We help IT Professionals succeed at work.

Slow Net Use mappings of NT drives

Medium Priority
379 Views
Last Modified: 2013-11-29
We have combination NT 4.0 and Netware 3.11/3.12 network.  Had it for many years, and nothing has recently been added or removed such as new servers, client configuration changes, etc.  We recently are having periodic problems running the login.bat file on the NT servers.  We perform drive maps via the NET USE command within login.bat.  They are taking 1-2 minutes per NET USE to complete the mapping. The problem comes and goes, but seems to be related to the Netware servers across the WAN.  There are 3 local netware servers and 2 over 384k frame relay links.  If we disconnect the WAN, the NET USE map is quicker but still not normal (1-2 seconds I consider normal).  Once a person is logged in, all software applications, file copies, etc are normal, its only the NET USE mapping.

Does anyone know how the NET USE actually functions or what its doing that might cause it to take so long?  It never actually times out.  Eventually the mapping takes place.  It can take someone with 5-6 maps up to 10 minutes to log in. Clients are using the Client for Microsoft Networks as the primary client.  

I'm really interested in exactly how the NET USE works in regards to protocol, process, etc.

Thanks
Comment
Watch Question

Commented:
Have you tried doing the net use manually?  Does it still take as long from a DOS prompt?  If that's the same, have you tried using the computer's IP address instead of name, such as:

net use t: \\10.1.1.1\c$

Also, I am assuming the problem is to any server and not just one particular server?

As far as how it works, there is not much special to it.  It does a name lookup (if you aren't using IP) and then sets up an SMB session with the server and maps the drive.  Like I said not much to it.

By my questions, I am trying to narrow it down to 1 of a few possible issues - the server, the login process, or the name lookup.

CERTIFIED EXPERT
Top Expert 2007

Commented:
Could t be your 384Kbps connection is acting up somehow ??

Have you tested it lately ??

Any router/rquipment changes ?


I hope this helps !
Top Expert 2004

Commented:
also make sure the binding orders are the same on the workstations as on the server. you want the protocol you wish to be used listed first (at the top)
CERTIFIED EXPERT

Commented:
which is you current master browser? is it probaly also on the other side of the 384k line?
(use browstat.exe from ResorceKit)
CERTIFIED EXPERT
Distinguished Expert 2019

Commented:
Which client for NetWare do you have? The Novell one tries to open NDS on M$ servers and M$'s one tries to open the registry on Novell servers. Both cause long timeouts depending on which version.

Author

Commented:
scraig84:
1)  Manual net use takes just as long.
2) Using the IP address instead of machine name does not help.
3)  Yes, it affects all servers, not just one or two.
StevenLewis:
1)  Yes the binding orders are the same.
SysExpert:
1)  We thought about the 384K connection, but all other WAN activity (that we can tell) is normal.
2) No router equipment changes.  The WAN router was power cycled prior to this problem happening as we needed to shut down power in the wiring room last week beyond the expected UPS time.  The timing is not quite right with when the problem surfaced however.  The equipment had been powered on for 6-7 hours before we had any problems.  I have thought about whether its related.   Again, no other issues other than slow NET USE mappings.
AndyAlder:
1)  We use the M$ client.
AHoffman:
1) Master browser in on the LAN side.

Another items of note:  
1) None of our Windows 2000 clients are affected.  Only 95/98SE.
2)  Any possiblility this could be a SAP problem?
3)  We have noticed today that shutting the WAN router off and back on seems to resolve the problem temporarily.  This leads me to either the router, or possibly a Netware server in one of the remote offices.  I would think that if it was a server than recycling the router would not work, but I don't really know.

Commented:
Do you have access to any type of sniffer such as Sniffer Pro or even Ethereal (free sniffer available at www.ethereal.com).  I would be interested to see the output from both the client and server ends of the connection.  

You mention SAP being a possibility which confuses me a bit.  We are talking about mapping drives to an NT server correct?  Also, are we using IP or IPX?  SAP is a protocol specific to IPX and NetWare service advertisement.  Also, do the NT servers run IPX?  Do they need it?

Also, when you recycle the router, how long does it fix the problem?  Only while its down, or a while after its up?  What type of time frame are we talking?
CERTIFIED EXPERT

Commented:
I would also try to remove the IPX protocol from the W9x workstations and switch over to TCP/IP (if your NetWare servers support it).

Author

Commented:
scraiq84:

I have downloaded and installed Ethereal.  Right now the NET USE is behaving.  When it goes south again (and I'm sure it will).  I'll do some packet capturing.

We are talking about mapping NT servers, but Netware servers are on the LAN/WAN.  One of my people doing research thought that the Netware server(s) may be the problem, but we are grasping at any straw.  No concerte evidence that an IPX protocol is the issue.  We run both IP and IPX as native Netware 3.11 does not support IP (I'm sure you knew that).  The NT servers to run Gateway Services for Netware to allow NT servers to map drives from Netware Servers and for some legacy Netware printing.

Recycling the router takes care of the problem for typically a few hours.  Then the problem returns.  Not just when the router is off.  Someone mentioned to me about SAP tables in the router?  Again, for what its worth.
Sr. Systems Engineer
CERTIFIED EXPERT
Top Expert 2008
Commented:
>Recycling the router takes care of the problem for typically a few hours.  

This would make me suspicious of a bug in the router IOS or a memory lead/memory card going south. Is it a Cisco router? Are you routing or bridging IPX? Can you turn on logging on the router and then as soon as the network starts to hang again, check the router logs.

It could also be a configuration issue on the router, something that was tweaked, but the config was not saved. When you powered it off, then back on, it reverted to a previous saved config?

Commented:
You mentioned Gateway Services for Netware - are the Net Use problems related to any share including native NT shares, or just GSN shares?  

Author

Commented:
IrMoore:
We are routing IPX.  I have a call in now to my WAN/Router Reseller with whom I have a maintenance agreement.  I will certainly suggest the logging idea.  Thanks.  

Scraig84:
Any share

Commented:
I meant to comment on the SAP tables earlier and got distracted - sorry.  The only way I can see your SAP tables being a problem is if they are getting overly large for the amount of memory you have in your router, or the transfer across the WAN is impeding other traffic.  In an environmetn as small as yours, the only way I could see this happening is if you have a lot of NT boxes (workstations and servers) with IPX bound to the Server service.  This is a common mistake.  Unless you are actually using the client for MS servers with IPX, this should be shut off.  Otherwise, every workstation and server advertises its services with SAPs, which can cause a pretty large table.  Generally though, this shouldn't be a problem unless there is a fairly large number of NT boxes in the company.
One other thing to note on the router side that may be related.  If the router is having problems with SAPs, the only way it should be affecting workstations is if it is answering Get Nearest Server (GNS) requests too quickly.  Generally if you have actual NetWare servers on your network, you want to crank up the timers a bit to ensure that an actual server answers these requests rather than the router.

Author

Commented:
The router recycling really lead us to it being the issue.  Our WAN folks did some route changes withing the configuration and we have not had the problem for a week (fingers are still crossed).  Irmoore seemed the closest to the final answer as he mentions "It could be a configuration issue", and maybe some tweak was never saved.

Explore More ContentExplore courses, solutions, and other research materials related to this topic.