FS#26704 - [cups] Should be restarted after suspend

Attached to Project: Arch Linux
Opened by Olivier Mehani (shtrom) - Monday, 31 October 2011, 23:39 GMT
Last edited by Andreas Radke (AndyRTR) - Monday, 16 April 2012, 14:56 GMT
Task Type Bug Report
Category Packages: Extra
Status Closed
Assigned To Andreas Radke (AndyRTR)
Architecture All
Severity Low
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

Description:

After supending the machine (or, actually, changing network attachment), cupsd is a bit lost and can't print anymore until it is restarted.

I'm experimenting with the following in /etc/pm/sleep.d/90cups, but there may be a more appropriate hook associated to the network interfaces.
#!/bin/sh
#
# 90cups: suspend/wakeup CUPS

case "$1" in
hibernate|suspend)
;;
thaw|resume)
rc.d restart cupsd
;;
*) exit $NA
;;
esac


Additional info:
* cups-1.5.0-1


Steps to reproduce:
* boot machine
* print something
* it works
* suspend
* resume
* print something
* cups queues the job, but doesn't print them
This task depends upon

Closed by  Andreas Radke (AndyRTR)
Monday, 16 April 2012, 14:56 GMT
Reason for closing:  None
Comment by Olivier Mehani (shtrom) - Wednesday, 08 February 2012, 16:22 GMT
  • Field changed: Percent Complete (100% → 0%)
Ok, that was a dirty solution. I started investigating, and openned a bug with CUPS at [0]. They reckon code to avoid this problem should be in since 1.4.0, but it seems to occur on a fresh Arch install with 1.5.0.

[0] http://www.cups.org/str.php?L4012
Comment by Olivier Mehani (shtrom) - Wednesday, 15 February 2012, 05:02 GMT
I have intermittent connectivity today (due to DHCP leases exhaustion which has nothing to do with ArchLinux), and notice the problem everytime I have lost my lease, and eventually reconnect: no printers available until I restart cupsd.
Comment by Andreas Radke (AndyRTR) - Saturday, 03 March 2012, 15:27 GMT
This seems to be no cups bug. Can you track it down to our initscripts or dhcp server? I need to know what packages and maintainers to assign this bug.
Comment by Olivier Mehani (shtrom) - Thursday, 08 March 2012, 04:30 GMT
Sure. It seems to survive multiple calls to `netcfg -r eth0'. Any hint about how to proceed further, though? Is there a debug mode for inistcripts or dhcp (client, by the way, not server) that I could activate?
Comment by Olivier Mehani (shtrom) - Sunday, 15 April 2012, 22:58 GMT
Hum, with cups-1.5.2-1 and netcfg-2.6.8-1, the problem seems to have disappeared now. I'd say the bug can be closed, and hopefully never opened again.

Loading...