Scanning hangs on HP-LaserJet-MFP-M28-M31

Bug #1815794 reported by Daniel Șerbănescu
16
This bug affects 3 people
Affects Status Importance Assigned to Milestone
HPLIP
New
Undecided
Unassigned

Bug Description

Hello,
I have a debian stable system (9.7) and I tried the 3.18 hplip driver and the latest 3.19 with the binary plugin for scanning.

Whenever I try to scan something either via USB or network the scanner starts and move its reading head a bit and afterwards stops and abandons the scanning process.

I have captured the output I found in /var/log/syslog (grepped the simple-scan output).
I have tried different distros (debian stable, testing, unstable, arch, manjaro) and scanning fails over all of them. No matter which scanning software I use it fails.

Can you help in figuring out what happens? I'll provide more details if needed. Just say how to get them.
Hope we get this issue sorted out.
Thanks
Daniel

Revision history for this message
Daniel Șerbănescu (dasj19) wrote :
Revision history for this message
Jarno (yannick-jarno) wrote :

Hi Daniel,

I open an issue https://answers.launchpad.net/hplip/+question/677937 on january 19th and i still not have any answer.
As you I try to work with an MFP M28a and hplip 3.19.1 ( and plugin for scanner)
Printer work very well but scanner starts to move and i get an Error during device I/O (code=9)
With HP tool hp-scan i get an /hp-scan: http.c 322: invalid http_status=503 in syslog.
Could you verify that you have same problem ? I would feel less alone :)
I tested on physical machine ubuntu, virtual debian, virtual opensuse , same problem .
regards
Yannick

Revision history for this message
brian_p (claremont102) wrote :

> io/hpmud/model.c 532: no laserjet_mfp_m28-m31 attributes found
> in /usr/share/hplip/data/models/models.dat

Find

  [hp_laserjet_mfp_m28-m31]

in models.dat and change it to

  [laserjet_mfp_m28-m31]

The message

How do you go on?

--
Brian.

Revision history for this message
brian_p (claremont102) wrote :

> How do you go on?

A success/failure report would be good to have.

--
Brian.

Revision history for this message
Dylan Schwilk (firepulaski) wrote :

I have the same exact problem with hplip 3.19 and an mfp m28-m31. hp-check shows everything as ok, scanner is seen, scanning begins and the scanner head/light moves, but process ends with same error. Printing works. I'm on kubuntu 18.04

Revision history for this message
brian_p (claremont102) wrote :

> I have the same exact problem with hplip 3.19 and an
> mfp m28-m31. hp-check shows everything as ok, scanner
> is seen, scanning begins and the scanner head/light
> moves, but process ends with same error. Printing
> works. I'm on kubuntu 18.04

And yet - the workaround given wasn't tried or the 'journalctl' output examined.

What is the point of having the same problem and not reporting on anything tried.

Printing working is totally irrelevant.

--
Brian.

Revision history for this message
brian_p (claremont102) wrote :

> Can you help in figuring out what happens?

Yes.

> I'll provide more details if needed.

Didn't quite work out that way, did it?

--
Brian.

Revision history for this message
Daniel Șerbănescu (dasj19) wrote :

Hi, sorry for the delay... My old launchpad account was not configured to my new email.. and did not see your responses until now.

So in Jano's response:
Running hp-scan results in :
error: SANE: Error during device I/O (code=9)
Looking into journalctl I can indeed see:
/hp-scan: http.c 322: invalid http_status=503

Revision history for this message
Daniel Șerbănescu (dasj19) wrote :

Now trying brian_p's change... I noticed there are two entries labeled "[hp_laserjet_mfp_m28-m31]" in /usr/share/hplip/data/models/models.dat ... I changed both to [laserjet_mfp_m28-m31]...

And could not scan again... I had to re-add the printer using Gnome's Control Center.

Then the same error happens again error: SANE: Error during device I/O (code=9)

I'll add a dump of /var/log/syslog with the next message

Revision history for this message
Daniel Șerbănescu (dasj19) wrote :

And added relevant journalctl.log after I applied the changes to the models.dat file.

Revision history for this message
brian_p (claremont102) wrote :

Apologies for being a bit grumpy, Daniel; thanks for the explanation and the second log. This log does not show the first two lines that are in your first log, so, to some extent, my suggestion worked.

Both the entries in models.dat have

plugin=0
plugin-reason=0

which implies no plugin is required.

I would try moving /usr/share/hplip/scan/plugins out of the way and run 'scanimage > image.pnm' as a test. (Check the scanner can be seen with 'scanimage -L).

--
Brian.

Revision history for this message
Daniel Șerbănescu (dasj19) wrote :

Hi brian_p thanks for your quick response...

From what i get now from the logs it seems that the plugins are required.

I moved /usr/share/hplip/scan/plugins to /usr/share/hplip/scan/plugins_old and tried a scan as you said.

Revision history for this message
brian_p (claremont102) wrote :

Hi Daniel,

I really have no solution for you; this is one for the HPLIP developers. They are the only ones with full knowledge of plugin operation. However, I eventually realised I had dealt with the issue regarding your MFP-M28-M31 a couple of times before:

https://bugs.launchpad.net/hplip/+bug/1813445
https://bugs.launchpad.net/hplip/+bug/1811504

In #1811504 I did suggest an alternative scanning method using scan.py. Maybe it might suit you. If you do try it (it doesn't take a long time and is quite easy), I would appreciate knowing the outcome.

More problem descriptions from other users:

https://answers.launchpad.net/hplip/+question/675574
https://answers.launchpad.net/hplip/+question/677268

With you, Jarno and Dylan that is seven affected people. As James Bond said - once is happenstance; twice is coincidence; three times is enemy action.

--
Brian.

Revision history for this message
Daniel Șerbănescu (dasj19) wrote :

Thanks a lot for your input brian_p.

I just tried this python.py (from this repo: https://github.com/kno10/python-scan-eSCL) and this is what happens.

The reading head of the scanner does complete its trip from start to finish however the scan.py script fails right after the reading head starts to read. Hence I do not get the result of the scan and I can not use the script again before I restart my printer.

I will investigate this into more detail when I get more time at hand.

Thanks again for your valuable inputs, I hope this issue gets fixed with the releases of new versions of hplip ... or I might start looking at the code that fails myself if time allows... I do not have any python expertise but I am a web developer and know my way around programming...

Anyway there seems to be some progress with this scan.py script, at least it looks like it initiates a full scan rather than hplip which quits after 1 centimeter...

Revision history for this message
Thomas Moreau (t-moreau) wrote :

Hi,
I bought a HP LaserJet Pro MFP M28w about a week ago and I'm also affected by this bug.

Here is what the documentation from HP says
Model Name : HP LaserJet Pro MFP M28w
Min. HPLIP Version : 3.18.4
Driver Plug-in : No
Support Level : Full
Print Mode : Mono
Scan to PC : Yes
PC Send Fax : No
Connectivity : USB,Network
https://developers.hp.com/hp-linux-imaging-and-printing/supported_devices/index

Ubuntu MATE 16.04 LTS (32 bit) is shipped with HPLIP 3.16.3 so I downloaded and installed the latest version (3.19.1) from HP
https://developers.hp.com/hp-linux-imaging-and-printing/gethplip

I was able to set up the printer, configure the WiFi through a temporary USB connection, print and copy. But I couldn't scan...

First of all, despite what the documentation above says, I've been asked to install a plug-in.
Authentication and validation of the plug-in fails. And even if I still want to install it, the installation fails.
I had to download and install manually from
https://www.openprinting.org/download/printdriver/auxfiles/HP/plugins/
This plug-in installation bug has been reported at least twice :
https://bugs.launchpad.net/hplip/+bug/1749690
https://bugs.launchpad.net/hplip/+bug/1749573

I can see and hear the scanner's sensor moving left to right but I get the "Error during device I/O (code=9)".

I installed in dual boot a fresh Ubuntu MATE 18.04 LTS (32 bit) that is shipped with HPLIP 3.17.10. Had the same problems after installing HPLIP 3.19.1.

I noticed that people reported that scanning was to slow on the same model
https://bugs.launchpad.net/hplip/+bug/1802381
https://bugs.launchpad.net/hplip/+bug/1809287
So I had some hope for scanning to work, albeit not as fast as one would expect... I installed HPLIP 3.18.10 and its matching plug-in on Ubuntu MATE 18.04 LTS (32 bit). Not a fresh install anymore though, as I installed HPLIP 3.18.10 after HPLIP 3.19.1. Same problem.

I removed the "hp_" prefix as suggested, with the following command
sudo sed -i 's/hp_laserjet_mfp_m28-m31/laserjet_mfp_m28-m31/g' /usr/share/hplip/data/models/models.dat
to no avail.

I wanted to try Ubuntu MATE 18.10 that is shipped with HPLIP 3.18.7. That would be the first system with the minimum HPLIP version requirement met by default (>=3.18.4). Unfortunately, 32 bit is not supported anymore by Ubuntu. Not only there is no i386 ISO available, but the system prevents me from upgrading from 18.04 (at least in the GUI, I don't think I tried with the CLI). That might be worth trying by someone with a 64 bit computer.

As a workaround, I use the modified version of scan.py by Lucazade. Works for me.

There was a feature called "webscan" on my now dead HP Photosmart 5524 that allowed me to scan from the printer's embedded web server (EWS). I haven't found a similar feature on my HP LaserJet Pro MFP M28w.

How do we get HP to acknowledge the problem and lift a finger ? Will having enough people hitting "affectsmetoo" get HP's attention ?

Full Linux support, as claimed in the documentation, is one of the main factors when purchasing a new printer.

Last but not least, I take my hat off to brian_p. Thank you !

Revision history for this message
brian_p (claremont102) wrote :

Hello Daniel and Thomas,

Daniel wrote

> How do we get HP to acknowledge the problem and lift a
> finger ? Will having enough people hitting "affectsmetoo"
> get HP's attention ?

Upstream's response has now been made in a number of reports on this tracker. It is a standard message and you can view it in Message #9 at

https://bugs.launchpad.net/hplip/+bug/1821932

Daniel also said (referring to webscan)

> I haven't found a similar feature on my HP LaserJet Pro MFP M28w.

So - does the advice given work for either of you?

--
Brian.

Revision history for this message
Ben Bipod (benbipod) wrote :

Consolidating multile reports about scan problems with LaserJet Pro MFP M28-M31 to bug #1811504 as it seems to be the most advanced and even has a workaround.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.