[ H(1) | H(2) | H(3) | Novell FAQ Home Page ]

H.39 Help Desk software for NetWare

HelpStar, (800)-563-4357

[Thanks to Lewis M and Richard Szabo for this info]

DOMINION by InterApps, Inc. Operates within Windows 3.1x, Windows NT and OS/2; compatible with Netware, VINES, LANtastic, LAN Manager, and NetBios-based LANS. A native Windows application that was designed by people in the Help Desk and network management field, it takes full advantage of the Windows environment, including support for full-motion video, CD-ROM and sound. It includes modules for problem, inventory, change and financial management. It's relational database system provides limitless storage and high speed retrieval of text, images, and binary data. It also provides the capabilities to work the SQL-based database engines like Oracle, XDB, SQL-Server, IBM's DB2 and others. Contact The Help Desk Institute, Suite 101 - 1755 Telstar Dr., Colorado Springs, CO 80920-1017, Tel: (719) 531-5138.

H.40 Deleting very large "sparse" files

Sparse files are basically files which have 'holes' in them. They appear to be a lot larger than the data they contain.

To repair sparse files, obtain TSparse & its text file at:

ftp://netlab1.usu.edu/pub/mirror/apps/tsparse.zip

and

ftp://netlab1.usu.edu/pub/mirror/apps/tsparse.txt

[Thx V.K.R., R.J.L., R.C. & S.R.#2]

H.41 Slow mount of 50 GB drive array

>A 50GB disk array is used with a Netware 3.12 server. However, the server takes 20 minutes to mount the disk array during the startup. How can I speed up the mounting?

How about a faster disk I/O channel? NetWare v4.1 also is much more efficient during the boot process, so perhaps an upgrade is in order.

>I know that it does a consistency check on the two sets of directory tables and FATs to verify that the two copies are identical.

Only one directory table. FATs are indeed verified.

>Are there any other [checks], and can those be turned off?

Yes, no. What may be the largest factor in the apparent slowness you describe is the rest of the volume mounting process. First, NetWare builds a copy of the entire FAT of every volume in RAM. Second, it builds an index, called a Hash Table, of the Directory Table and caches the index in RAM. Believe me, you don't want NetWare to *not* do those things!!! Not with as much as 50 GB of data!

These functions are core to the NetWare operating system, and there's no way to turn them off.

[Thanks to Roger Kresge for this info]

H.42 Clearing "Not-Logged-In" license robbers

This isn't a problem on NetWare 4.x, as NOT-LOGGED-IN's don't count against your user count. For NetWare versions prior to 4.x, try:

ftp://ftp.iastate.edu/pub/pc/novell/utilities/nliclear.zip

[Thx B.S. and S.M.D. for this info]

Also, SHUTDOWN will forcefully disconnect users, with an optional time delay and accompanying warning messages.

	ftp://ftp.cdarchive.com/pub/WIN_UTIL/16A/SHUTDOWN.ZIP
or
	ftp://ftp.cdrom.com/.4/os2/sysutil/shutdown.zip
or
	ftp://ftp.demon.co.uk/.d1/mirrors/os2/sysutil/shutdown.zip

[Thx S.R.#2]

ZZClear is a freeware NLI clearer. If the number of NLI's is less than 2 [i.e. all N-2 connections are logged in], it will force the NLI clear. Has options for not clearing the boss's connection or other station requiring protection from clearing.

	ftp://ftp.uni-marburg.de/pub/netware/nlms/zzclear.zip
and
	ftp://ftp.uni-marburg.de/pub/netware/nlms/zzclear.txt

[Thx J.P. & S.R.#2]

H.43 Running alternatives to MONITOR on the server console

Novell wrote SS.NLM and CPU-UTIL.NLM and put them in the Public Domain "back at the dawn of time". When loaded (one autoloads the other) they create a histogram of CPU utilization which makes a great screen saver for monochrome server consoles. You can find them on:

ftp://ftp.icfrnet.unp.ac.za/pub/novell/

[Thanks to Ron Neely & S.R.#2 for this info]

H.44 NLMs allowing access to NetWare Server C drive

The NetWare NLM SDK contains NWSHELL.NLM that allows DOS type access to any partition on the server, including the DOS partition. It asks you to log in before it lets you on the Netware partitions and will also let you log in to other servers. It is located on the CD as: novsdkcd_4:/nlm/novm/nwshell

[Thx J.B.]

Also, there is MountDOS.NLM for NetWare 4.1 that mounts DOS partitions as NetWare volumes which can then be mapped from any workstation. MountDOS eases updating and backing up the DOS partition software. It supports DOS and NetWare commands including: dir, md, cd, NCOPY, NDIR, VOLINFO, FILER; utilizes the Netware cache, supports access to multiple DOS partitions, provides file locking, has tight security restrictions, supports NETX, VLM & Client32 access, and NDS Users/Groups. Midnight Technologies, CIS: 76451,1527, 76451.1572@Compuserve.COM

[Thanks to Alain Bessette for this info]

Avoid NCL.NLM, it appears to crash servers when unloaded.

Try JCMD, it simulates many DOS commands at the console prompt. It can be found at:

	ftp://netlab2.usu.edu/apps/jcmd_135.zip
For Netscapers:
	ftp://netlab2.usu.edu/sys/anonftp/apps/jcmd_135.zip

[Thx Joe D.]

The Novell Consulting Toolkit has such an NLM called Network Command Center located at:

http://members.aol.com/dssol/home.html

"I have yet to use it, but heard good things about it at a Netware Conference." - D. R.B.

[Thanks Dante RandBenson & S.R.#2 for this info]

Filer.nlm is a file manager designed to allow network administrators the flexibility to access Netware and DOS file systems from a Netware 3.x or 4.x file server console. Freeware:

http://ourworld.compuserve.com/homepages/cdubois/filer.htm

H.45 NLM to delay loading of AUTOEXEC.NCF NLMs

DELAY.NLM is the program. Just 4 KB, it can be downloaded at:

	ftp://netlab2.usu.edu/netwire/novuser/07/delay.zip
For Netscapers:
	ftp://netlab2.usu.edu/sys/anonftp/netwire/novuser/07/delay.zip

	[Thanks to Jiri Polach for this info]
or
	ftp://ftp.novell.co.uk/pub/netwire/novuser/07/delay.zip
or
	ftp://ftp.pht.com/.5/gamehead/oldies/delay.zip
or
	ftp://ftp.data-io.com/dataio/betaalgs/delay.zip
or
	ftp://ftp.cis.ohio-state.edu/pub/jain/papers/delay.zip

	[Thx S.R.#2]

H.46 Checksum errors showing up on MONITOR.NLM screen

This indicates Bad Packets.

The first place to look at is your wiring plant. Not all twisted pair wiring is created equal, and less stays that way over time (no squeezing it, no tight bundles, need to recrimp the crummy RJ45's from time to time). If the wiring plant was incorrectly installed (above, plus wrong pairing, plus wrong punch down blocks, excessive length, or even not even Cat 3 wire) then all kinds of troubles ensue. Twisted pair lan wiring is NOT Telco junk and can't be treated that way; and it is more susceptible to maltreatment than coax.

Before jumping too far, also check the hubs (replace with a spare) since flakey hardware occurs too. To check the wiring plant either you really know what you are doing so visual inspection tells a lot, or you rent/acquire a high calibre cable tester (hand held box) with perhaps an operator attached (see below for story).

The checksum is at the very end of each packet and that's the place least likely to be damaged in transit unless the system is well out of spec.

Packets do get damaged, no matter how good the wire/fiber and electronics are. Damage can be subtle, enough so to pass the hardware CRC check since software is also involved. Thus we should be running IPX with checksums turned ON. We can't use IPX checksums with Ethernet_802.3, so those are losers. UDP/IP packets are similar, and Sun runs NFS without UPD checksums (but don't you make that mistake). Lest folks think I'm just talking, have a look at an serious IP machine and see the errors revealed by netstat.

I provide my student lab consultants with a show and tell on lan wiring. That lab uses coax (better than twisted pair, cheaper too). I demo ordinary packets with a good Tektronix sampling scope, plus variable attenuators, coax stub extensions, inserting another kind of coax, etc. Twisted pair is much more difficult to demonstrate, but near end cross talk (NEXT), a major spec/issue of lan cable, is easy to show by watching the receive pair as the board transmits.

More background info: Twisting wire isolates a pair from the world better than not twisting the wires, yet radiation still leaks away. The impedance of the wire changes as the twisting changes, and the coupling to the world (other wires) changes too. Thus squeezing wires in a bundle increases NEXT, a very bad thing. Unwrapping wires to fit RJ45 plugs and punch down blocks does the same, so great care is needed to do this just right (not to mention using the proper punch down blocks).

Transmitted data goes on one pair, received data on another. If we scramble the wires so twisted pairs are not Tx or Rx then major trouble from NEXT (cross talk) ensues.

As we go from 10Mb/s Ethernet to 100Mb/s (and you will go too) these problems scale up by a factor of 10. Cat 3 cable is just adequate, Cat 5 is better, and the difference is in the way wires are bundled in the jacket. So how come there is such a cost difference? Because they can get away with it, that's why, and some [net-god] system managers insist upon validation of NEXT, impedance and attenuation when the wiring is installed.

[Thx Joe D.]

Note:	100Base-T and 100VG-AnyLAN can run on: Cat 3, 4, and 5 UTP;
		Type 1 STP; and Fiber.
	100Base-T maximum hub-to-node distance is: 100 meters (UTP & STP),
		and 2000 meters (Fiber).
	100VG-AnyLAN max. hub-to-node distance is: 100 meters (Cat 3 UTP),
		200 meters (Cat 5 UTP), and 2000 meters (Fiber).

	[Thx S.R.#2]

H.47 Using EMM386 with Boot ROM (ie. diskless) workstations

You need to use the undocumented /y=[d:][path]EMM386.EXE switch to point to where a copy of EMM386.EXE can be found _after_ the boot image has been discarded.

[Thanks to David Brisbane and George Spencer for this info]

H.48 Fixing NetWare partition problems when Mac support NLMs are used

Reboot the server. Run SERVER.EXE with -na -ns, load your hard disk drivers, load VRepair, then switch back to the console (using Alt-ESC) and load V_MAC.NLM as it is needed for proper repair of Mac namespace (fork) stuff.

[Thanks to Joerg Trawinski for this info]

H.49 BSOD (Black Screen of Death) fix

If you are finding that machines running Windows and connected to NetWare are halting with just a black screen (of death) then:

Replace the device=*vtd entry in the [386 Enh] section of SYSTEM.INI with device=vtda.386 (vtda.386 is available as WW0863.EXE from ftp.microsoft.com)

Add a TimerCriticalSection=1000 or =10000 entry to the [386 Enh] section of SYSTEM.INI, as per the March 95 issue of Novell App Notes:

"If you are running a version of DOS Requestor earlier than version 1.20, or if you are running a version of VNETWARE.386 dated earlier than 8-17-94, you will need to add to the [386 Enh] section of SYSTEM.INI a "TimerCriticalSection=10000 line. The VNETWARE.386 driver for the DOS Requestor 1.20 defaults to 1000, which better supports mobile products."

[Thanks to Wayne Fee and F.H.S. for this info]

Consider downloading NOV-BSOD.TXT (Section S.4) for more complete information.

H.50 Running a World Wide Web server on NetWare

Note: Web server software is included with IntranetWare at no extra cost compared with just NetWare 4.11.

http://www.glaci.com/info/glaci-httpd.html

[Thanks to Walter Blocher for this info]

http://www.american.com

American Internet Corporation, 4 Preston Court, Bedford MA 01730, Voice: 617-271-9200, Fax: 617-275-4930, info@american.com or sales@american.com.

[Thanks to Michael Haag for this info]

http://www.cpu.lublin.pl/yawn/

or

Email: info@cpu.lublin.pl

[Thanks to Artur Urbanowicz for this info]

H.51 Testing the strength/security of your users' passwords

CRACK works on copies of the bindery and uses a dictionary to attempt password cracks. It is available at:

ftp://ftp.mersinet.co.uk/pub/novell/freeware/crack.zip

The extra (huge) dictionary is also there as:

ftp://ftp.mersinet.co.uk/pub/novell/freeware/huge.zip

[Thx S.R.#2]

SMARTPASS is another password checker for NetWare and uses a database of 130,000 common words. It is available at:

ftp://ftp.uni-weimar.de/pub/network/security/smrtpass.html

[Thx E.S. & S.R.#2]

H.52 Faster (but riskier) re-mirroring after server crash

When the server abends and needs to remirror the drives, you can speed up the remirroring process by a factor of 3 or 4 if you disable "read after write verify." Many will argue this practice, but when the server crashes most users are very irate when they can't log into the server due to a remirror process...Of course, when the remirror process finishes, be sure to set "read after write verify" back on! You can change this via MONITOR.

[Thx J.N.]

H.53 Sharing a modem across a NetWare network

Novell Connect v2.0, there are 2 port and 8 port versions.

WINPort from LanSource. Contact roger@lansource.com.

[Thanks to Michael D. Perry for this info]

Spartacom make a nice product, the name is SAPS. Check out:

www.spartacom.nl

[Thanks to Frans Meyer for this info]

H.54.1 Using the Debugger to determine the cause of the ABEND

  a.	Write down the Abend message.
  b.	Drop into the debugger using the 
	key sequence.
  c.	Display the running process using the .R debugger command and note
	the running process information.
  d.	Display your location using the ? command and note the NLM and
	function information (if available).
  e.	Display the control registers using the RC command and note the
	contents of CR2.
  f.	Quarantine the running thread by setting EIP to
	CSleepUntilInterrupt (case sensitive).
  g.	Try to restart the server using the G command.
  h.	If you're successful, retry your client connections, close your
	applications, and allow all pending I/Os to flush to disk. If
	you're unable to restart the server at this point, use the
	information you've gathered to ascertain which module may be
	generating the fault, reboot and proceed to step j.
  i.	Down the server using the DOWN console command.
  j.	Begin troubleshooting by removing the buggy software module, or
	load the module inside a protected domain using domain.nlm.

[Thx D.B.]

H.54.2 Help for Debugger commands

At the debugger prompt type .h to get help on all debugger commands.

# .h
.A		Show the abend or break reason
.C		Do a diagnostic core dump to disk
.D		Show page directory map for current debugger
		domain
.D addr		Show page entry map for current debugger domain
.F		Toggle ON/OFF the developer option flag
.G		Show the GDT
.H		Show this dot help screen
.I		Show the IDT
.I2		Show the IDT for Processor 2
.M		Show loaded module names and adresses
.L off 		Show linear address given page map offsets
.LA laddr [cr3]	Find all aliases of linear-address
.LP paddr [cr3]	Find all linear mappings ofphysical-address
.LX addr	Show page offsets & values used for translations
.P		Show all process anmes and adresses
.P[L] addr	Show address as a process control block
.R		Show the running process control block
.S		Show all screen names and adresses
.S addr		Show address as a screen structure
.T addr		Show address as a TSS structure
.TSsegnum	Show GDT[segnum] as a TSS structure
.V		Show server version

Where: addr is the address off is the offset laddr is the linear address paddr is the physical address [Thx D.B.]

H.55 Network management utilities

Managewise is the market leader - there's a huge advantage in not needing $20,000 workstations just to run the console. From what I've seen it's fairly comparable to HP Openview. IBM also has an entry, but the name escapes me.

[Thx R.G.]

DOMINION by InterApps, Inc. Operates within Windows 3.1x, Windows NT and OS/2; compatible with Netware, VINES, LANtastic, LAN Manager, and NetBios- based LANS. A native Windows application that was designed by people in the Help Desk and network management field, it takes full advantage of the Windows environment, including support for full-motion video, CD-ROM and sound. It includes modules for problem, inventory, change and financial management. It's relational database system provides limitless storage and high speed retrieval of text, images, and binary data. It also provides the capabilities to work the SQL-based database engines like Oracle, XDB, SQL-Server, IBM's DB2 and others. Contact The Help Desk Institute, Suite 101 - 1755 Telstar Dr, Colorado Springs, CO 80920-1017, Tel: (719) 531-5138.

H.56 Only see 2 GB of disk space in DOS/Win when U know HD/volume is bigger

DOS uses a 32-bit signed integer to keep track of free disk space. The largest quantity which a 32-bit signed integer can represent is 2 GB less one byte. Therefore, that's the most free disk space that you'll see the DOS DIR command report. That doesn't mean that the space isn't available; use CHKVOL (3.x and below) or NDIR /VOL (4.x) to confirm that there actually is more disk space available to you than DOS indicates. With DIR, you will see your free disk space appear to stay the same, at slightly under 2 GB, until such time as your actual free disk space drops below 2 GB, and then the result of DIR will be correct.

[Thx S.M.D.]

H.57 Patching NetWare to handle the Year 2000 and beyond

See:

http://www.novell.com/p2000

[Thx John Canfield]

[ H(1) | H(2) | H(3) | H(4) | Novell FAQ Home Page ]