11

Y'know they way windows print queues will occasionally stop working with a print job at the head of the queue which just won't print and which you can't delete? Anyone know whats going on when this happens?

I've been seeing this since the NT4 days and it still happens on 2008. I'm talking about standard IP connected laser printers - nothing fancy. I support a lot of servers and loads of workstations and see this happen a few times a year.

The user will call saying they can't print. When you examine the print queue, which in my case will generally be a server based queue shared out to the workstations, you find a print job which you cannot cancel. You also can't pause it, reinitialize it, nothing.

Stopping the spooler is the usual trick and works sometimes. However I occasionally see cases which even this doesn't cure and which a reboot is the only solution. Pause the queue, reboot, when it comes back up the job can then be deleted. Once gone the printer happily goes back to its normal state. No action is ever necessary on the printer.

I regard having to reboot as last resort and don't like it. What on earth can be going on when stopping the process (spooler) and restarting it doesn't clear a problem?

Its not linked to any manufacturer either. I've seen this on HPs, lexmark, canon, ricoh, on lasers, on plotters.... can't say I ever saw this on dot matrix.

Anyone got any ideas as to what may be going on.

Ian

Ian Murphy
  • 1,329
  • 4
  • 19
  • 29
  • 1
    With my print server it seems to be jobs sticking at 64.0kb spooled for no apparent reason. Doesn't happen regularly enough for us to nail it down though - it's well known printers are evil. EEEEVIIIIIILLLLLL. – tombull89 Oct 23 '12 at 09:59
  • 2
    Million dollar question right here – Dan Oct 23 '12 at 10:14
  • 1
    @tombull89 LOL. Whenever I hear the words can't print its like another nightmare begins. I think we've all lost hours of our lives to these infernal machines. – Ian Murphy Oct 25 '12 at 08:55
  • We used to schedule a bounce of the print daemon regularly but it didn't help. How many times can I +1 this question?! $1,000,000 question! – Lizz Jan 05 '13 at 06:43
  • 1
    In my experience, you can resolve this problem without rebooting by deleting the actual files from the spool directory while the spooler is stopped. I realize that doesn't answer your question, which is why I'm leaving it as a comment instead of an answer, but it seemed worth mentioning that you probably don't need to reboot. – davidcl Mar 03 '13 at 18:56
  • [Obligatory](http://blog.lib.umn.edu/elsie/elsiespeaks/2009/11/computer-dialogue.html). – Mike G Mar 04 '13 at 20:51
  • I have seen strange issues like this when a printer is set to half-duplex instead of full-duplex in it's network settings. –  Mar 04 '13 at 20:48
  • Buggy drivers. Cr@ppy printers. Stupid users. Duff client software. And feckless Level 1 / 2 support staff. How many times have us folks in the server team had to explain to people that a printer won't print if you can't ping it. FFS! Sorry, rant over. – Simon Catlin Mar 04 '13 at 22:15

1 Answers1

2

My experience is usually a buggy print driver. I can't tell you how many times I've had to stop the print spooler service on Windows systems, navigate to the print spooler directory (usually C:\WINDOWS\System32\spool\PRINTERS), and delete the spool files there, then restart the print spooler service.

In my experience, it seems to mostly always be related to a font issue, which is a nightmare in the world of scanned PDFs.

Sorry I don't have a fix. Just wanted to let you know I feel your pain.

churnd
  • 3,977
  • 5
  • 33
  • 41
  • 1
    Why are there any Adobe print products on a print server? – MDMarra Mar 04 '13 at 21:04
  • I was behind you until you said the about about Adobe. Buggy print drivers are my stock answer. – Evan Anderson Mar 04 '13 at 21:26
  • I meant the way adobe desktop products generate PDF's on client systems, which get sent to sprint servers. I agree it was poorly worded. – churnd Mar 04 '13 at 23:21
  • Can you be more descriptive? I don't know a whole lot about PDF creation. Are you saying that adobe file printers make malformed files that Windows print servers choke on? What's malformed about them and why hasn't adobe fixed it over the course of the last decade? – MDMarra Mar 05 '13 at 12:23
  • My experience mainly centered around people trying to print off scanned PDF's. They scan the PDF, optimize it and/or OCR it in Adobe, then later try to print it off for whatever reason. My memory is fuzzy, but my research at the time led me to believe the optimizing/OCR'ing of the PDF was imbedding certain glyphs or fonts that the print driver would choke on. I could never narrow it down to something precise & I haven't done it in a few years (thank God). The ultimate underlying issue was definitely the print driver, but just triggered most by Adobe PDFs. – churnd Mar 05 '13 at 13:45
  • @churnd That makes very little sense. A malformed PDF *may* cause some software to choke, but there is zero reason that OCR or "optimizing" (usually a fancy way of saying "converting images to lower resolution / color depth") should cause problems, provided the PDF being produced is still a *valid PDF*. If Adobe's software is not generating valid PDFs (or your print driver that supposedly handles them is not dealing with valid PDFs correctly) that's a whole new world of problems... – voretaq7 Mar 05 '13 at 19:10
  • I didn't say the aforementioned PDF's were not valid, just that they were choking the print drivers. The problem I dealt with was reproducible every time with scanned then optimized/OCR'd PDF's. If you didn't optimize/OCR the PDF, it didn't happen. Why? I don't know. My best guess is a buggy print driver coupled with some funky fonts/characters OCR inserted. You ever try to copy/paste some of the stuff OCR generates from a PDF to a rich text editor? It's not all letters & numbers. :) – churnd Mar 05 '13 at 20:58
  • 1
    @churnd I can confirm that OCR done by various Adobe products is downright evil. Some of the glyphs are malformed which causes issues in the Windows font-engine which is called by the printer-driver to render the text. As the driver (and thus also the font-engine) is run in the context of the print-spooler process this can cause all sorts of havoc. – Tonny Mar 05 '13 at 22:44
  • 2
    There are a shitload of bad Tier-2 printer-drivers around that are based on the sample source-code once published in the MS SDK for Windows 2000. The SDK code had a bug: It didn't release a gdi-handle after rendering and printing a page. As a result the print-spooler process would loose a handle for each page printed. After 10.000 handles (hard limit set in the NT kernel, even in Win2012/Win8) the print-spooler crashes. You will only notice this if you print more than 10.000 pages between reboots. – Tonny Mar 05 '13 at 22:54
  • @tonny, 10000 pages between reboots isn't a very high figure, so its surprising we don't see this more often. I have systems at clients which only reboot once or twice per year and they print plenty (accounts and billing depts). That could well be the cause - after all these years. The content of the print job didn't seem likely as the driver should be just treating it as binary data... usually. – Ian Murphy Mar 06 '13 at 13:40
  • @Ian That 10.000 pages issue only affects Tier-2 drivers with that particular bug. Most printer-manufacturers are doing it right these days, especially on x64 Windows because you can't use Tier-2 drivers anymore on those. If they need to build a new tier-3 driver for x64 anyway they usually also compile this as a x86 driver so Tier-2 drivers are a dying breed. I have seen the bug in older (before 2010) HP, Zebra, Minolta and Epson drivers. Current versions of those are OK, but there are a lot of print-servers out there still running old drivers. – Tonny Mar 06 '13 at 14:31
  • hmm, thats a shame - because I've seen it with x64 in the last couple of months. In this case I know the server is only running recent drivers because both server and printers are all post-2010 and I know all the drivers are pretty much the most recent versions available on that server. – Ian Murphy Mar 07 '13 at 14:45