[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference noted::pwnt

Title:	PATHWORKS for Windows NT
Notice:	See note 11.64 for a temp pointer to the PWNT V4.1B SSB kit
Moderator:	VAXCPU::michaud

Created:	Fri Oct 30 1992
Last Modified:	Thu Jun 05 1997
Last Successful Update:	Fri Jun 06 1997
Number of topics:	1009
Total number of notes:	3881

1003.0. "DECnet dies, resource issue?" by VMSNET::BAUER () Tue Apr 01 1997 14:23

    
    Does this look like a resource issue? A bug,  incompatablity or what?
    
    
    I have a customer with an NT cluster of 2 Prioris zx6000 SMP  systems
    running Windows NT Version 3.51 with Service patch 5 and PATHWORKS for
    Windows NT v4.1b.  Each PC has 2 (one for clustering) DECPCI Fast
    Ethernet DECchip 21140 ethernet devices.
    
    They have Oracle SQL version 7.3 running on NT and Oracle SQL version
    7.1 running on the VAX.
    
    They do some processing on the VAX, then run a batch program that
    inputs the data from the database on the VAX to the NT systems.
    
    It all worked ok with minor loads during their startup phase, but now
    they are in full production, and DECnet can't seem to handle the load
    
    It now puts a few entries into the NT's database, then stops.
    DRWATSON.LOG gets exception c000005 errors on PWSVC.EXE.
    
    At this point, if you go into NCP, even a  SHOW EXEC fails with a 
    listener response error. Can't loop, you get a file open error.
    Can't look at counters.
    
    He has to reboot (sometimes more than once) for DECnet to come back
    alive.  
    
    After the reboot, the NCP counters (LINE, CIRCUIT, EXEC) are clean.
    
    While the first node in the cluster is down, the failover pc will also
    experience the problem.
    
    
    I found an update of PWSVC.EXE DATED 9/20/96 and gave it to the
    customer (he already had the latest NBD, NBDLAN, PWSOCK, DNA4.SYS and
    FAL.EXE).
    
    This keeps the exception messages from coming into DRWATSON, but the
    symptom is still there.
    
    The only thing at all weird in the Event Log is in the System log,
    there are thousans of event 2009 errors, which say server could not
    expand a table because table reached maximum size'. These are RED 
    stop sign messages.
    
    The customer doesn't believe these messages are the root cause, since
    the failover pc does not get them, and it loses DECnet as well.

T.R	Title	User	Personal Name	Date	Lines
1003.1	should I ask here? or in pw32?	SUFRNG::WSA028::DOWNING_L		`Mon Apr 07 1997 15:04`	2
	Not much activity in this conference anymore. Should I move my question over to the pw32 conference?
1003.2		SPELNK::curless	You're not done til you CLD/QAR it	`Mon Apr 07 1997 17:42`	17
	Well, unless the customer wants to upgrade... Actually, I'd rather: 1) You file a CLD 2) Stop giving the customer unoffical "pieces and patches". We have had to (several times) have customers reinstall because they installed "new components" that would "solve the problem", when those fixes either required OTHER components that were not added, or ... where not really related to the issue. Have you talked to them about upgrading? Jeff (the other one)
1003.3		JAMIN::WASSER	John A. Wasser	`Tue Apr 08 1997 10:34`	11
	> Not much activity in this conference anymore. > Should I move my question over to the pw32 conference? Only if the question is about PATHWORKS 32. For PATHWORKS V4.1 for Windows NT, this is still the place for unofficial answers and "as time permits" support. If you want real answers or real support, report through official channels.