| Title: | Digital PATHWORKS 32 |
| Moderator: | SPELNK::curless |
| Created: | Fri Nov 01 1996 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 337 |
| Total number of notes: | 1612 |
Hi,
Customer has 50 new Venturis PC's type 5166
Type: FR-A75WW-BD with 32 Mb memory and a Matrox VGA adapter.
Version CPU Boards:
54-24588-02 Rev M03, 54-24588-02 Rev M04 and 54-24544-02 Rev K06
Bios Versions: 4.05 Phoenix Bios and 1.09 FX51661
DE500 adapter. (we also tried a DE435 with same result)
About 30 PC's have a problem with the new license tsr of Pathworks32 running
Windows 95 (English AND Dutch version).
The license retrieving from the license server works without any problems
according to the pwlog file and the license server log file.
When the pc connects to a pathworks server we get in the
PWRK$LICENSE_REGISTRAR_<node>.LOG the following error:
6-MAY-1997 09:38:14.00 MESSAGE:
PingConnect [4,1341]: Cannot receive from client license responder
Client Address: "PC760 R", Transport: NetBEUI
XTI error in IC at line 3126
XTI error code t_errno = 9 ("event requires attention")
XTI event code = 0X0010 ("disconnect received")
XTI system error code = 0
when we look with IRIS we see that the server makes a connection to the client
but the client does not present any license data back to the server. After
a couple of seconds we see a disconnect from the server.
We have installed the debug version of PNLT32.EXE and PNLR32.EXE.
We get the following output:
Log started at 1997-05-05 13:12:20.650
Copyright (c) 1993-1997 Digital Equipment Corporation
Service name is: "PC760 R"
LANA 0x00: Starting first NCBADDNAME
LANA 0x00: Starting thread
Ready on 1 LANAs.
PWCLR_running() == FALSE
GetLastError()=
"De syntaxis van de bestandsnaam, directorynaam of volumenaam is onjuist.
"
PWCLR_start...
LANA 0x00: Starting NCBLISTEN
Entering message loop
Received LIC_GET_TABLE IPC
Received LIC_LOAD_TABLE IPC
LANA 0x00: NCBLISTEN retcode = 0x00
LANA 0x00: Starting NCBRECV
LANA 0x00: NCBRECV retcode = 0x0A, length=164
LANA 0x00: Starting NCBLISTEN
LANA 0x00: NCBLISTEN retcode = 0x00
LANA 0x00: Starting NCBRECV
LANA 0x00: NCBRECV retcode = 0x0A, length=164
LANA 0x00: Starting NCBLISTEN
LANA 0x00: NCBLISTEN retcode = 0x00
LANA 0x00: Starting NCBRECV
LANA 0x00: NCBRECV retcode = 0x0A, length=164
LANA 0x00: Starting NCBLISTEN
Another thing we notice:
On a proper working PC we see using the System Monitor that the Pathworks
license responder takes 100% Kernel Processor Usage.
On the failing pc we don't see 100% but just the normal activity as if the
license TSR is not started.
Output of IRIS trace:
IRIS capture data Page 1
05/06/97 15:38:49
M 0 0.0000s PC760<- 42.103 NETBEUI Name_Recognized DestName=
"PC760 ." SrcName="SENS_
CLUSTER "
1 0.0003s 42.103<- PC760 DLL UI SABME-C DSAP=F0 SSAP=F0
2 0.0010s PC760<- 42.103 DLL UI UA-Resp DSAP=F0 SSAP=F1
3 0.0012s 42.103<- PC760 DLL S RcvRdy DSAP=F0 SSAP=F0
4 0.0019s PC760<- 42.103 DLL S RcvRdy DSAP=F0 SSAP=F1
5 0.0021s 42.103<- PC760 NETBEUI Session_Initialize RSN=1
LSN=12 NB2.0 NOACK=YES MaxRcv=1468
XC=002D RC=000C
6 0.0028s PC760<- 42.103 DLL S RcvRdy DSAP=F0 SSAP=F1
7 0.0031s PC760<- 42.103 NETBEUI Session_Confirm RSN=12
LSN=1 NB2.0 NOACK=YES MaxRcv=1478
XC=000C RC=000C
8 0.0033s 42.103<- PC760 DLL S RcvRdy DSAP=F0 SSAP=F1
9 0.0036s 42.103<- PC760 SMB C NegProt
10 0.0062s PC760<- 42.103 NETBEUI Data_Ack RSN=12 LSN=1
XC=0028
11 0.0064s 42.103<- PC760 DLL S RcvRdy DSAP=F0 SSAP=F1
12 0.0227s 42.103<- PC760 NETBEUI Name_Recognized DestName=
"PWRK$LSTLOT1 P" SrcName="PC760
R"
13 0.0238s 42.103<- PC760 NETBEUI Name_Recognized DestName=
"PWRK$LSTLOT1 P" SrcName="PC760
R"
14 0.0246s PC760<- 42.103 NETBEUI Session_Initialize RSN=9
LSN=2 NB2.0 NOACK=YES MaxRcv=1478
XC=0009 RC=002D
15 0.0247s 42.103<- PC760 NETBEUI Session_Confirm RSN=2
LSN=9 NB2.0 NOACK=YES MaxRcv=1468
XC=002D RC=0009
16 0.0255s PC760<- 42.103 DLL S RcvRdy DSAP=F0 SSAP=F1
17 0.0263s PC760<- 42.103 NETBEUI Data_Only_Last RSN=9
LSN=2 XC=5B18 RC=002D NOACK
ACKWDATA
18 0.0265s 42.103<- PC760 DLL S RcvRdy DSAP=F0 SSAP=F1
19 8.1908s PC760<- 42.103 NETBEUI Session_End RSN=9 LSN=2
XC=3DF8 RC=D317
20 8.1910s 42.103<- PC760 DLL S RcvRdy DSAP=F0 SSAP=F1
21 8.5131s PC760<- 42.103 SMB R NegProt Dialect=-1
22 8.5134s 42.103<- PC760 NETBEUI Data_Ack RSN=1 LSN=12
XC=0027
23 8.5138s PC760<- 42.103 NETBEUI Session_End RSN=12 LSN=1
XC=9998 RC=D4AB
any clue's??
I will post an IPMT next week if we don't find any solution very soon.
regards,
Paul
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 274.1 | JAMIN::WASSER | John A. Wasser | Mon May 12 1997 13:39 | 18 | |
PNLT32.EXE debug log: > LANA 0x00: NCBLISTEN retcode = 0x00 > LANA 0x00: Starting NCBRECV > LANA 0x00: NCBRECV retcode = 0x0A, length=164 > LANA 0x00: Starting NCBLISTEN The connection arrives (LISTEN completes with status 0) but an attempt to receive the request from the License Registrar completes with an error code of 0x0A meaning Session Closed. As far as the Licnese Transponder can tell, the License Registrar opened a connection and then closed it without sending any data. From the network trace it appears that the License Registrar sends something (Packet 17) but that does not get through the NetBEUI stack to the License Transponder. | |||||
| 274.2 | What about 100% cpu usage? | UTRTSC::BAKKER | Running arround in circles | Tue May 13 1997 03:55 | 8 |
John,
Thanks for your comment. One small question. How can you explain the
fact that you don't see the 100% load on this pc while this is the
"normal" behaviour of the license responder?
Paul
| |||||
| 274.3 | JAMIN::WASSER | John A. Wasser | Tue May 13 1997 14:45 | 6 | |
> How can you explain the fact that you don't see the 100% load on this
> pc while this is the "normal" behaviour of the license responder?
It is not a behavior of the License Transponder. It is
a behavior of the NetBIOS interface of the DECnet stack
under Windows 95.
| |||||
| 274.4 | Not Agree and found SOLUTION!!! | UTRTSC::BAKKER | Running arround in circles | Thu May 15 1997 06:19 | 36 |
John,
Sorry but I don't agree. The pc's I have seen don't even use DECnet but
they have 100% Kernel activity!
I have found the real reason!!!!!
SOLUTION:
If you have e.g. a VENTURIS FX, it is equipted with a DUAL IDE
controller.
This is the source of the problem. When you suffer from the problem
metioned in (.0) you have to go the Device manager and look for your IDE
controller. You will probably find two ! (exclamation marks) next to
the two IDE controllers (single Fifo). To solve this error you have to
start REGEDIT:
HKEY_LOCAL_MACHINE\
System\
CurrentControlSet\
Services\
VxD\
IOS\
Remove from this point the key NOIDE. This will fix the problem.
Restart Windows95 but beaware that the restart finished without a hang
otherwise do the change in SAFEMODE.
regards,
Paul Bakker
P.S. John After this change I did get back my 100% kernel activity and
I had ONLY NETBEUI enabled.
| |||||
| 274.5 | VMSNET::N_WILKINSON | 16 bit? We don't need no stinkin' 16 bit! | Thu May 15 1997 08:40 | 9 | |
I'll side with Paul here. I also have 100% CPU utilization with no DECnet stack (as reported in note 123.11 and following). Something other than the DECnet stack IS causing 100% CPU utilization. By the way, I have a Celebris 590 with dual IDE controllers but do not have duplicates as -1 does. Nelson Wilkinson ALF CSC | |||||
| 274.6 | DECnet is just one possible cause, not the only one. | JAMIN::WASSER | John A. Wasser | Thu May 15 1997 15:03 | 12 |
> John, > Sorry but I don't agree. The pc's I have seen don't even use DECnet but > they have 100% Kernel activity! I did not mean to imply that DECnet was the only possible cause of 100% CPU utilization. It is just the only cause I know of that is associated with Licensing. > I have found the real reason!!!!! I think it would be closer to say that you have found yet another reason for the CPU utilization to peg at 100%. | |||||