View previous topic :: View next topic |
Author |
Message |
The_Great_Sephiroth Veteran
Joined: 03 Oct 2014 Posts: 1602 Location: Fayetteville, NC, USA
|
Posted: Tue Jan 23, 2018 3:38 pm Post subject: CUPS keeps pausing a printer... |
|
|
Got a strange problem. I have three HP OfficeJet Pro 8720 printers at the main office. They're plugged into a Gentoo print/scan server via USB. Scanning works flawlessly, but one of the three keeps going into a paused state, and refuses to resume. If I log into the box and resume it, it shows processing for a few seconds, then it pauses again. If I clear the queue and resume it it still pauses after a few seconds. The others pause occasionally for no reason, but a simple resume kicks them off again. Logs are not useful at this point as no errors or warnings have been logged. How do I begin to diagnose this issue? _________________ Ever picture systemd as what runs "The Borg"? |
|
Back to top |
|
|
LIsLinuxIsSogood Veteran
Joined: 13 Feb 2016 Posts: 1179
|
Posted: Wed Jan 24, 2018 7:49 am Post subject: |
|
|
Just to recap is the problem with just 1 out of 3 printers or with all 3 of them? If it is just the 1 then that might be an issue with how the printer was installed, and it may need to be reinstalled. Or else it might suggest a hardware issue with that printer. If it is all 3 of them then I would think that suggests either something with cups server configuration or else might also be a system dependency like udev that is doing something with the hardware to mess things up.
One thing that I personally don't like about CUPS (although it does seem to work pretty great for most everything) is its look and feel and especially with local printers, the device management pages can get sort of annoying to have to really look closely and see, ok is that a local port (yes or no)...I'm not saying you should change anything in terms of the hardware configuration, but in case you haven't yet attempted to connect the problem printer(s) via network cable cat5 or cat6 ethernet could be worth a try and may actually improve the flow to other local ports if that somehow lessens the load on the pci bus via USB ports.
Can you provide the information for package cups (emerge -pv cups) and emerge --info as well...Not sure if there are more than one boxes involved, but it didn't sound like it was more than one.
Also this could help....possibly which was the first hit on google for debugging cups...
https://support.apposite.com.hk/Knowledgebase/Article/View/478/0/how-to-enable-debug-in-cups
https://fedoraproject.org/wiki/How_to_debug_printing_problems |
|
Back to top |
|
|
The_Great_Sephiroth Veteran
Joined: 03 Oct 2014 Posts: 1602 Location: Fayetteville, NC, USA
|
Posted: Wed Jan 24, 2018 2:54 pm Post subject: |
|
|
They use the same driver since they are the same exact model. CUPS has paused them all at one point. Now it just isn't resuming one. The printer hardware is fine since we can copy and fax with it and even print to it directly from our phones. That leave the only re-installation at the USB cable, which we have checked, unplugged, and plugged in again already. The fact that udev works fine using the same driver with two identical printers plugged in the same way kind of rules it out in my mind, since the only difference is the serial number on them.
I don't know what you're talking about with lessening the load on a PCI bus because three USB printers and a PS/2 keyboard are all that's plugged in. This is a modern 64bit system with 4GB of RAM in it. If Linux cannot handle that I'm going back to DOS or Windows 9X. We're not spending the money to run Cat6 through the building to make these things run on Ethernet. Too much time and money would be involved.
Finally, of the links you sent the second isn't even valid. This is a print server. All of my servers are shell-only, so the Fedora article cannot work. You access my servers via SSH and then do things the old-school way. I also already stated that I have reviewed my logs and found nothing odd. The log level was already increased and is the first thing any seasoned user would do when having an issue that the default log level does not identify. I assumed that would be evident. Since it isn't, I am stating that I am already using debug output. My issue is that it isn't showing me an issue, so I came here for more advanced advice.
So to recap:
- Logs are already on debug level
- All three printers are identical and use the same driver
- Only one has stopped resuming at this time
- All three have paused at different times
- All three are USB directly into the PC using 6ft cables
- Cannot run more Ethernet to that room
- When WiFi is enabled, phones can print to the trouble printer without a hitch
This is where I am stuck. _________________ Ever picture systemd as what runs "The Borg"? |
|
Back to top |
|
|
The_Great_Sephiroth Veteran
Joined: 03 Oct 2014 Posts: 1602 Location: Fayetteville, NC, USA
|
Posted: Wed Jan 24, 2018 4:21 pm Post subject: |
|
|
OK, I believe I fixed it. After reverting to normal log level and restarting cups it FINALLY spewed an error about a cache file having an invalid value. I removed the file in question and restarted cupsd again, and all is good. Apparently there's a print job in that file with an invalid value. The only question I have remaining is why cupsd wasn't telling me that to begin with, especially with debug level logging enabled. _________________ Ever picture systemd as what runs "The Borg"? |
|
Back to top |
|
|
LIsLinuxIsSogood Veteran
Joined: 13 Feb 2016 Posts: 1179
|
Posted: Thu Jan 25, 2018 1:08 am Post subject: |
|
|
The Great Sephiroth, I did not mean to suggest you did not already have debugging turned on, sorry about making such a suggestion about that.
The fact that the error messages were not visible from within the server validates my point about the CUPS interface being sort of difficult and useless at times even.
Although I guess since you haven't posted the error message that led you to resolving the issue, and more details like when you restarted the service not just what was the message exactly but then how you proceeded, i.e. just deleting the file mentioned in the message or going into the file and removing some of it only...also did you have to restart cups afterwards for the change to take effect? |
|
Back to top |
|
|
The_Great_Sephiroth Veteran
Joined: 03 Oct 2014 Posts: 1602 Location: Fayetteville, NC, USA
|
Posted: Fri Jan 26, 2018 12:07 am Post subject: |
|
|
It logged a message about an invalid value on line <some number> in a job in file /var/cache/something. I stopped cupsd, removed the file, and started cupsd. All is good now. No clue how the value became invalid. The server is on BTRFS RAID1 so even if one disk was dying the odds of two disks having the exact same sector go bad is astronomically low. My guess is the HP driver did something to that one print job that CUPS hated. _________________ Ever picture systemd as what runs "The Borg"? |
|
Back to top |
|
|
LIsLinuxIsSogood Veteran
Joined: 13 Feb 2016 Posts: 1179
|
Posted: Sat Jan 27, 2018 6:26 am Post subject: |
|
|
Quote: | The server is on BTRFS RAID1 |
I would also assume that it is not disk related, because then you would be experiencing problems maybe in other applications in addition to this one. Not to mention what you describe seems like a bad variable or something being registered in the print job for cups. Has it happened again yet since removing the file? If so, you will need to share more info like emerge --info cups and some other stuff so that someone may help diagnose the issue.
I would suggest reinstalling cups server and set it up again. |
|
Back to top |
|
|
The_Great_Sephiroth Veteran
Joined: 03 Oct 2014 Posts: 1602 Location: Fayetteville, NC, USA
|
Posted: Sat Jan 27, 2018 3:52 pm Post subject: |
|
|
Considering the file contained an invalid value it could easily be a bad disk. If the file was stored on a failing sector it would result in a value being bad and thus the same issue. However, BTRFS does bit-rot protection so I was not worried about that. Also, as stated previously, the issue is resolved. It wasn't CUPS. It was an invalid value in a print job. The driver renders said jobs, in this case the HP driver. It may have done something that CUPS didn't like. It was only one print job so nothing was lost. _________________ Ever picture systemd as what runs "The Borg"? |
|
Back to top |
|
|
LIsLinuxIsSogood Veteran
Joined: 13 Feb 2016 Posts: 1179
|
Posted: Sun Jan 28, 2018 12:40 am Post subject: |
|
|
If solved, then please change the subject to mark it as solved. |
|
Back to top |
|
|
|