You are not logged in.
This problem is not solved as such, but it has gone away. The HP 3830 is described as a bi-directional printer (as most are now) and communicates with the computer in both directions. That allows for hand-shaking, and I suspect hand-shaking is required because of the multiple functions of the printer which can also be used as a copier and a scanner.
Obviously the hand-shaking protocol requires something as sophisticated as CUPS, and the more primitive LPR can not handle it. Also the printer does not react kindly to simply having a character string dumped on it, when its response is ignored.
I have set up CUPS on one Devuan machine, and used HPLIP to set up the printer. So now I can entertain myself by learning the mysteries of CUPS, and getting the printer available on all three machines in my setup. One complication to be dealt with is that the cups-bsd package is required to provide the familiar lpr commands, and it conflicts with lpr -- one has to make a choice.
The printer doesn't print what is sent, but it does respond with its electronic jingle noise. I am sure the problem is some printer control command, and that that must be in the driver files that come with hplip.
As I can in fact print, the objective at this to point is just to find out what is happening when the printer is connected to Linux. One option is for me to set up CUPS on the computer that doesn't usually have the printer connected to it, but it might also be interesting to capture the print file on windows and look at what characters are in it.
As I said, it is a game!
I am running the test as root. The only user is myself, and in my usual persona I belong to the lp group.
My system consists of two devuan machines, one windows machine and a wifi access point. All the machines can access one another's file systems.
It isn't the end of the world if I have to connect the printer to windows, but I am trying to avoid the complexity of CUPS for one lousy printer, and I also like to be in control and understand what is happening. That's my main reason for using Linux where I can access good usable compilers and do some things for myself. It's my form of computer game -- it's entertainment.
I think all that they do is spool the output and then send it to the printer. I get the same null response using a raw printer defined in /etc/printcap. This mechanism worked perfectly well with an older HP printer, so that I could in effect write my own simple filter for text files. I suspect the HP 3830 has a much more sophisticated set of controls since it can be used as a copier and a scanner. I think I am missing some command that simply says "This is a computer output file; just print it." The printer is responding audibly, but the proper handshaking is not happening.
My HP 950C printer is dying, and I am attempting to set up a new HP 3830 in the same way. The old printer had a parallel port connection and I used a USB adapter to connect it to a Devuan computer. It was identified by the system as /dev/usb/lp0 and handled by LPR. I reverted to LPR because CUPS repeatedly dropped the device, and in any case it seemed to me to be unnecessarily complicated for a system with a single fixed printer.
The HP 950C was set up in raw mode on a Devuan machine, and could be used with some embedded PCL commands to print text files. For fancy printing it was connected as a remote samba printer to Windows, using the local Windows device driver (one thing that Windows do handle well). My new HP 3830 works very well with Windows, and I am attempting to get it to work in raw mode on Devuan.
Devuan recognizes a new device /dev/usb/lp1 but attempts to print using "cat filename > /dev/usb/lp1" produce nothing. The file "filename" contains a PCL reset command, a short character string, and a formfeed. I must be missing something simple. If there is any sort of connection to the printer, something should appear, if only garbage. Suggestions please.
A follow-up for those who might be concerned about possible problems with transition; there are none. My second machine is now running Devuan ascii with the 4.9.0-9-amd kernel; it is the web-server, and any misgivings I might have had about apache2 were unfounded. The only interruption to the half hour process was apt-get dist-upgrade terminating "unable to fetch some archives." The suggested repetition of the command worked to completion, and I suspect a network or server glitch.
Faced with a Debian upgrade to 9.9, including a kernel upgrade to 4.9.0-9, I decided it was time to convert the first of my two "production" machines to Devuan. The whole process was easy and painless, and took no more time than the Debian upgrade would have. There were 72 upgraded packages, and 7 new ones to give a final total of 911. My concerns about possible changes to packages that require some tricky customized configuration were unfounded.
Everything works as before, and I am very impressed with the way the Devuan system is maintained and documented. My congratulations to the maintainers.
Having done a simple test on migration from Debian stretch to Devuan ascii on a virtual machine, I am planning for changing over two "production" desktop computers. Just as a matter of interest, since the migration process was straightforward, it would be useful to know what fraction of Debian packages are affected by the elimination of systemd.
My test was on a "bare bones" system with 438 packages, and only 23 packages listed by dpkg -l show with a Devuan identifier. Is the ratio going to be the same on machines running a more comprehensive selection of software?
Note that I am not using any kind of desktop environment; my only gui is to run an fvwm window manager using xterms, and drop-down menus for some task selection.
As a follow-up, checking for reports on this problem elsewhere, it seems the cause is that in the boot process file systems are mounted before the VirtualBox Guest Additions kernel module is loaded. This sequence of events is verified by examination of kern.log
Ironically systemd has a mechanism to deal with this, but it is much simpler to run a mount command on infrequent occasions where the shared folder is required.
In any case, the problem won't arise on a real machine; the virtual machine was used for testing the migration process.
This topic may belong with configuration, but it arises in migration from Debian stretch to Devuan. I am running Devuan in a VirtualBox machine set up on Windows 10. The virtual machine was cloned from a Debian machine, and converted using the migration procedure. After one minor problem with adapter names, everything now works as before, except for the automatic mounting of a shared folder with the host machine.
The shared folder is not mounted by the boot process on Devuan, but it can be mounted manually without any problem. In the original machine, before migration, the folder was mounted automatically as specified in the VirtualBox settings. One solution to the problem would be to include the mount in /etc/fstab but it should be handled by VirtualBox Guest Additions.
Is it possible that some systemd debris, that has not been cleaned out of the virtual machine, is getting in the way?
Thanks for the help; I should have known, as I had dealt with the opposite problem when Debian upgraded from jessie to stretch.
All I needed to do was edit the interface name in /etc/network/interfaces and restart networking (now easy to do with the familiar SystemV).
I must repeat again how easy the migration process was; Devuan is obviously the way to go, and things are now arranged back in a familiar way.
I am experimenting with migration from Debian stretch to Devuan on a VirtualBox machine running on a Windows 10 laptop. I have a Debian machine there, and made a clone in order to test the migration process without damaging a working system (the whole point of virtual machines). Everything works well until reboot after the final dist-upgrade; on rebooting there is no longer a network connection.
In passing I should say I am impressed with the ease of the migration process; my test machine has only 430 Debian packages, essentially X11 and compilers, plus the standard utilities, and the process only required 69 upgraded packages. That suggests to me that where there is no systemd contamination, Debian packages are essentially left intact.
The virtual machine network connection is simply through dhcp connected directly to the laptop via NAT. For some strange reason, the adapter is called enp0s3 instead of eth0 which one might expect. I have not done much exploration of the problem yet, and the only things that occur to me are (a) a different adapter name, or (b) no running dhcp client. I assumed that the migration process does upgrade/replace the dhcp client.
I did not install wicd as the existing /etc/network/interfaces file is precisely what is needed, assuming the adapter is identified correctly. With only one connection to the network, it doesn't seem that a network manager is required.
Suggestions for fixing this problem would be much appreciated.
Does anyone know where Devuan icon files can be found?
Ideally I would like 32x32 Windows ico format to use for a link to a virtual machine running Devuan.
Thanks.
The migration process looks no more difficult than the usual distribution upgrade process. It should also be easy to test on a virtual machine without running into trouble.
My VirtualBox runs on a Windows laptop, mainly because dual booting is not recommended, and Windows does at least handle drivers and wifi without trouble -- useful on the road to get into my home machines via ssh. All I use the laptop for is to run Linux in a virtual machine and use it as a remote terminal for my real computers.
I have done programming and worked with unix workstations as a physicist, most of my working life, and in retirement have been running Linux for several years on two or three machines. Many years ago I switched from RedHat to Debian because I did not like being told how I should do things. I am currently running the latest version of Debian on two desktop machines, and in a VirtualBox on a Windows laptop. I am getting disenchanted with the increasing complexity of Debian, having just had to deal with restarting inetd, and am interested on the possibility of moving to Devuan.
This post is primarily to introduce myself, but also to seek some reassurance that I could run my current set of some 800 packages on Devuan. They include compilers, apache2, samba and the usual utilities. As an old dog who does not want to learn new tricks unless I have to, I do not like desktop environments and graphical interfaces; my choice is to use fvwm and run xterms, plus x11 programs I have written, as I find entering commands usually is faster and works better than answering multiple questions on dropdown menus.
Probably the best way to familiarize myself with Devuan is to start another virtual machine.