Linx client stops working

Question

Linux Client v1.1 (for NLD) running on OES SP2. Our OES server is
running GroupWise, and to get a backup I've written a batch script to
log into a NetWare server, "map" a drive and then start a DBCopy.
Everything works fine until approximately halfway through the copy, when
around 8Gb of data has been copied. Then the "mapped drive" disappears,
and the copy starts failing.
If I try a "nwconnections" command, I get the following:
bcw-groupwise:~ # nwconnections
Cannot get connection list: Cannot scan Connection References.
Error:NWCCScanConnRefs:
If I try to login:
bcw-groupwise:~ # nwlogin
Tree: bcw_group
Server: bcw_1
Unable to locate tree bcw_group
However, if I do a ncl_control restart, I can log back in normal, map
drives without problems etc.
Why does the client "die" after copying large amounts of data, and how
can I fix it?

Answer 1

On Wed, 08 Mar 2006 13:45:03 +0000, Kenny Anderson wrote:
> Linux Client v1.1 (for NLD) running on OES SP2. Our OES server is
> running GroupWise, and to get a backup I've written a batch script to
> log into a NetWare server, "map" a drive and then start a DBCopy.
>
> Everything works fine until approximately halfway through the copy, when
> around 8Gb of data has been copied. Then the "mapped drive" disappears,
> and the copy starts failing.
>
> If I try a "nwconnections" command, I get the following:
>
> bcw-groupwise:~ # nwconnections
> Cannot get connection list: Cannot scan Connection References.
> Error:NWCCScanConnRefs:
>
> If I try to login:
>
> bcw-groupwise:~ # nwlogin
> Tree: bcw_group
> Server: bcw_1
> Unable to locate tree bcw_group
>
> However, if I do a ncl_control restart, I can log back in normal, map
> drives without problems etc.
>
> Why does the client "die" after copying large amounts of data, and how
> can I fix it?
IMHO, It probably has something to due with the clients ability to keep
the SLP data and connections pathways open while moving so much data. I
would try eliminating anything in your script that relies on SLP or DNS to
resolve, use the ipaddress of the trees root master(to Login) and the ip
address of the groupwise server. This should eliminate timeouts on
requests for services, caused by collisions in "ack" packets. JT