Chris Breemer writes:
@
@ I took the new version and tested it. Seems to run into problems rather soon:
<snip working example deleted >
Now look I put the new toy out there, but I didn't expect you to break
it.
@ Works great, no sweat. Now something a little more daring :
@
@ : RUN REMSH; info="felix -l chris ls -l"
@
@ This hangs forever, although there are only 112 files to display. I aborted
@ it after some 5 minutes (I'm not being impatient, am I ... ?)
@ So what can be the snag here ?
Hmm, I thought I'd tried this. Oh yeah, Well this is embarassing.
Here's the deal. the HPUX implementation has the option -n to force
disabling of STDIN. For the longest time I was planning on hard coding
this true on the MPE iX version. At the last minute, I changed it back to
the more traditional Unix implementation to allow you guys to make that
choice for yourself.
shell/iX> REMSH hpcsyn24 -l casc -n ls
works just fine
as does run remsh;info="hpcsyn24 -l casc -n ls"
So my recomendation at this time is, if it hangs, try running
with the -n option.
I'll have a look and see if I can find out what the hang is and if
there's a way to work around it. I may go back in and have to force the
map STDIN to bit bucket (/dev/null) for the production version. Any
thoughts on that from you guys?
@ >
@ (sob) I for one have been _craving_ for a remshd for a long time ! There was a version
@ on Jazz but it did not work, as other on this list reported.
@
@ If I really want to
@ integrate the 3k with other machines I should be able to initiate commands/jobs on the 3k
@ from other machines, like we do on other servers. I've been kludging with vt3k but
@ that's so ugly... Why oh why is remshd dropped ? Will it ever be supported ? Offering
@ remsh without remshd seems like on of these half mesaures again to me.
@
@ Pleez, HP ?
Well, remshd was not dropped. So far as I know it's never been planned.
The reason you're seeing the half measure, is this is an attempt to get
job initiated commands sent to non-mpe machines. In other words, I was
unable to get telnet client to run under the MPE job environment with a
reasonable amount of work. However, remsh client came over relatively
easily.
If y'all want remshd, we'll need to build a business case and get it
funded through one of the solution teams. I can enter the base SR for
tracking, but it will be perceived as more important if one of you takes
the time to enter it via normal support channels. If you do enter an SR
tell them you want it against the INTERSRV product. That's the product
number we use for tracking the Internet Services products introduced with
5.5 (except telnet).
If you want to register support for implementing remshd but don't want to
enter an SR, you can send me e-mail and I'll start builing a list. If
there's enough of you sending me mail, I'll enter the SR myself. (Chris
you're already on my list <grin>).
I looked at the remshd issue enough to know it's not a quick and easy
implementation. That's why it's likely to be something one of us can't
add via CPE or Gee-job routes. Sorry.
Actually while I think of it, how would y'all feel about a remshd that
did not accept rlogin connections? Mapping MPE TIO into non MPE TIO is
really tough to do.
Seriously thanks for bringing the problem with remsh to my attention.
Look forward to hearing from more of you,
cas
CSY Networking
--
=======================================================================
[log in to unmask] (Cas Caswell) By the way: I said it, not my company.
=======================================================================
|