLM | IP service discovery / Directory Agent | Scopes and directory replicas affect availability. |
| NSS.NLM | Novell Storage Services | Do not unload with active NSS volumes or dependent services. |
| NWPA.NLM | Storage driver architecture | HAM/CDM storage access depends on it. [S3] |
| MONITOR.NLM | System monitoring | Can be loaded and unloaded as a utility. |
| NWCONFIG.NLM | Server configuration | Installation and driver operations can alter NCF files. |
| DSREPAIR.NLM | Directory diagnostics/repair | Repairs can change replicated state. |
| CONLOG.NLM | Console logging | Configure rotation; unlimited logs can consume SYS. |
| PORTAL.NLM / HTTPSTK.NLM | Remote Manager and HTTP stack | Legacy web/TLS exposure requires containment. |
| TIMESYNC.NLM / XNTPD.NLM | Time synchronization | Use one planned time model; eDirectory depends on stable time. |
| SMDR.NLM / TSAFS.NLM / SBCON.NLM | Storage Management Services backup | Coordinate application and directory-aware backup. |
| SNMP.NLM | Monitoring agent | Legacy community-based SNMP is not suitable across untrusted networks. |
Illustrative NCF skeletons Back to contents
NOT PASTE-READY: Driver names, load order, bindings, addresses, and SET parameters must come from the target server’s known-good configuration and matching manuals. The skeletons show separation of concerns only.
REM C:\NWSERVER\STARTUP.NCF – schematic only
REM Pre-mount SET parameters validated for this exact release
SET <pre-mount parameter> = <validated value>
REM Platform, storage adapter, and device modules from known-good media
LOAD <platform>.PSM
LOAD <adapter>.HAM <validated parameters>
LOAD <device>.CDM
REM SYS:SYSTEM\AUTOEXEC.NCF – schematic only
FILE SERVER NAME <SERVERNAME>
LOAD CONLOG ARCHIVE=YES MAXIMUM=<validated-kilobytes>
REM Load/bind LAN and protocols or invoke generated network configuration
<known-good network configuration>
REM Start directory, discovery, storage, management, and applications
<service-specific NCF files>
MOUNT ALL
REM SYS:SYSTEM\SHUTDOWN.NCF – schematic only
REM Stop application services in reverse dependency order
<application stop commands>
REM Flush/close product-specific engines before DOWN completes
Operations runbook Back to contents
Daily, weekly, and monthly checks Back to contents
Operational cadence
| Cadence | Checks | Evidence to retain |
| Daily | Server up time; health summary; ABEND.LOG; SYS/pool free space; mirror/RAID/path state; backup completion; time state; critical service availability | Alert record, console/health snapshot, backup result |
| Weekly | Directory sync for dynamic trees; CONSOLE.LOG and SYS$LOG.ERR review; NLM/application errors; packet buffers; LAN errors; salvage growth; sample restore | DSTRACE/DSREPAIR report, capacity trend, restore evidence |
| Monthly | Directory health for static trees; replica and partition inventory; DS versions; schema/obituary status; account review; recovery media and cold-image verification | Signed health report and configuration archive |
| Before major change | Full directory health check; application-consistent backup; boot/config export; driver/media check; rollback rehearsal; maintenance communications | Change record, hash/manifest, rollback decision point |
Controlled maintenance shutdown Back to contents
- Confirm a current usable backup and record the current VERSION, up time, active modules, volume/pool status, mirror/RAID state, time status, and directory health.
- Notify users and application owners; quiesce or stop databases, message stores, print services, and backup jobs through their supported procedures.
- Issue DISABLE LOGIN. Review MONITOR connections and open files; have users close data and log out rather than clearing active sessions blindly.
- Dismount only the volumes required by the maintenance procedure. Confirm cluster or shared-storage ownership where applicable.
- Issue DOWN for an orderly shutdown. DOWN flushes cache, closes files, updates file-system structures, and runs SHUTDOWN.NCF if present. [S2]
- Wait for the completion message or return to DOS before powering off or rebooting hardware.
- After startup, validate volumes, directory synchronization, SLP/SAP discovery, applications, clients, logging, and backups before re-enabling normal access.
Change preparation checklist Back to contents
- Exact server, support pack, eDirectory, NLM, driver, and hardware/virtual hardware versions recorded.
- STARTUP.NCF, AUTOEXEC.NCF, SHUTDOWN.NCF, SYS:ETC configuration, driver set, and application NCF files copied and hashed.
- Storage map records devices, partitions, pools, volumes, namespaces, quotas, trustee metadata, cluster resources, and free space.
- Directory map records tree, partitions, replicas, Master roles, time sources, SLP scopes/DAs, CA host, licensing, and schema extensions.
- Application-consistent backup and a directory-aware backup completed; representative restore tested.
- Rollback is time-bounded, resourced, and tested; the point beyond which rollback is unsafe is explicit.
- Management access remains available if clients, SLP, DNS, or the normal AUTOEXEC.NCF path fails.
Troubleshooting playbooks Back to contents
Server will not start or SYS will not mount Back to contents
- Capture the screen and preserve BOOT$LOG.ERR. Classify the failure as before STARTUP.NCF, during driver load, during storage discovery, during SYS mount, or after AUTOEXEC.NCF begins.
- Boot with SERVER -NA when SYS can mount but post-mount services fail. Use SERVER -NS or a known-good alternate startup file only when prepared to load required storage support manually.
- Compare STARTUP.NCF, platform support, HAM/CDM drivers, firmware, virtual hardware, and device presentation with the known-good baseline.
- Confirm the controller/LUN/device is present and stable before attempting file-system repair. A missing or changing device is not a metadata-repair problem.
- For a Traditional volume, use VREPAIR only while unmounted and after preserving evidence/backup. For NSS, use NSS status, NSSMU, and the appropriate VERIFY/REBUILD procedure; never VREPAIR an NSS volume.
- Once SYS mounts, start AUTOEXEC.NCF services in controlled groups to isolate the failing module or binding.
Clients cannot find a server Back to contents
Discovery fault isolation
| Check | IP/SLP environment | IPX/SAP environment |
| Basic reachability | PING/TPING, routing, VLAN/firewall, DNS | Frame type, external network number, router path |
| Service discovery | SLP scope, DA list, SYS:ETC\SLP.CFG, port 427 | DISPLAY SERVERS, SAP filters, hop count |
| File service | NCPIP.NLM, TCP/UDP 524, server object address | NCPIPX.NLM, IPX binding/socket |
| Client settings | Preferred tree/server, SLP DA/scope, protocol order | Preferred server, frame type, network number |
| Directory dependency | Replica reachability, NDAP service registrations, time | Directory SAP service and route |
Authentication or login script failure Back to contents
- Separate authentication failure from post-authentication login-script failure. Test a minimal login without application mappings where possible.
- Confirm user distinguished name, context, preferred tree/server, password status, account restrictions, and client date/time.
- Check server time and DSREPAIR Time Synchronization; then verify the required partition replica is reachable and synchronized.
- For legacy clients/applications, verify BINDERY CONTEXT and unique short names.
- Trace the container, profile, user, and default login scripts in order. Enable MAP errors and remove conditionals temporarily in a test account, not in production for every user.
- Verify that the user is a trustee of the Profile object and that mapped servers/volumes are reachable through the selected protocol. [S9]
Access denied or files are invisible Back to contents
- Confirm the exact path, server, volume, namespace, protocol, and user identity. Alias and Directory Map objects can conceal the real target.
- Use RIGHTS with /NAME and /I, or the client’s Current Effective Rights view, to identify direct, group, security-equivalent, and inherited rights.
- Inspect IRFs at every level from the relevant parent to the target. Confirm that Read and File Scan exist for visibility and content access.
- Inspect target file/directory attributes: Read Only, Hidden, Delete Inhibit, Rename Inhibit, Purge Immediate, and protocol-specific enforcement.
- Check user/directory/volume quotas, free pool space, open-file and record locks, ownership, and application-level permissions.
- Remember that rights on the eDirectory Volume object do not substitute for file-system trustee rights stored on the volume.
Slow response or high utilization Back to contents
- Establish whether the delay is client-only, service-specific, server-wide, site-specific, or time-of-day dependent.
- Use Novell Remote Manager or MONITOR for CPU, service processes, packet receive buffers, memory, connections, open files, LAN errors, and disk activity.
- Check mirror/RAID state, controller errors, LUN latency, low SYS/pool space, salvage backlog, and concurrent backup/antivirus/application jobs.
- Review SLP timeouts and DA availability for slow logins; review directory replica placement and time for slow authentication.
- Record DISPLAY MODIFIED ENVIRONMENT. NetWare defaults were tuned as a balanced system; do not copy old tuning folklore without evidence. [S4]
- If No ECB Available Count grows, investigate dropped packets, driver/TSM compatibility, and packet receive buffers; more buffers consume memory. [S4]
Volume or pool is full Back to contents
- Identify whether the constraint is a user quota, directory quota, logical volume quota, physical NSS pool, Traditional partition, SYS, or underlying storage device.
- Stop the process generating data before deleting evidence or expanding storage.
- Review salvageable files and purge only under an approved retention decision. Salvage is shared-capacity recovery, not free space.
- For NSS, check every volume in the pool and account for overbooking. Extend the pool only after validating device size, RAID/path protection, backups, and vendor limits.
- For SYS, remove or rotate logs and temporary/support-pack content only when ownership is known. Do not delete hidden directory or product files by pattern.
- After remediation, restore alert thresholds, logging rotation, quota controls, and capacity trend monitoring.
Abend or repeated restart Back to contents
- Preserve the abend screen, ABEND.LOG, CONSOLE.LOG, core dump if configured, MODULES list, application logs, and the exact preceding change/workload.
- Prevent an uncontrolled restart loop. Automatic restart can hide recurring abends; check ABEND.LOG and server up time routinely. [S2, S4]
- Identify the faulting NLM, address space, thread, and dependency chain. Determine whether it ran in the kernel or a protected space.
- Reproduce only in an isolated clone with matching data and versions. Do not swap NLMs across support packs merely because file names match.
- If a protected application space faults, review restart/no-restart policy and product recovery semantics before reloading it.
- Treat resulting file-system or application inconsistency separately; a recovered kernel does not prove data consistency.
eDirectory synchronization errors Back to contents
- Run the directory health sequence: versions, time synchronization, replica synchronization, schema synchronization, obituaries, and external references.
- Verify IP/IPX reachability, NCP/NDAP service addresses, SLP/SAP discovery, DNS, firewall rules, and the replica ring’s server objects.
- Use DSREPAIR Report Synchronization Status and DSTRACE filters to collect the error and affected partition; seek All Processed = Yes for healthy rings. [S10]
- Resolve time, connectivity, name/address, disk-space, and version defects before repairing the directory database.
- Back up directory state and record replica roles before partition, replica, or obituary repairs. Coordinate changes across every server in the ring.
Backup and disaster recovery Back to contents
What a usable backup must preserve Back to contents
Recovery asset inventory
| Asset | Preserve | Why |
| Boot environment | DOS partition/image, SERVER.EXE, STARTUP.NCF, AUTOEXEC.BAT, drivers | Needed before SYS and network services are available. |
| SYS configuration | AUTOEXEC.NCF, SHUTDOWN.NCF, SYS:ETC, NLM/application config, logs | Reconstructs service identity and load order. |
| Directory | Supported NDS/eDirectory backup, schema, partitions/replicas, certificates | A file copy of the live DIB is not a supported directory backup. |
| File data | Files plus trustees, IRFs, ownership, attributes, namespaces, quotas, links | A generic SMB copy can lose NetWare metadata and security. |
| Applications | Vendor-consistent database/message-store backup and transaction logs | Volume-level consistency does not ensure application consistency. |
| Storage map | Controller/LUN/RAID, devices, partitions, pools, volumes, cluster resources | Required to present the same data in the same ownership model. |
| Software entitlement | Install/overlay media, support packs, patches, drivers, licenses, keys | Downloads and activation services may no longer be obtainable. |
| Operational evidence | Runbooks, credentials escrow, dependencies, test results, hashes | Turns backup media into a repeatable recovery. |
BACKUP SEMANTICS: Use Storage Management Services or another NetWare-aware product for trustee and namespace fidelity, plus an application-aware method for databases and directory services. Test the exact restore path; a successful backup job is not evidence of recoverability.
Recovery rehearsal Back to contents
- Create an isolated recovery network with no route to production and a controlled time/DNS/SLP design.
- Recover the boot environment and virtual/physical hardware drivers, then start with normal NCF files suppressed if necessary.
- Present storage consistently and recover SYS before application/data volumes. Validate TFS/NSS type before any repair action.
- Restore the directory using its supported method and intended replica topology. Avoid creating duplicate server or tree identities on a connected network.
- Restore file data with trustees, IRFs, ownership, attributes, quotas, and namespaces, then application data with vendor consistency checks.
- Test representative authentication, login scripts, rights, mappings, locks, salvage, print, backup, and application transactions.
- Record recovery time, manual decisions, missing assets, and new hashes. Update the runbook and repeat until another engineer can execute it.
Printing and ancillary services Back to contents
Printing generations Back to contents
NetWare printing models
| Model | Core objects/services | Client experience |
| Queue-based printing | Print Queue, Printer, Print Server; PSERVER; CAPTURE/NPRINT | LPT redirection or queue submission; common in 3.x/4.x. |
| NDPS | Broker, Manager, Printer Agent; NDPSM | Directory-discovered printers, driver distribution, status and notification. |
| iPrint | IPP-based print services and web installation | Browser/client printer installation and IP transport; mature in 6.x/OES. |
- Inventory printer agents, gateways, drivers, queues, ports, DNS names, and application dependencies before migration.
- A user can authenticate and map drives successfully while printing fails through an independent Broker/Manager/gateway path.
- Legacy printer drivers are executable code. Preserve them for recovery but do not deploy them to unsupported modern clients without containment and testing.
Other common services Back to contents
Common optional services
| Service | Typical role | Migration/containment note |
| DNS/DHCP | Directory-integrated network services | Export zones, subnets, options, and service-object dependencies. |
| NetStorage | Web access to NCP/CIFS-backed files | Isolate old web/TLS; login-script support is partial. |
| iFolder | User file synchronization | Inventory clients, stores, policies, and conflict behavior. |
| Apache/Tomcat/MySQL/PHP | Web/application platform on 6.5 | Version-specific security/consistency; migrate rather than expose. |
| GroupWise | Messaging and collaboration | Use product-specific domain, post-office, and agent migration. |
| Btrieve/Pervasive | Transactional database engine | Coordinate shutdown, logs, locks, and version compatibility. |
| Cluster Services | Failover for volumes and services | Preserve virtual NCP identity, scripts, preferred nodes, and shared storage. |
| SMS backup | TSA/SMDR/SBCON backup framework | Record agent, media, catalog, encryption, and restore dependencies. |
Containment, preservation, and migration Back to contents
Minimum containment pattern Back to contents
- Place NetWare and any dependent legacy clients on a dedicated VLAN or virtual switch with no direct Internet route.
- Default-deny at the firewall. Permit NCP, SLP, DNS, NTP, backup, directory, print, and application flows only between documented endpoints.
- Use a hardened jump host with the compatible Novell Client and management tools. Reach the jump host through modern MFA and encrypted remote access.
- Send logs and monitoring outward through a controlled relay or poll from a collector; do not install untested modern agents into the NetWare kernel.
- Keep offline, immutable copies of installation media, patches, drivers, configuration, license artifacts, system images, and data backups.
- Set an explicit retirement date and risk owner. Containment reduces exposure; it does not make unsupported code supportable.
Migration sequence Back to contents
Staged migration
| Stage | Activities | Exit criterion |
| Discover | Inventory services, directory roles, applications, volumes, trustees, clients, print, protocols, and dependencies | Authoritative dependency and data map approved |
| Stabilize | Patch to the approved terminal level, fix time/replication/storage errors, test backup and restore | Healthy, repeatable source baseline |
| Design | Select supported OES/Enterprise Server, Windows/Linux, SaaS, or application-specific destinations; map identity and rights | Target architecture and rollback signed off |
| Pilot | Migrate representative users/data/printers/apps with metadata-aware tools | Functional, security, performance, and recovery tests pass |
| Coexist | Introduce target services, NCP/CIFS/client changes, DNS/SLP updates, and staged data synchronization | Users operate on target with measured exceptions |
| Cut over | Quiesce source, final sync, redirect mappings/services, validate rights and applications | Business acceptance and rollback decision closed |
| Retire | Remove applications, replicas, service objects, licenses, routes, and storage in supported order | No hidden dependency; evidence and retention complete |
METADATA-AWARE COPY: Use the supported migration/consolidation tool or an NSS/NCP-aware process when trustee assignments, IRFs, ownership, namespaces, Macintosh metadata, quotas, or application attributes matter. Generic drag-and-drop or SMB copies are not equivalent. [S12, S15]
Virtualization and historical preservation Back to contents
- NetWare 6.5 SP8 documented VMware and Xen guest deployments, but compatibility depends on virtual CPU, storage, network adapter, and driver choices. [S1]
- Preserve the original disk images before converting formats. Work on a verified copy and record hashes before and after transformation.
- Keep the virtual NIC disconnected during the first boot of a clone to prevent duplicate server names, internal network numbers, tree identities, or replica activity.
- Match old virtual hardware where possible. A newer hypervisor’s default controller or NIC may have no NetWare driver.
- Capture console video/screens, configuration, volumes, application behavior, and client workflow as part of preservation, not only a bootable VM.
- If the goal is evidence or data extraction rather than continued service, prefer an offline, read-only recovery workflow over production resurrection.
Appendix A – Common paths and files Back to contents
Locations commonly encountered on NetWare 4.x-6.5
| Location | Contents / use |
| C:\NWSERVER\SERVER.EXE | NetWare server loader/kernel image. |
| C:\NWSERVER\STARTUP.NCF | Pre-SYS SET parameters and platform/storage drivers. |
| C:\NWSERVER\BOOT$LOG.ERR | Boot messages/errors according to logging configuration. |
| C:\ABEND.LOG then SYS:SYSTEM\ABEND.LOG | Abend record before and after restart/copy. |
| SYS:SYSTEM | Core NLMs, AUTOEXEC.NCF, SHUTDOWN.NCF, utilities, application start files. |
| SYS:PUBLIC | Client utilities and management program files. |
| SYS:LOGIN | Files accessible during login and pre-authentication workflows. |
| SYS:ETC | Network/service configuration and logs, including SLP.CFG and CONSOLE.LOG. |
| SYS:ETC\CONSOLE.LOG | Default CONLOG output. |
| SYS:SYSTEM\DSTRACE.DBG | Directory trace output when trace-to-file is enabled. |
| SYS:_NETWARE | Hidden/system directory containing directory database and security data; never treat as ordinary file content. |
| SYS:SYSTEM\SYS$LOG.ERR | Common system error log location on many releases. |
| volume-root\VOL$LOG.ERR | Traditional volume error/repair log commonly found at a volume root. |
VERSION VARIANCE: Paths can be redirected, clustered, or changed by products and support packs. Treat this appendix as a discovery list, then confirm with CONFIG, SEARCH, module parameters, NCF files, and the matching manual.
Appendix B – Glossary Back to contents
Glossary
| Term | Definition |
| Abend | Abnormal end: a NetWare fault or exception that can suspend a thread, fault an address space, or stop/restart the server. |
| Bindery | Per-server flat database of users, groups, properties, and services used primarily by NetWare 2.x/3.x. |
| Bindery context | One or more eDirectory containers exposed as a flat Bindery view for legacy clients/applications. |
| CDM | Custom Device Module in the NetWare Peripheral Architecture storage stack. |
| DIB | Directory Information Base: the local NDS/eDirectory database. |
| Directory Map | eDirectory object that represents a path and reduces hard-coded mapping dependencies. |
| Distinguished name | An object’s unique hierarchical name in the eDirectory tree. |
| eDirectory | Later name and evolution of Novell Directory Services (NDS). |
| HAM | Host Adapter Module: storage adapter driver in NWPA. |
| IRF | Inherited Rights Filter: blocks selected rights inherited through a hierarchy. |
| IPX/SPX | Legacy Novell routed network and connection-oriented transport protocol suite. |
| NCP | NetWare Core Protocol: client/server file and network service protocol. |
| NCF | NetWare Command File: a server-side batch/configuration file. |
| NDS | Novell Directory Services: distributed directory introduced with NetWare 4. |
| NDPS | Novell Distributed Print Services, the directory-based successor to queue printing. |
| Namespace | File-name and metadata representation for DOS, LONG/Windows, Macintosh, NFS, or other clients. |
| NLM | NetWare Loadable Module: executable server component linked into the runtime. |
| NSS | Novell Storage Services: journaling file system and storage-pool/volume architecture. |
| NWPA | NetWare Peripheral Architecture for HAM/CDM-based storage drivers. |
| ODI | Open Data-Link Interface used by classic Novell client LAN/protocol stacks. |
| OES | Open Enterprise Server, the Linux-based successor platform for eDirectory, NCP, NSS, iPrint, and related services. |
| Partition | Contiguous subtree replicated as a unit in NDS/eDirectory. |
| PSM | Platform Support Module for processor/chipset/platform integration. |
| Replica | Copy of a directory partition held by a server; types include Master, Read/Write, Read-Only, and Subordinate Reference. |
| Replica ring | Set of servers holding replicas of the same partition. |
| SAP | Service Advertising Protocol used to advertise services in IPX networks. |
| Salvage | Recovery of files retained after deletion but before purge/reclamation. |
| SLP | Service Location Protocol used for IP service discovery and registration. |
| SMS | Storage Management Services: NetWare backup architecture using agents such as TSA and SMDR. |
| TFS | Traditional NetWare File System, distinct from NSS. |
| Trustee | User, group, or object assigned rights to a target directory, file, or directory object. |
| VLM | Virtual Loadable Module client architecture used by DOS NDS-aware clients. |
| Volume | Named NetWare file-system container such as SYS or DATA, exposed through NCP and represented by an eDirectory object in NDS-era systems. |
Appendix C – Official source set Back to contents
Sources were selected from surviving Novell, Micro Focus, NetIQ, and OpenText documentation. They were accessed on 8 August 2026. Product pages and document hosts can move; retain local archival copies where licensing permits.
[S1] NW 6.5 SP8 Installation Guide. Open official source
[S2] NW 6.5 SP8 Utilities Reference. Open official source
[S3] NW 6.5 SP8 NLM Reference. Open official source
[S4] NW 6.5 SP8 Server Operating System Administration – Troubleshooting. Open official source
[S5] NW 6.5 SP8 NSS File System Administration Guide. Open official source
[S6] NW 6.5 SP8 File Systems Management Guide – attributes and trustees. Open official source
[S7] NW 6.5 SP8 Planning and Implementation Guide. Open official source
[S8] NetIQ eDirectory 8.8 SP8 Administration Guide – eDirectory Rights. Open official source
[S9] Novell Login Scripts Guide. Open official source
[S10] NDS/eDirectory Health Check Procedures – Cross Platform. Open official source
[S11] NW 6.5 SP8 Novell Remote Manager Administration Guide. Open official source
[S12] Open Enterprise Server – Coexistence and Migration of File Services. Open official source
[S13] OpenText Product Support Lifecycle. Open official source
[S14] SLP Design and Implementation Guidelines. Open official source
[S15] Novell Server Consolidation and Migration Toolkit. Open official source
EDITION NOTE: This reference deliberately avoids prescribing hardware-specific driver lines, destructive repair options, or a current migration destination without an environment inventory. Those decisions must be made against the exact server state and the current support/interoperability matrix.