ADSM-L

NetWare TCP/IP problems

2015-10-04 18:05:24
Subject: NetWare TCP/IP problems
From: Arian de Wit
To: ADSM-L AT VM.MARIST DOT EDU
Hi All,

I am using the latest version of DSMC.NLM (2.01g, March 19, 1997) on
our NetWare 4.10 server, which has the latest TCP/IP installed (the
TCPN04 patch date August 1997).

Backups seem to be hanging on some files (often, but not always,
large ones), and the server times out after 300 seconds, dropping the
connection. This gives the following sequence of errors in the
client's ADSM error log:

23-09-1997 13:40:45  TcpRead: Zero byte buffer read.

23-09-1997 13:40:45  sessRecvVerb: Error -50 from call to 'readRtn'.

23-09-1997 13:40:45  cuConfirm: Received rc: -50 trying to receive
ConfirmResp verb

23-09-1997 13:40:45  TcpFlush: Error 4 sending data
on Tcp/Ip socket 4294967295.

23-09-1997 13:40:45  sessSendVerb: Error
sending Verb, rc: -50

23-09-1997 13:40:45  ANS4017E Session rejected:
TCP/IP connection failure


This sort of problem has been mentioned on this list before (see
below), and the suggestion was to increase the server's timeout from
60 seconds to 300. This got it past some of the files it had trouble
with, but it still has problems with others, some big some small.

The files are definitely not corrupted -- they are still usable in
their respective apps. NetWare volume compression is on, but ADSM
compression is not.

Does anyone know of any problems with NetWare's TCP/IP or the
NetWare ADSM client, or any configuration changes I could try?


Thanks a lot,

Arian


P.S. Here's bits of the previous posting that I found related to
this:

>>>>>>>>>>>>>>>>>>>
Check the messages in your ADSM server log at the time the client got
this error (remember to allow for the likelihood that your server and
client's time clocks are not exactly in sync). Look for this message:

        07/17/1997 14:50:09  ANR0481W Session 5218 for node xxxx
terminated - client did not respond within nn seconds.

I found that most of the problems occurred when clients were
compressing very large files.  Sometimes the client (especially
Windows clients) just gets tied up compressing a large file, and takes
too long to get back to the server. The server then times out the
connection with the ANR0481W message. Then the client IP stack (I
think) issues those errors below cleaning up the debris from the
disconnected session.

I searched the archives of this list for ANS4017E and found several
people recommended changing the server COMMTIMEOUT parm from 60
seconds (default) to 300.  I made that change and it has eliminated
most of the timeouts.

You can also have clients time out when they lose their brains for
other reasons - a disk or directory error that causes the backup
client to loop or lock up, for example.  Those you resolve it by
correcting the problem in the client's file system.

Hope this helps.

> ----------
> Subject:      Errors on Windows/NT Client with TCP/IP
>
> I'm finding that some of my scheduled backups are "dieing" with the
> following messages in DSMERROR.LOG.
>
> (20 minutes later, the backup restarts, and completes quite
> happily)!
>
> Any suggestions from the Net?
>
> 1997 23:18:37  TcpRead: Error reading data, rc: 10054
> 1997 23:18:37  sessRecvVerb: Error -50 from call to 'readRtn'.
> 1997 23:18:37  cuConfirm: Received rc: -50 trying to receive
> ConfirmResp
> verb
> 1997 23:18:37  TcpFlush: Error 10038 sending data on Tcp/Ip socket
> 4294967295. 1997 23:18:37  sessSendVerb: Error sending Verb, rc: -50
> 1997 23:18:37  ANS4017E Session rejected: TCP/IP connection failure
> 1997 23:18:38  ANS4017E Session rejected: TCP/IP connection failure
>
> Thanks
<<<<<<<<<<<<<<<<

 ------------------------------------------------------------------
Arian de Wit, Network Administrator
NZ National Institute of Water & Atmospheric Research (NIWA)
Gate 10 Silverdale Road, Box 11-115, Hamilton, New Zealand
Phone (+64) 7-856 1769; Fax (+64) 7-856 0151
eMail a.dewit AT niwa.cri DOT nz; Web http://www.niwa.cri.nz


<Prev in Thread] Current Thread [Next in Thread>