I could use a bit of help on a printing issue, please.

We have a number of Linux servers (RHEL 5 & 6) that we need to print from, and we have the following options:

1. Setup the printers on Linux in CUPS, individually host-by-host.
2. Use Samba to have the Linux systems print to the printer queues that have already been set up on a Windows server.
3. Use LPR/LPD to have the Linux systems print to the printer queues that have already been set up on a Windows server, WITHOUT SETTING UP A PRINTER QUEUE ON THE LINUX SERVERS.
4. Setup a Linux CUPS print server parallel to the Windows one.

We've done (1), but it is proving to be a bit difficult to manage, and is replicating work already done.

We don't like (2) because it again requires us to do setup on the Linux systems.

We don't like (4) because it is replicating the Windows printing configuration (and the business won't be wanting to shift the Windows printers to Linux so the Windows ones have to be there).

So, I've been exploring (3) - using LPD and just can't quite get it to do what I want.

I have a HP laserJet printer "pr202" that supports LPD and has a builtin queue "test".
I have a Windows server "ms579p" with LPD configured and with a printer queue "pr202" set up.

Printing directly to the Laserjet from the Linux command line works:

date | lp -h pr202 -d test

And issuing a print command on Windows works:

lpr -S ms579p -P pr202 testfile

But I don't seem to be able to print from Linux to the Windows queue:

lp -h ms579p -d pr202

I can telnet and establish a connection from the Linux server to port 515 on the Windows server, so it does not appear to be a firewall issue.

If I set up a (CUPS) print queue on the Linux server to the LPD interface on the Windows printer and it successfully prints.

So, either (a) I am using the lp command wrongly, (b) the "-h" option isn't working, (c) the CUPS lp command isn't talking LPD to the Windows server, or (d) Windows doesn't like what it's getting from Linux.

I've also tried specifying "-h ms579p:515" and variations on the Windows UNC "//ms579p/pr202" and the URI format "lpd://ms579p/pr202", just to see if they work, but no joy.

There was an "rlpr" command to issue remote prints, but it seems to be deprecated now.

So, can anybody help me out here, please?

To recap, basically I want to set up queueless printing from Linux to Windows (ie. issue remote prints from the command line without doing any configuration on the Linux systems).

Moved: This thread is more suitable in <Red Hat> and has been moved accordingly to help your thread/question get the exposure it deserves.

Thanks!

What do you mean by “rlpr” is deprecated - you can still download and install it.

Hi.
I've looked around and it doesn't seem to be being maintained anymore, and I've tried several download links but they turn out to be dead.
Do you use it?
Thanks.

I just tried it for curiosity on my Mac. I download rlpr-2.06 and it compiled and worked out-of-the-box.

It looks like Mandriva and Ubuntu support it, but I couldn't see Red Hat/CentOS/Fedora canned releases.
I found a source for it at: ftp://metalab.unc.edu/pub/Linux/system/printing/.
(It is kind of funny that the one you picked up is under an "old" directory, which is sort of where I was going with the comment about it being deprecated.)
I'll get it onto a Red Hat box and give it a go.
But the probem I have is that I'm not sure the business will go for it as a viable piece of key infrastructure.

Well, if it’s stable and still working maybe there is no need to maintain it, unless you want to add features. Nevertheless I would also be interested whether it can be done without the application. Maybe local spooling is so common nowadays, that no one want to print directly any longer to a remote server. Of course, having it in a chain of spoolers it’s hard to cancel a print job (local spooler, remote spooler and finally the one which is nowadays in the printer itself).

Do you want to use you print spooler for a special reason, or why not printing directly to the printer even by rlpr?

Well - I compiled it and it worked 2nd time (1st time I got the syntax wrong!).

The company I work for is a global organization with printers in multiple countries, and we would prefer to spool prints via a single (or clustered) print server, to give manageability. There is an M$ email, document, etc. infrastructure in place and the business applications (databases, etc.) need to share their printers. The particular application I'm looking at at the moment is an Oracle based one and it issues prints via a shell script somewhere in its gubbins.

We'll have to do some testing to see whether this approach is robust enough.