[Bug 441121] [NEW] seahorse sucks CPU, and RAM, causes swap storm, makes system completely unresponsive

[Stephen Warren]

Public bug reported:

Binary package hint: seahorse

I've been attempting to get "bzr viz" to show GPG signatures on
revisions. So, I ran "bzr viz" against a small local branch (or a branch
from http://bazaar.launchpad.net/~bzr/bzr-stats/trunk/), selected
various revisions, and also selected the "Signature" tab. Most of the
time, bzr-gtk will simply say "Signature key not available" (that seems
like another bug...) when switching to view a revision with a signature.
However, sometimes, bzr-gtk will hang, and after a few seconds spew the
following backtrace:

[swarren at esk test1]$ bzr viz
No handlers could be found for logger "bzr"
Traceback (most recent call last):
  File "/usr/lib/python2.6/dist-packages/bzrlib/plugins/gtk/revisionview.py", line 486, in _switch_page_cb
    self.signature_table.set_revision(self._revision)
  File "/usr/lib/python2.6/dist-packages/bzrlib/plugins/gtk/revisionview.py", line 210, in set_revision
    self.show_signature(crypttext)
  File "/usr/lib/python2.6/dist-packages/bzrlib/plugins/gtk/revisionview.py", line 229, in show_signature
    (cleartext, key) = seahorse.verify(crypttext)
  File "/usr/lib/python2.6/dist-packages/bzrlib/plugins/gtk/seahorse.py", line 83, in verify
    keyset[key] = Key(key)
  File "/usr/lib/python2.6/dist-packages/bzrlib/plugins/gtk/seahorse.py", line 99, in __init__
    fields = openpgp.GetKeyFields(key, ['fingerprint', 'trust', 'flags', 'display-name', 'location'])
  File "/var/lib/python-support/python2.6/dbus/proxies.py", line 140, in __call__
    **keywords)
  File "/var/lib/python-support/python2.6/dbus/connection.py", line 622, in call_blocking
    message, timeout)
dbus.exceptions.DBusException: org.freedesktop.DBus.Error.NoReply: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.

Right after this happens, seahorse-agent (or seahorse-daemon, whichever
it is) will start sucking a lot of CPU, and also growing its RAM usage
very fast. If it isn't killed quickly, it'll consume all physical RAM
and grow into swap. The CPU usage and resulting swap storm make the
system almost entirely unresponsive; even switching VTs and text-mode
login doesn't work. Sometimes, switching VTs and CTRL-ALT-DEL will
gracefully restart. (As an aside, epic fail for the kernel; OOM killer
ended up killing some completely irrelevant and tiny Apache process,
then not bothering to ever kill anything else, like the runaway seahorse
process)

Final note: I have the WAR from
https://bugzilla.gnome.org/attachment.cgi?id=135042&action=edit applied,
so as to avoid https://bugs.launchpad.net/bzr-gtk/+bug/377476. I don't
expect this makes a difference, but you never know.

ProblemType: Bug
Architecture: i386
DistroRelease: Ubuntu 9.04
NonfreeKernelModules: nvidia
Package: seahorse 2.26.1-0ubuntu1 [modified: usr/share/dbus-1/services/org.gnome.seahorse.service]
ProcEnviron:
 LANG=en_US.utf8
 SHELL=/bin/bash
SourcePackage: seahorse
Uname: Linux 2.6.28-15-generic i686

** Affects: seahorse (Ubuntu)
     Importance: Undecided
         Status: New


** Tags: apport-bug i386

-- 
seahorse sucks CPU, and RAM, causes swap storm, makes system completely unresponsive
https://bugs.launchpad.net/bugs/441121
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.

-- 
ubuntu-bugs mailing list
ubuntu-bugs at lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs