Arch Linux

Is this related to certain kind of printers (IPP, local USB,...)? What's your setup.

Please enable debug logging in cupsd.conf and check if some error msg pops up.

good day!

cupsd on my computer (package cups-2.0.0-2) -- begins use cpu 100% *always* when I go to "http://127.0.0.1:631/admin" and push button "Find New Printers".

then I must do "sudo systemctl restart org.cups.cupsd" to stop 100% cpu using.

my file "cupsd.conf" -- is https://gist.github.com/anonymous/d3b293d712ddf62b1d09

I tryed to view what the operation doing in infinite cycle (via using utility "strace")

[code]
$ sudo systemctl stop org.cups.cupsd org.cups.cupsd.path org.cups.cupsd.socket

$ sudo strace /usr/bin/cupsd -l

### go to "http://127.0.0.1:631/admin" and push button "Find New Printers" ###

... ... ...
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
epoll_wait(4, {{EPOLLHUP, {u32=2683298992, u64=140598437733552}}}, 4096, 1000) = 1
... ... ... (repeat infinite times) ... ... ...
[/code]

I disabled network (disabled WiFi card) on my computer. but no any difference.

no USB-printers connected on my computer.

I can confirm this after boot cupsd takes 100% CPU for like 1 min after that its looks normal

This is happening also when printer configuration is changed.

I have just made a test and I also have a never ending
"epoll_wait(4, {{EPOLLHUP, {u32=842727904, u64=140137035661792}}}, 4096, 1000) = 1"

Remember that there are not error messages in demsg, journal and cups log.
Deleting /etc/cups and /var/cache/cups and reinstalling cups wil not solve the problem.

Same problem here, also with no local printers but only network printers discovered by cups-browsed.service. I've set the LogLevel to debug in cupsd.conf. Attached are my /var/log/cups/{access_log,error_log,page_log} files.

Note that the "cupsd -l" process just became sane again. But I can trigger another period of 100% CPU usage by clicking the "Find new printers" button on the administration page on localhost:631.

I believe I had the same issue (cups locking up a cpu core when printing a document and never actually printing)

(gnome-control-center:12460): printers-cc-panel-WARNING **: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name org.fedoraproject.Config.Printing was not provided by any .service files
(gnome-control-center:12460): printers-cc-panel-WARNING **: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name org.freedesktop.PackageKit was not provided by any .service files
(gnome-control-center:12460): printers-cc-panel-WARNING **: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name org.fedoraproject.Config.Printing was provided by any .service files

Installed "system-config-printer" and I no longer get high cpu consumption and printing works (pretty sure I didnt have it installed for 1.7.5 and printing worked perfect)

Im having trouble adding the virtual pdf printer (works through web interface so propably a gnome bug)

I have system-config-printer installed but still get the high CPU consumption.

I can confirm that installing system-config-printer doesn't solve the issue. I have it installed but still get high CPU loads.

I had a couple of shared printers. I disabled sharing and changed the cups configuration file disabling Browsing (Browsing Off) but the issue remains. Every time I print and on boot CPU load goes to 100%.

Just to point out: I have gnome-manual-duplex installed from sourceforge and it causes some errors. In the error log I have several lines with "[cups-deviced] Bad line from "gmd":". Also there are warning lines like: "CreateProfile failed: org.freedesktop.ColorManager.AlreadyExists:profile id 'Canon_MP250_series-Gray..' already exists" for every printer I have installed.

I'm not sure if this is a bug in cups or in cups-filters because it seems to be triggered when browsing for printers.

Can you please tell us what printers are you using? All IPP or a certain company backend?

You can try to disable str4500.patch if this affects your bug.

If this all doesn't help look for known upstream bugs or file one and leave here the STR cups tracker ID.

These are the printers I have configured. They're all local USB printers, no network printers. Most of the time they are not physically connected to the laptop (some are in different physical spaces).
Canon MP250 series - CUPS+Gutenprint v5.2.10
GnomeManualDuplex Virtual Printer Generic CUPS-PDF Printer
HP Deskjet 1510 Series hpijs, 3.14.10 Idle
HP LaserJet Professional p1102, hpcups 3.14.10, requires proprietary plugin
Generic CUPS-PDF Printer

How would I disable str4500.patch?

You can use abs to rebuild the package and comment out that patch.

For me all printers are network printers discovered here at my university's network by cups-browsed. Most of them have a connection like socket://printername, some are printers shared by people's computer, e.g., ipps://SomeUsers-MacBook-Pro.local:631/printers/Brother_DCP_J315W.

Ok, AndyRTR, I've now built and installed cups-2.0.0-3 without the str4500.patch but that doesn't make a difference. cupsd -l still takes 100% CPU for some minutes after startup and then again when pressing the "Find new printers" button (and probably also when other actions are performed).

Ok, there hasn't been an existing upstream bug report so I filed this new one: https://www.cups.org/str.php?L4521+P-1+S-2+C0+I0+E0+Q

Someone affected should file an upstream bug with debug logs: https://cups.org/str.php

Andreas, that's what I did, no? Do you say that I should attach my logs also at the upstream bug in addition to linking to the logs of this report?

I also experience problems after the CUPS update. First, the full load of one CPU core by /usr/bin/cupsd -l whenever trying to add a printer through system-config-printer, as described in this bug. The corresponding strace looks differently, though (see attachment). Second, system-config-printer crashes after entering root authentication, so I cannot add printer through system-config-printer.

after install lastest version of cups: cups 2.0.1-1 and libcups 2.0.1-1
after login to system proccess /usr/bin/cupsd -l take 100% one of core CPU.
after restart cups service (sudo systemctl restart org.cups.cupsd) all is fine.

I confirm the problem as aforementioned by @Tomasz

I can confirm this behaviour: If I switch on the usb-connected printer, cpu load on one core raises and stays on 100% for exactly 5 minutes. Then the load return to normal, regardless if the printer is idle or printing. After switching off the printer load raises again on one cpu core for 5 minutes, then back to normal.

Everybody affected should help providing info to the upstream report: https://www.cups.org/str.php?L4521 !

CUPS takes 100% of one core everytime I power my printer or scanner on. How can I fix it? I'm using CUPS 2.0.1. Please help!

hi, my problem was solved with modified client.conf file(/etc/cups/client.conf)
cups looks for „ServerName /run/cups/cups.sock“
i change the line to „ServerName localhost:631“

edit: i remove „/usr/share/cups/model/foomatic-db-ppds/Kyocera/ReadMe.htm“ before i change client.conf,
maybe its both necessary.

ServerName localhost:631 doesn't work for me

@ Antonio: Did You restart Your computer? It works fine for me, but after an restart. :)

@ Peter Mayr: You're great, thx man! :)

Yes, i rebooted

I tried the solution suggested by Peter Mayr and it seems it is partially effective. I mean, when I rebooted, Cups did not generate high CPU load at all but when I click "Find new printers" in Cups admin, the problem returns.

I tried removing the "Listen /run/cups/cups.sock" line from the Cups configuration, but this doesn't seem to stop Cups generating high CPU load.

I don't have any high load on boot. I only had it when powering on my scanner, or printer. Now, after changing the line it's gone. It worked for me, I can't say that'll work for everyone.

I have modified the line in client.conf to „ServerName localhost:631“. Results:

1) High cpu on reboot: fixed
2) High cpu on printing: fixed
3) High cpu on changing configuration at localhost:631 (for example printer defaults): NOT fixed

If it is any help I have also been getting 100% CPU load for a minute or two when printing - this evening I switched on the usb printer and printed a page - relevant dmesg output was as follows:

[110948.082470] usb 3-4: new high-speed USB device number 4 using xhci_hcd
[110948.256955] usb 3-4: ep 0x3 - rounding interval to 8 microframes, ep desc says 10 microframes
[110948.256962] usb 3-4: ep 0x84 - rounding interval to 8 microframes, ep desc says 10 microframes
[110948.256966] usb 3-4: ep 0x1 - rounding interval to 8 microframes, ep desc says 10 microframes
[110948.256970] usb 3-4: ep 0x82 - rounding interval to 8 microframes, ep desc says 10 microframes
[110948.263876] WARNING! power/level is deprecated; use power/control instead
[110948.270885] usb 3-4: ep 0x1 - rounding interval to 8 microframes, ep desc says 10 microframes
[110948.270895] usb 3-4: ep 0x82 - rounding interval to 8 microframes, ep desc says 10 microframes
[110948.271488] usblp 3-4:1.1: usblp0: USB Bidirectional printer dev 4 if 1 alt 0 proto 2 vid 0x04E8 pid 0x342B
[110948.271525] usbcore: registered new interface driver usblp
[110949.437654] usblp0: removed
[110949.438829] usb 3-4: ep 0x1 - rounding interval to 8 microframes, ep desc says 10 microframes
[110949.438837] usb 3-4: ep 0x82 - rounding interval to 8 microframes, ep desc says 10 microframes
[110949.440316] usblp 3-4:1.1: usblp0: USB Bidirectional printer dev 4 if 1 alt 0 proto 2 vid 0x04E8 pid 0x342B
[111009.819386] usblp0: removed
[111018.403335] usb 3-4: ep 0x1 - rounding interval to 8 microframes, ep desc says 10 microframes
[111018.403342] usb 3-4: ep 0x82 - rounding interval to 8 microframes, ep desc says 10 microframes
[111018.403744] usblp 3-4:1.1: usblp0: USB Bidirectional printer dev 4 if 1 alt 0 proto 2 vid 0x04E8 pid 0x342B
[111018.403812] usblp0: removed
[111018.561760] usb 3-4: reset high-speed USB device number 4 using xhci_hcd
[111018.561780] xhci_hcd 0000:00:14.0: Setup ERROR: setup context command for slot 3.
[111018.561784] usb 3-4: hub failed to enable device, error -22
[111018.721809] usb 3-4: reset high-speed USB device number 4 using xhci_hcd
[111018.721831] xhci_hcd 0000:00:14.0: Setup ERROR: setup context command for slot 3.
[111018.721835] usb 3-4: hub failed to enable device, error -22
[111018.882033] usb 3-4: reset high-speed USB device number 4 using xhci_hcd
[111018.895677] xhci_hcd 0000:00:14.0: xHCI xhci_drop_endpoint called with disabled ep ffff8800d5735d80
[111018.895681] xhci_hcd 0000:00:14.0: xHCI xhci_drop_endpoint called with disabled ep ffff8800d5735dc8
[111018.895684] xhci_hcd 0000:00:14.0: xHCI xhci_drop_endpoint called with disabled ep ffff8800d57353c0
[111018.895686] xhci_hcd 0000:00:14.0: xHCI xhci_drop_endpoint called with disabled ep ffff8800d5735408
[111018.895693] usb 3-4: ep 0x3 - rounding interval to 8 microframes, ep desc says 10 microframes
[111018.895696] usb 3-4: ep 0x84 - rounding interval to 8 microframes, ep desc says 10 microframes
[111018.895700] usb 3-4: ep 0x1 - rounding interval to 8 microframes, ep desc says 10 microframes
[111018.895704] usb 3-4: ep 0x82 - rounding interval to 8 microframes, ep desc says 10 microframes
[111018.895828] usb 3-4: ep 0x1 - rounding interval to 8 microframes, ep desc says 10 microframes
[111018.895832] usb 3-4: ep 0x82 - rounding interval to 8 microframes, ep desc says 10 microframes
[111018.896180] usblp 3-4:1.1: usblp0: USB Bidirectional printer dev 4 if 1 alt 0 proto 2 vid 0x04E8 pid 0x342B

The print was completed OK but the CPU on one of the 4 cores was pegged at 100% for around a minute or two. I have not tried to change the line in /etc/cups/client.conf as in some of the comments. Any chance this is related to problems with multifunction printers connected to usb3 ports when scanning?

In my case I have the socket enabled instead of the service:
$ sudo systemctl status org.cups.cupsd.service
● org.cups.cupsd.service - CUPS Scheduler
Loaded: loaded (/usr/lib/systemd/system/org.cups.cupsd.service; disabled; vendor preset: disabled)
Active: active (running) since Mon 2014-12-22 23:06:43 GMT; 2min 19s ago
Docs: man:cupsd(8)
Main PID: 375 (cupsd)
CGroup: /system.slice/org.cups.cupsd.service
└─375 /usr/bin/cupsd -l

Dec 22 23:06:43 home1 systemd[1]: Started CUPS Scheduler.
$ sudo systemctl status org.cups.cupsd.socket
● org.cups.cupsd.socket - CUPS Scheduler
Loaded: loaded (/usr/lib/systemd/system/org.cups.cupsd.socket; enabled; vendor preset: disabled)
Active: active (running) since Mon 2014-12-22 23:06:42 GMT; 2min 23s ago
Listen: /run/cups/cups.sock (Stream)

Dec 22 23:06:42 home1 systemd[1]: Starting CUPS Scheduler.
Dec 22 23:06:42 home1 systemd[1]: Listening on CUPS Scheduler.

When I changed the /etc/cups/client.conf as suggested above and rebooted then there was a systemd service failure. On returning the config file to the original and rebooting there was 100% CPU load on one processor for a minute or so and then on another CPU for a minute or so before it settled down. On turning off the printer I had another bout of CPU load pegged at 100%. Maybe this needs an upstream fix?

I thought I would try to see if there was any problem with the configs that had been setup for my printer prior to cups version 2.0. So with the printer off I deleted the printer from the KDE settings. Initially it refused to delete the printer but then allowed it. I then switched on the printer and the system set up a new printer which then allowed me to configure it for sharing. The error log showed:

$ cat /var/log/cups/error_log
E [22/Dec/2014:22:47:17 +0000] Missing value on line 1210.
E [22/Dec/2014:22:47:17 +0000] Missing value on line 2350.
E [22/Dec/2014:22:47:17 +0000] Missing value on line 2362.
E [22/Dec/2014:22:47:17 +0000] Missing value on line 2374.
E [22/Dec/2014:23:06:43 +0000] Missing value on line 1210.
E [22/Dec/2014:23:06:43 +0000] Missing value on line 2350.
E [22/Dec/2014:23:06:43 +0000] Missing value on line 2362.
E [22/Dec/2014:23:06:43 +0000] Missing value on line 2374.
E [23/Dec/2014:09:58:49 +0000] Missing value on line 1210.
E [23/Dec/2014:09:58:49 +0000] Missing value on line 2350.
E [23/Dec/2014:09:58:49 +0000] Missing value on line 2362.
E [23/Dec/2014:09:58:49 +0000] Missing value on line 2374.
E [23/Dec/2014:10:21:40 +0000] [Client 11] Returning HTTP Forbidden for CUPS-Delete-Printer (ipp://mike@localhost:631/printers/SCX-4500W-Series) from localhost
E [23/Dec/2014:10:22:32 +0000] [cups-driverd] Bad driver information file "/usr/share/cups/model/foomatic-db-ppds/Kyocera/ReadMe.htm"!
W [23/Dec/2014:10:22:36 +0000] CreateProfile failed: org.freedesktop.ColorManager.AlreadyExists:profile id 'SCX-4500W-Series-Gray..' already exists
W [23/Dec/2014:10:22:49 +0000] CreateProfile failed: org.freedesktop.ColorManager.AlreadyExists:profile id 'SCX-4500W-Series-Gray..' already exists

I then printed a mail from Thunderbird which behaved normally, with the print emerging within a few seconds, and no high CPU load. Switching off the printer then gave the CPU pegged at 100% for some minutes again until I used systemctl stop org.cups.cupsd. I also noted from the error log that there was a Bad driver information file, so I deleted it even though I have no Kyocera printer on my system, and never have!

Is there any other diagnostics that I can generate that might be helpful in diagnosing where this bug originates? My dmesg output from the point where I switched off the printer in the sequence above is:
[40255.411712] usb 3-4: USB disconnect, device number 5
[40255.411988] usblp0: removed
[40492.352365] usb 3-4: new high-speed USB device number 7 using xhci_hcd
[40492.526960] usb 3-4: ep 0x3 - rounding interval to 8 microframes, ep desc says 10 microframes
[40492.526966] usb 3-4: ep 0x84 - rounding interval to 8 microframes, ep desc says 10 microframes
[40492.526970] usb 3-4: ep 0x1 - rounding interval to 8 microframes, ep desc says 10 microframes
[40492.526973] usb 3-4: ep 0x82 - rounding interval to 8 microframes, ep desc says 10 microframes
[40492.527230] usb 3-4: ep 0x1 - rounding interval to 8 microframes, ep desc says 10 microframes
[40492.527235] usb 3-4: ep 0x82 - rounding interval to 8 microframes, ep desc says 10 microframes
[40492.527780] usblp 3-4:1.1: usblp0: USB Bidirectional printer dev 7 if 1 alt 0 proto 2 vid 0x04E8 pid 0x342B
[40493.615901] usblp0: removed
[40493.616703] usb 3-4: ep 0x1 - rounding interval to 8 microframes, ep desc says 10 microframes
[40493.616711] usb 3-4: ep 0x82 - rounding interval to 8 microframes, ep desc says 10 microframes
[40493.619412] usblp 3-4:1.1: usblp0: USB Bidirectional printer dev 7 if 1 alt 0 proto 2 vid 0x04E8 pid 0x342B

After then switching off the printer again the CPU also pegged at 100% again for several minutes swapping between CPUs that showed 100%. After waiting a couple of minutes with no sign of the CPU high load terminating I stopped the cups service in order to prevent the CPU load remaining at 100%. At this point an additional line appeared in dmesg:

[43442.945126] perf interrupt took too long (2509 > 2495), lowering kernel.perf_event_max_sample_rate to 50100

Since this bug has been around ever since the update to cups version 2.0 through to the current 2.0.1-1 it would be nice to get this sorted out.

I'm having the same problem and the only remedy I have until now is to uninstall gutenprint.
Only then can I get to the printer driver list and install a driver, which is slow because one core still remains at 100%. With Gutenprint installed, both cores are at 100% and even after half an hour nothing happens.
I still have to restart the org.cups.cupsd.service after the installation because one CPU core stays at 100% and only drops to normal after a restart.
This should only be a temporary solution because now I have to use the HP driver for my old HPLaserjet6L, which gives me a poorer quality and accuracy than the Gutenprint driver.

I posted some additional information to the thread at https://bbs.archlinux.org/viewtopic.php?pid=1490124#p1490124

Whether this helps since this points to possible involvement of socket activation I am unsure but if anyone can explore that possibility it might narrow down where the main issue is?

Please check with latest cups and cups-filters updates.

I still get 100% CPU load on 1 core (process cupsd) every time I add or modify a printer using the CUPS admin interface (http://localhost:631/).

No high load problems when printing, though.

I solved it on my way by disconnecting the printer from usb port and attached it to the network.

Please check cups 2.0.2-2 in testing. It should solve some high load issues.

Please check 2.0.2-3 - fixes the cpu load when searching for new printers for me.

It looks like the issue has been fixed in 2.0.2-3.