Re: Using QSH to concatenate stream files on remote system

[CRPence <CRPbottle@xxxxxxxxx>]

  So how about using cat to stdout instead of writing to /g [in /QNTC]; 
i.e. drop the redirect.  Does that function well, or result in the same 
error?  If that also fails the same, perhaps try cat a couple local file 
with redirect to /g to implicate the write instead of the read as origin 
for the encountered difficulty.

  Not sure where the msg CPDB050 is logged [apparently the interactive 
joblog invoking the QSHell facility vs one of the /batch/ jobs that 
implements the work], but are the full message details [F6=Print or 
spooled joblog] available... to clarity if perhaps the errno 3447 
ETIMEDOUT condition is specific to the readdir or something else.?

  I did not find any APARs with msgCPDB050 rc3447 nor various kwd 
combinations with either errno3447 ETIMEDOUT

  I guess I would not be surprised to find that directory related 
processing functions differently than the actual I/O; i.e. such that 
WRKLNK could navigate fine, but the actual file I/O might fail.  An 
EBUSY condition or similar, might be relevant to one but not another, so 
the methods are sufficiently different they might encounter an impasse 
distinct from one another.?

  Note: I recall one or two environment variables related to how /QNTC 
server connections are cached\tracked.  Probably best to look into what 
if any exist and settings... and try to find any documented that might 
change the effects.  I did run across a reference to QZLC_SERVERLIST 
with a value of "1" or "2", and that was referenced in the following 
link, but I did not research further:
<http://www.ibmsystemsmag.com/ibmi/administrator/systemsmanagement/A-Look-at-File-Systems/?page=2>

  With that a quick look in InfoCenter also found QIBM_ZLC_NO_BROWSE:
<http://publib.boulder.ibm.com/infocenter/iseries/v5r4/topic/ifs/rzaaxqntcvariables.htm>
_QNTC environment variables_
"The network browsing behavior of QNTC can be controlled by two 
environment variables. Support for these environment variables began in 
i5/OS® V5R4. Use the ADDENVVAR CL command to create these environment 
variables.
 ..."

Regards, Chuck

On 01-Feb-2014 13:08 -0800, Peter Dow wrote:
> Yes, I was, i.e. I did a
>
> cd /g
>
> before doing the cat command.
>
> On 1/31/2014 12:26 PM, Jack Kingsley wrote:
> > Peter, when you start to do this CAT sequence, are you in the g
> > directory??
> >
> > On Fri, Jan 31, 2014 at 2:52 PM, Peter Dow wrote:
> >
> > > I'm working on a v5r4 system that has a symbolic link to a remote
> > > system:
> > >
> > > Object link Type
> > > g SYMLNK->DDIR
> > >
> > > Content of Link . . . : /QNTC/OP-2003-SQL1/Galt2008
> > >
> > > There are several files in that folder that I want to
> > > concatenate, so from an OS400 command line, I started QSH and
> > > entered:
> > >
> > > cat asp_445.csv asp_445_canada.csv asp_445_no.csv > asp_445_all.csv
> > >
> > > It sat there forever, so I canceled it and looked at the joblog,
> > > where I see that I got a CPDB050 with error code 3447, indicating
> > > a timeout waiting for the remote host to respond.
> > >
> > > When I do a wrklnk '/g', I immediately see the folder and can
> > > look at all the files within it. So what's the deal? wrklnk uses
> > > a different method of accessing /g than qsh? If so, what can I do
> > > about it?