Comment 30 for bug 341827

Having lasted for a month or so since the last printer hang, and in particular none since I manually upgraded to hplip-3.9.8 backported from F11, we had an incident yesterday which required a CUPS restart.

The incident timeline started at around 19:00 BST yesterday. Print Jobs had been submitted, but nothing seemed to be happening.

lpstat showed that the printer had been in a disabled state since Sep 12th, though no printing had been attempted at that time.

My first action was simply to "cupsenable" the printer, which initially appeared to start the queued print Jobs flowing again.

However, after printing the first page of the first queued job, the printer apparently hung. Looking at the syslog at that point, I noted that I could see the familiar "did not claim interface 1 before use" message repeating every 10 seconds.

I therrefore restarted cups using my /usr/local/bin/cupsrestart script, which duly dumped an lpstat/ps listing to Email as attached, before restarting CUPS. Once CUPS restarted, the outstanding print Job all printed successfully to completion.

The debug/status messages which I managed to capture at that time are attached.

Sadly, I failed to note down the exact timestamps of performing the cupsenable, nor when exactly the printer had come up in disabled state. Judging from the logs, the cupsenable action corresponds to this message:

...
Sep 15 19:31:06 topaz Photosmart_2570_series?serial=MY65T112BM04DY: io/hpmud/hpmud.c 348: [15869] hpmud_init()
...

Seemingly things started to go wrong just after this message:

...
Sep 15 19:31:38 topaz kernel: usb 1-7: reset high speed USB device using ehci_hcd and address 3
Sep 15 19:31:38 topaz kernel: usblp0: USB Bidirectional printer dev 3 if 1 alt 0 proto 2 vid 0x03F0 pid 0x4E11
...

And the cupsrestart was finally performed here:

...
Sep 15 19:35:37 topaz sudo: anne : TTY=pts/1 ; PWD=/home/anne ; USER=root ; COMMAND=/usr/local/bin/cupsrestart
...

Remember, as ever, that the copy of hplip I have running is a backported copy of 3.9.8 from F11 with additional debugging enabled via "#define HPMUD_DEBUG 1" in io/hpmud/hpmudi.h.

Last nights incident was slightly more unusual in that the printer subsystem had already ended up in a disabled state, but the behaviour of the system after I had invoked cupsenable seems to be identical to previous incidents.