[comp.sys.hp] Canceling remote print jobs

gordon@maxwell.waterloo.edu (Gordon R. Strachan) (05/15/91)

Okay, I admit defeat, I have looked in every manual I could find but to no
avail.  Here is the situation, a Sun machine (I have tried this from a Sun
SLC and a 386i) submits a print job to an HP9000/345 running HPUX V7.0.  The
printer is connected to the HP9000/345.  The user on the sun then attempts to
cancel the print job.  If he types lprm -Pprinter, nothing happens.  If he
types lprm -Pprinter userid or lprm -Pprinter jobID then the following error
is retured

Maxwell: cfA034bbcr: Permission denied

Where maxwell is the HP9000/345's host id bbcr is the Sun's host id and 034
is the job id.  As near as I can tell from any of the manuals, there is no
setting to control the permissions and removing jobs from the remote host.
The rlpdaemon is running (launched from the inetd with the -i and -l options)
and logs the rmjob command as follows: 

rlpdaemon: bbcr.waterloo.edu requests rmjob rwriter bbcr:gordon 34

I can cancel jobs from any of the nodes within the HP cluster, it is only
from outside the cluster that we have problems.  Also, the cancel is not
restricted.

Anybody seen this one before?  Got any idea how to fix it?

Thanks
Gordon

gaute@stud.cs.uit.no (Gaute Nessan) (05/15/91)

In article <1991May14.192739.546@sunee.waterloo.edu> gordon@maxwell.waterloo.edu (Gordon R. Strachan) writes:
>...  The user on the sun then attempts to
>cancel the print job.  If he types lprm -Pprinter, nothing happens.  If he
>types lprm -Pprinter userid or lprm -Pprinter jobID then the following error
>is retured
>
>Maxwell: cfA034bbcr: Permission denied
>
>...
>
>Anybody seen this one before?  Got any idea how to fix it?

The problem is that the rlpdaemon uses gethostbyaddr() to get the hostname
of the peer host. gethostbyaddr returns a domain qualified name (aaa.bbb.edu).
That name is different from the hostname in the spooling queue, and you get
the error message "Permission denied".  I have a patched version available
by anonymous ftp to hpserv1.cs.uit.no (HP/rlpdaemon.Z).

>Thanks
>Gordon

Gaute
-- 
//// Gaute Nessan                 // Email: gaute@staff.cs.uit.no            /
/// Computer Science Department  //                                         //
// University of Tromsoe        // Fax: +47 83 44580                       ///
/ N-9000 Tromsoe, NORWAY       // Phone:+47 83 44053                      ////

gordon@maxwell.waterloo.edu (Gordon R. Strachan) (05/16/91)

In article <1991May15.142107.25132@mack.uit.no> gaute@stud.cs.uit.no (Gaute Nessan) writes:
>In article <1991May14.192739.546@sunee.waterloo.edu> gordon@maxwell.waterloo.edu (Gordon R. Strachan) writes:
>>...  The user on the sun then attempts to
>>cancel the print job.  If he types lprm -Pprinter, nothing happens.  If he
>>types lprm -Pprinter userid or lprm -Pprinter jobID then the following error
>>is retured
>>
>>Maxwell: cfA034bbcr: Permission denied
>
>The problem is that the rlpdaemon uses gethostbyaddr() to get the hostname
>of the peer host. gethostbyaddr returns a domain qualified name (aaa.bbb.edu).
>That name is different from the hostname in the spooling queue, and you get
>the error message "Permission denied".  I have a patched version available
>by anonymous ftp to hpserv1.cs.uit.no (HP/rlpdaemon.Z).

After I posted this I thought about it some more and I was wondering if that
was the problem (I had noticed that the rlpdaemon log was using both unqualified
and fully qualified domain names).  Thank you for the responce, I will FTP the
code.  I have a dream that some day Unix will actually figure out domain names
and will understand that they can be qualified or not and still be the same
host.  Then we won't have this problem with every other program.

>
>Gaute
>-- 

Gordon

lienhart@hpfcdc.HP.COM (Bob Lienhart) (05/17/91)

>The rlpdaemon is running (launched from the inetd with the -i and -l options)
>and logs the rmjob command as follows: 
>
>rlpdaemon: bbcr.waterloo.edu requests rmjob rwriter bbcr:gordon 34
>
>I can cancel jobs from any of the nodes within the HP cluster, it is only
>from outside the cluster that we have problems.  Also, the cancel is not
>restricted.

What shows up in lpd.log when you try to cancel a request from
outside the cluster.  Also what shows up when you submit a request
from within the cluster and outside the cluster.

What version of HP-UX are you running and on what HP9000 series
model?  What is the result of "what /usr/lib/rlpdaemon"?

NOTE:  The request must be cancelled from the same system that
submitted the request and by the same user on that system.

Bob Lienhart