AdvancesServer 7.3A2 [bd]

[Doug Phillips]

OpenVMS 7.3-2 update 3
AdvancedServer 7.3A2 (as BD)
TCP/IP v5.4 eco2

Primary domain controller is a WinNT v4 and clients are various 95/XP/W2K.

Just updated system to above from VMS 7.2-1 w/ AS 7.3, all settings in AdvancedServer (ie. domain, transports, etc) were left the same but now:

1. it doesn't show up in client's Network Places/Neighborhood, but find computer will find it.

2. it shows as off-line [bd] from admin show computers, and but it shows as [BD] from admin show computer/full (name).

3. client's autoconnect drive map shares don't, and show X'ed in MyComputer but when clicked open, the first asks for Aministrator password and opens, and the rest will then open when clicked.

I googled the BIND problem and finally got PWIP & KNB up but is there another gotcha that I haven't found?

Doug

[DICTU OpenVMS]

Not sure about this, but after upgrade the config file's are moved to a save place and fresh ones are installed. Take a look in PWRK$LMROOT:[LANMAN.DATAFILES].

[Antoniov.]

Doug,
do you upgraded pwrk licenses from CA07.02 to CA07.03?
If you didn't upgraded, do you enabled old license use?
If you would use old license, you cannot manager server based license model.

Antonio Vigliotti

[Paul Nunez · Accepted solution]

Hi Doug,

The first problem is a browser issue. When Advanced Server starts it will broadcast a host announcement message on the local subnet and the local master browser should receive that announcement and add the server to its browse list.

Browsing can be tough to troubleshoot. You need to find the local master browser on the same subnet as the Advanced Server. One way to find that system is to do:

$ @sys$startup:pwrk$define_commands
$ nbshow knbstatus 1d

You should get a response which displays a list of NetBIOS names registered by the responding system. Due to a bug in the nbshow utility, you MUST examine the list to verify the name you searched for is in the list (the bug can result in a response from a system which does NOT own the queried name). The local master browser will own the netbios name \0x1D. An example, for the domain named MNOSNAT:

$ nbshow knbstatus mnosnat 1d
Adapter status for mnosnat on
ID: 00 50 8B 11 9C 19 Version 0.0
Time up: 0 days 0 hours 0 minutes
Packets sent: 0 CRC errors: 0
Packets received: 0 Alignment errors: 0
Retransmitted packets: 0 Collisions: 0
Resources exhausted: 0 Aborted transmissions: 0
Ncbs: Free : 0 of 0; maximum configurable: 0
Sessions: In use: 0 of 0; maximum configurable: 0
Adapter packet size: 0
Local name table (11 names):
Name Soc Num Status
HOLTZ x00 1 Unique Registered
MNOSNAT x00 2 Group Registered
MNOSNAT x1c 3 Group Registered
HOLTZ x20 4 Unique Registered
MNOSNAT x1b 5 Unique Registered
MNOSNAT x1e 6 Group Registered
MNOSNAT x1d 7 Unique Registered
HOLTZ x03 8 Unique Registered
^^__MSBROWSE__^ x01 9 Group Registered
INet~Services x1c 10 Group Registered
IS~HOLTZ^^^^^^^ x00 11 Unique Registered

You can see in the above list of names the entry for the local master browser (actually, it will own 2 names):

MNOSNAT x1d
^^__MSBROWSE__^ x01

Once you've identified the local master browser, you'll want to logon to that system and look at Network Neighborhood to determine if the Advanced Server name is present. If not, this is where the problem lies - either the local master browser isn't receiving the host announcement broadcast (most likely) or the Advanced Server isn't sending one (i.e, because the registry parameter srvhidden=yes).

If the Advanced Server name is in the Network Neighborhood display on the local master browser, the next place to look is the Domain Master Browser, which is always the domain PDC. Is the Advanced Server name listed in its Netowrk Neighborhood? The PDC is suppose to periodically communicate with its local master browsers on other subnets (the PDC itself is also the local master browser on its subnet) and exchange browse lists, thus the PDC gathers a network-wide list of domain server resources and passes that complete list to each of the local master browsers. You can look at Network Neighborhood on the local master browser of other subnets to determine if that is occurring.

On the Advanced Server itself, you just need to be sure the IP Broadcast address is correct.

The problem with mapping drives at logon is difficult to diagnose based on the information provided. Is there any error on the client side? Is the problem consistent across all client O/Ses and does it occur every time they logon? Any related errors seen in:

$ admin show event/full/since=
$ admin/analyze/since=
$ type/tail pwrk$lmlogs:pwrk$lmmcp_.log
$ type pwrk$logs:pwrk$knbdaemon_.log
$ type sys$manager:tcpip$pwipacp_.log

You may need to get a network trace of a client failing:

$ tcptrace/full/packet=10000/buffer=300/output=anyfile.txt

Then have the user logon the client. Stop the trace after the logon has completed and you see the drives 'x'd out. You can send the trace to me for analysis: paul.nunez@hp.com.

Hope this helps,

Paul

[Brad McCusker]

Clarification:

Antoniov wrote:

>do you upgraded pwrk licenses from CA07.02 >to CA07.03?

He upgraded from Advanced Server V7.3 to Advanced Server V7.3A. No upgrade of change in license is required for that upgrade.

Regards,

Brad McCusker

[Doug Phillips]

Thanks Paul,

The knbdaemon log was reporting a failure to find a domain (LANGROUP) that once had a trust relationship with this one but has since been "emptied". I removed the trusts but don't think that has anything to do with my problem. I do need to get all of the references to that domain removed, though (any shortcut hints?).

However... I promoted the AdvancedServer bd to PD. Then "admin show computers" listed everyone else lowercase ( [ws] [bd] ) except the AS [PD].

So, I promoted the WinNTv4 back to PD and now show computers lists everyone's status correctly.

The workstations autoconnect to the AS shares, but now they don't autoconnect to the WinNTv4 shares without a password. There are only 5 ws (3 XP & 2 Win2k) that I can play with except by special arrangement after hours and they all act the same AFAICT (they connect to different shares).

I'll have to work on this from the NT server side when I can get back in there.

I think I need to work on the TCP/IP BIND service a bit more and restart TCP/IP because I'm getting some indications that the server isn't really running (tcpip show hosts returns E-BIND_NOSERVERS, tcpip show services indicates bind is disabled)although tcpip$config shows that both server is enabled and started and the resolver is configured. I think I can straighten that out but I'll be back if not.

As a reminder, this network was working fine prior to the upgrade mentioned in the original post.

I'll leave this thread open for awhile so please (anyone) feel free to comment.

Thanks,
Doug

[Paul Nunez]

Hi,

You're right, the knbdaemon log has nothing to do with the problem and now that you've removed the trust to the langroup domain, you shouldn't see those messages any longer (you might also see if there are any references in pwrk$lanman:lmhosts.)

Glad the browser issue is resolved.

As for the BIND setup on the OVMS system, keep in mind you have 2 components - the BIND Server and the BIND Resolver. My impression is you're trying to configure just the resolver so you avoid the pwip bug when it's not enabled.

$ TCPIP SHOW SERVICES shows the BIND Server service while $ TCPIP SHOW NAME shows the BIND resolver configuration ($ TCPIP SHOW CONFIG NAME shows the permanent resolver configuration).

So if you have other DNS servers on the network, you probably don't need/want the BIND Server running on the OVMS system (this is pure conjecture on my part :), you just need to configure the BIND resolver to use those other DNS servers.

And while everything was working prior to the OVMS upgrade, there's really very little chance that could affect the clients automapping drives to the NT system :) (well, I guess if the OVMS box were the DNS server and cleints relied on DNS to find the NT box, then the upgrade could be the culprit).

Paul

[Doug Phillips]

**CORRECTION**

It's Advanced Server v7.3a ECO 3,
CPQ AXPVMS ADVANCEDSERVER V7.3-A3

After running okay for a while, the AdvancedServer again shows as [bd] and all other symptoms are back.

The only log that shows anything strange is the knbdaemon log:

Tue Nov 16 15:44:23 2004 get_phys_addr: Searching known Physical device list
Tue Nov 16 15:44:23 2004 get_phys_addr: EWA2: PH Address: AA-00-04-00-05-04
Tue Nov 16 15:44:23 2004 IP Address: 192.168.1.100
Tue Nov 16 15:44:23 2004 NBNS IP Address: 192.168.1.9
Tue Nov 16 15:44:23 2004 ip_brdcst_address : 192.168.1.255
Tue Nov 16 15:44:23 2004 KNB Driver linked with successfully
Tue Nov 16 15:44:23 2004 KNB Driver : Starting knblink ...
Tue Nov 16 15:44:23 2004 KNB Driver : Starting knblmhost ...
Tue Nov 16 15:44:23 2004 KNB Driver : Starting knbbindresolver ...
Wed Nov 17 14:29:33 2004 resolve_name: LMHOSTS Lookup for 1B [20] failed knbcb 9
Wed Nov 17 14:29:41 2004 resolve_name: LMHOSTS Lookup for 1B [ 0] failed knbcb 9
Wed Nov 17 14:29:48 2004 resolve_name: LMHOSTS Lookup for 1D [20] failed knbcb 8
Wed Nov 17 14:29:51 2004 resolve_name: LMHOSTS Lookup for 1D [ 0] failed knbcb 8

There are no events recorded since PWRESTART and the other logs just show the startup.

I don't know enough about nbshow to interpret its display, but here's what it says. (the [BD] AS computer is MICKEY, the [PD] WinNT is GRUMPY and the domain is GRAPHICS and there are no other servers and there are no sub-nets)

$nbshow knbstatus graphics 1d
Adapter status for graphics on
ID: 00 00 F8 1E E3 E1 Version 0.0
Time up: 0 days 0 hours 0 minutes
Packets sent: 0 CRC errors: 0
Packets received: 0 Alignment errors: 0
Retransmitted packets: 0 Collisions: 0
Resources exhausted: 0 Aborted transmissions: 0
Ncbs: Free : 0 of 0; maximum configurable: 0
Sessions: In use: 0 of 0; maximum configurable: 0
Adapter packet size: 0
Local name table (14 names):
Name Soc Num Status
GRUMPY x00 1 Unique Registered
GRUMPY x20 2 Unique Registered
GRAPHICS x00 3 Group Registered
GRAPHICS x1c 4 Group Registered
LANGROUP x00 5 Group Registered
GRAPHICS x1e 6 Group Registered
GRUMPY x03 7 Unique Registered
ADMINISTRATOR x03 8 Unique Registered
E^Gg^^^:^^"`ZJT x71 9 Unique Registered
@GRUMPY x20 10 Unique Registered
GRUMPY x01 11 Unique Registered
GRAPHICS x1b 12 Unique Registered
GRAPHICS x1d 13 Unique Registered
^^__MSBROWSE__^ x01 14 Group Registered

Note that LANGROUP shows up here, too, so that dead domain is still referenced someplace.

Here's mickey's status:

$nbshow knbstatus mickey
Adapter status for mickey on
ID: AA 00 04 00 05 04 Version 3.0
Time up: 1 days 21 hours 14 minutes
Packets sent: 0 CRC errors: 0
Packets received: 0 Alignment errors: 0
Retransmitted packets: 0 Collisions: 0
Resources exhausted: 0 Aborted transmissions: 0
Ncbs: Free : 65535 of 65535; maximum configurable: 65535
Sessions: In use: 17 of 45; maximum configurable: 45
Adapter packet size: 1497
Local name table (10 names):
Name Soc Num Status
MICKEY x20 1 Unique Registered
MICKEY x00 2 Unique Registered
GRAPHICS x00 3 Group Registered
GRAPHICS x1c 4 Group Registered
GRAPHICS x1e 5 Group Registered
MICKEY#D x00 6 Unique Registered
MICKEY#B x00 7 Unique Registered
MICKEY_29F x00 8 Unique Registered
MICKEY_2A1 x00 9 Unique Registered
PWRK$LMICKEYR01 x50 10 Unique Registered

Any suggestions as to what to look for next to get this browser problem resolved?

Doug
ps. Okay, so GRUMPY wasn't a good name but he's an Alpha AS1000 who's been working just fine for 8 years;-)

[Paul Nunez]

Hi Doug,

The nbshow output shows the NT PDC is the master browser. Is the Advanced Server also missing from the PDC's nethood?

There's no easy fix. You can use some NT tools such as browstat or browmon to look at browser status or get network traces and look for the host announcement messages from the Advanced Server...

Advanced Server could be broadcasting them to the wrong address or a whole host of other odd, unexpected things could be occurring.

Paul

[Doug Phillips]

Thank you for your help, Paul.

I'm going back on-site tomorrow so I'll get to look at the NT server then. I can get to the VMS AS system remotely but not the NT and the set admin/server doesn't tell me what I need to know.

This old Alpha NT system is probably going to be replaced soon because there are things we want to do that aren't available for Alpha NT. But, for now it's still a workhorse and I do need to get this problem solved.

I'll post again when I know more. Thanks again

Doug