--- Begin Message ---
- To: Debian Bug Tracking System <submit@bugs.debian.org>
- Subject: cups: Millions of symlinks to .ppd file created in /tmp
- From: Carl Worth <cworth@debian.org>
- Date: Sun, 30 Nov 2014 12:03:17 -0800
- Message-id: <8761dwo3qy.fsf@yoom.home.cworth.org>
Package: cups
Version: 1.7.5-5
Severity: critical
Justification: breaks the whole system
A few days ago, I was using my laptop at the house of a family member,
and needed to print something. I clicked through the add printer
dialogs (in Gnome) and the printer was detected, added, and my
document printed just fine. [And let me interrupt this bug report to
congratulate everyone involved. This operation used to be quite
painful in the past---it's great that it can go so smoothly now.]
Then, yesterday I noticed that my laptop was acting in an
uncharacteristically sluggish way. There was a strange delay of
several seconds between launching an application and its window first
appearing. It was clear that there wasn't any excessive CPU
utilization, so this felf IO related, but it wasn't obvious to me what
was going on.
Eventually, this morning I decided to see if a reboot would bring my
system back to normal. It didn't.
Instead, the system failed to boot. I was greeted with a boot message
saying:
a start job is running for Create Volatile files and directories
This message was accompanied by a timer "[ 10s / no limit ]" and the
timer just continued ticking off time. I let it sit for 2 or 3 minutes
before giving up and deciding it was never going to boot.
I duckduckgo'ed the message and found the following Debian forum
threads from last month:
http://forums.debian.net/viewtopic.php?f=10&t=118008
I was able to at least boot to a shell by editing my "linux" option
within grub, changing "ro" to "rw init=/bin/bash".
At this point I could investigate /tmp, where I found over 6 million
symlinks of apparently automatically-generated hexadecimal digits, all
linking to the same .ppd file. The target of these symlinks had the
same name as the printer model which I had added a few days ago.
I successfully got my system to reboot again after deleting all of the
symlinks:
cd /tmp
find . -name '5477*' | xargs rm
On the next reboot the "a start job is running for Create Volatile
files and directories" did pause and count up for a few seconds,
(perhaps 4 or 5), but then proceeded to boot. I went into the printer
settings and deleted the printer, (I'm no longer at that
family-member's house and shouldn't need to print to that queue again
any time soon).
I've also verified that /tmp/ now only contains a single symlink to a
.ppd file, (corresponding to the one remaining print queue that I do
use on a regular basis).
So I still don't know much about what process went wild and created
millions of symlinks, but it's definitely a bug that can result in
some pretty painful side effects.
I looked through the bugs in cups and didn't see any talking about
"symlinks" or "/tmp" so I hope this isn't a duplicate bug entry.
If the bug is known to be fixed in subsequent versions, that will be
great to know. If not, if there is anything I can do to debug further,
I will be happy to do so.
(If I had noticed the errant process creating the many symlinks while
my systerm was still usable, then I certainly could have done some
more investigation.)
Please let me know what further information I might be able to provide
that would be useful.
Thanks,
-Carl
-- System Information:
Debian Release: jessie/sid
APT prefers unstable
APT policy: (500, 'unstable'), (500, 'oldstable'), (1, 'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386
Kernel: Linux 3.16-3-amd64 (SMP w/4 CPU cores)
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Versions of packages cups depends on:
ii cups-client 1.7.5-5
ii cups-common 1.7.5-5
ii cups-core-drivers 1.7.5-5
ii cups-daemon 1.7.5-5
ii cups-filters 1.0.61-2
ii cups-ppdc 1.7.5-5
ii cups-server-common 1.7.5-5
ii debconf [debconf-2.0] 1.5.53
ii ghostscript 9.06~dfsg-1.1+b1
ii libavahi-client3 0.6.31-4
ii libavahi-common3 0.6.31-4
ii libc-bin 2.19-11
ii libc6 2.19-11
ii libcups2 1.7.5-5
ii libcupscgi1 1.7.5-5
ii libcupsimage2 1.7.5-5
ii libcupsmime1 1.7.5-5
ii libcupsppdc1 1.7.5-5
ii libgcc1 1:4.9.1-18
ii libstdc++6 4.9.1-18
ii libusb-1.0-0 2:1.0.19-1
ii lsb-base 4.1+Debian13+nmu1
ii poppler-utils 0.26.5-2
ii procps 2:3.3.9-8
Versions of packages cups recommends:
ii avahi-daemon 0.6.31-4
ii colord 1.2.1-1+b1
ii cups-filters [ghostscript-cups] 1.0.61-2
ii printer-driver-gutenprint 5.2.10-3
Versions of packages cups suggests:
pn cups-bsd <none>
ii cups-pdf 2.6.1-14
ii foomatic-db-compressed-ppds [foomatic-db] 20141016-1
ii hplip 3.14.6-1+b1
ii printer-driver-cups-pdf [cups-pdf] 2.6.1-14
ii printer-driver-hpcups 3.14.6-1+b1
ii smbclient 2:4.1.11+dfsg-2
ii udev 215-5+b1
-- debconf information:
cupsys/raw-print: true
cupsys/backend: lpd, socket, usb, snmp, dnssd
--- End Message ---
--- Begin Message ---
Version: 1.7.5-6
Le lundi, 1 décembre 2014, 07.01:23 Carl Worth a écrit :
> On Mon, Dec 01 2014, Brian Potkin wrote:
> > On Sun 30 Nov 2014 at 12:03:17 -0800, Carl Worth wrote:
> >> I looked through the bugs in cups and didn't see any talking about
> >> "symlinks" or "/tmp" so I hope this isn't a duplicate bug entry.
> >
> > It is. :) But not to worry.
>
> Ah, well.
>
> > https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=764253
> >
> > Updating and testing would be good.
>
> Thanks for that information, I will certainly let you know if I see
> any future problems, (and I'll also try to do that in bug reports for
> the affected package).
This issue got fixed in the system-config-printer version 1.4.6-1 and
the 1.7.5-6 cups upload also applied fixes on its end to make sure it
wouldn't be affected again.
I'm therefore hereby closing this bug as fixed in the 1.7.5-6 version;
please don't hesitate to re-open it if it still occurs in a later
version.
Cheers,
OdyX
--- End Message ---