| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
| |
To collect /etc/os_net_config directory.
Resolves: #1088
Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
|
|
|
|
|
|
|
| |
Fixes: #920
Signed-off-by: Marcus Linden <mlinden@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
| |
Only collect logs by wildcard if --all-logs is given.
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
| |
Colecting logs and configs for the daemon and its proxy.
Resolves: #1080
Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Collect iptables and ip6tables for any table present in
/proc/net/ip*_tables_names.
Relevant to: #1092
Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
|
|
|
|
|
|
| |
Relevant to: #1092
Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
|
|
|
|
|
|
|
|
|
| |
To prevent timeouts of "df" commands when using a huge amount of
direct mappings on autofs.
Resolves: #1093
Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When pressing Ctrl+C during user input (name and case id), sosreport
treats that as an empty string and starts to collect data. It must
terminate instead.
Catch the KeyboardIntterupt, print a newline (to avoid munging the
terminal too much), and then re-raise the exception to cause the
process to exit.
Allow all other exceptions to propagate up and do not attempt to
continue.
Resolves: #1094
Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Call "ip -d address" to collect all details instead of
"ip address" only.
Resolves: #1084
Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
|
|
|
|
|
|
|
|
| |
Collect quota information and stats of users with a key on the system.
Resolves: #1085
Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Reverse sort the file list by modification time, to ensure that
the most recent data is included when limits apply.
Fixes: #1082
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add Tripleo Pike opinionated config paths to be collected
for services, when running in containers. Each service has a
config dir in /var/lib/config-data/puppet-generated on the host
which gets bind mounted read only into the container.
Closes #1054
Signed-off-by: Martin Schuppert <mschuppe@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bad ZIP files can raise zipfile.BadZipfile exception.
Fixes:
Traceback (most recent call last):
File "/home/msrb/projects/sos/sos/sosreport.py", line 1252, in setup
plug.setup()
File "/home/msrb/projects/sos/sos/plugins/jars.py", line 72, in setup
if Jars.is_jar(path):
File "/home/msrb/projects/sos/sos/plugins/jars.py", line 98, in is_jar
with zipfile.ZipFile(path) as f:
File "/usr/lib64/python2.7/zipfile.py", line 770, in __init__
self._RealGetContents()
File "/usr/lib64/python2.7/zipfile.py", line 842, in _RealGetContents
raise BadZipfile("Bad magic number for central directory")
BadZipfile: Bad magic number for central directory
Signed-off-by: Michal Srb <michal@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
Other known locations can be added via:
-k jars.all_known_locations
Additional user-specified locations can be added via:
-k jars.append_locations=...
Signed-off-by: Michal Srb <michal@redhat.com>
|
|
|
|
|
|
|
| |
The Python listdir() function never returns the special '.' and
'..' directory entries. Do not test for them.
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
| |
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
| |
This is a new plugin used to collect data for SAS adapters.
It collects a list of adapters and then puts information on
each adapter into the sosreport.
Signed-off-by: Kyle Mahlkuch <Kyle.Mahlkuch@ibm.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This moves several collected outputs over from using curl to using the
etcdctl command, as well as adding some addition collection.
Only statistic information is now obtained via curl.
Also we now attempt to determine the actual listening address based
on etcd.conf, and if that is not possible the plugin assumes
version-specifc defaults.
Closes: #1059.
Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
|
|
|
|
| |
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
| |
A single-valued tuple needs a ',' before the closing paren to
distinguish it from a value in parentheses.
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If a command is not found in Plugin.get_command_output() and the
run is not configured to use an alternate sysroot (either via
policy based autodetection or command line --sysroot=/path) the
plugin would log messages like:
[plugin:gluster] command 'gluster' not found in / - re-trying in host root
This is incorrect: without an alternate sysroot the attempt should
fail immediately at this point (since the 'retry' will attempt to
find the command in the same namespace).
This is caused by an incorrect test for alternate sysroot:
if chroot and self.commons['cmdlineopts'].chroot != 'always':
The 'chroot' variable is always True here, unless the caller has
explicitly disabled it, or we are re-trying a command in a run
with an alternate sysroot defined.
Instead test that 'root' is not None and that it is not equal to
'/'.
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
| |
Ignore IO errors reported by /proc and /sys.
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add a new list/tuple member to the Plugin class that contains a
list of executable commands that will enable the plugin if any
are present.
For example, a plugin:
class MyPlugin(Plugin, RedHatPlugin):
commands = ('mycmd1', 'mycmd2')
Will be automatically enabled if either 'mycmd1' or 'mycmd2'
is present and executable in the policy defined PATH for the
run.
Related: #1051.
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
New plugin to collect OpenStack Aodh configurations and logs.
Signed-off-by: Sachin Patil <psachin@redhat.com>
(minor style fix)
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
| |
This is an addition to pull request 1046 to collect gnocchi logs
correct when service is running in a container.
Signed-off-by: Martin Schuppert mschuppert@redhat.com
|
|
|
|
|
|
|
|
|
|
|
|
| |
The tool always requires PIN to be entered manually. That contradicts the aim
of sosreport to collect data as automatically as possible.
Once smartcar community allows providing the PIN e.g. via an env.variable,
we can add the command back.
Resolves: #1045
Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The output of ovs-ofctl varies depending on the version of OpenFlow you
query it with. Using the currently supported versions, retrieve
ovs-ofctl output for each ovs-bridge depending on which specified
versions of OpenFlow are in use.
Signed-off-by: Robb Manes <rmanes@redhat.com>
(minor style tweaks)
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Adds a new plugin for docker-distribution (registry v2).
Collects the journal, registry config and a view of the contents of
images in the registry.
Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
|
|
|
|
|
|
| |
We're using the "openstack" command to stream the ui log file to stdout.
Signed-off-by: Honza Pokorny <honza@redhat.com>
Signed-off-by: Adam Stokes <battlemidget@users.noreply.github.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Red Hat Certificate System 9 (Dogtag 10) has been out for a while.
This commit adds the ability to capture those logs and config
files for each subsystem.
Added pki-base because redhat-pki or dogtag-pki do not have to be
installed to build a Certificate Authority.
Removed "/usr/share/java/pki" check because CS 9 uses the same
directory.
Signed-off-by: Brian Gribble <bgribble@redhat.com>
Signed-off-by: Adam Stokes <battlemidget@users.noreply.github.com>
|
|
|
|
|
|
|
| |
Fix small typo in the chroot section of the man page (alwyas -> always)
Fixes #1035
Signed-off-by: Adam Stokes <battlemidget@users.noreply.github.com>
|
|
|
|
|
|
|
|
|
| |
With Tripleo Pike the default ENV changed from OS_TENANT_NAME to
OS_PROJECT_NAME. The commands of the openstack plugins against
the env should run when either OSP_TENANT_NAME or OS_PROJECT_NAME
is set.
Signed-off-by: Martin Schuppert mschuppe@redhat.com
Signed-off-by: Adam Stokes <battlemidget@users.noreply.github.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add Tripleo Pike opinionated logs paths to be collected
for services, when running in containers. This is a temporary
and will be reworked for Queens, like switching those to syslog
or fluentd shipping logs to Elasticsearch cluster.
Partial bug: https://bugs.launchpad.net/tripleo/+bug/1700909
Related blueprint: https://review.openstack.org/#/c/462900
Signed-off-by: Bogdan Dobrelya <bdobreli@redhat.com>
Signed-off-by: Adam Stokes <battlemidget@users.noreply.github.com>
|
|
|
|
|
|
| |
collecting the /sys/kernel/config/target
Signed-off-by: venkata edara <redara@redhat.com>
Signed-off-by: Adam Stokes <battlemidget@users.noreply.github.com>
|
|
|
|
|
|
| |
gluster_block plugin collects config/log files related to blocks
Signed-off-by: venkata edara <redara@redhat.com>
Signed-off-by: Adam Stokes <battlemidget@users.noreply.github.com>
|
|
|
|
| |
Signed-off-by: Adolfo Jayme Barrientos <fitojb@ubuntu.com>
Signed-off-by: Adam Stokes <battlemidget@users.noreply.github.com>
|
|
|
|
| |
Signed-off-by: Louis Bouchard <louis.bouchard@canonical.com>
Signed-off-by: Adam Stokes <battlemidget@users.noreply.github.com>
|
|
|
|
| |
Signed-off-by: Adam Stokes <battlemidget@users.noreply.github.com>
|
|
|
|
|
|
|
|
|
|
| |
This includes information for users wishing to submit PR's and is visible during
their creation of the PR. This allows them to validate that they are providing
all necessary requirements prior to submitting and should reduce the amount of
back and forth we do asking users to follow our guidelines.
Fixes #1021
Signed-off-by: Adam Stokes <battlemidget@users.noreply.github.com>
|
|
|
|
|
|
|
| |
We've had a lot of contributions from outside companies, organizations, and
individual contributors. I'd like to update this file to acknowledge those new
contributors and continue to encourage and promote this products ecosystem.
Signed-off-by: Adam Stokes <battlemidget@users.noreply.github.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Collect the entire /sys/class/dmi/id/ directory from the
hardware plugin. Today, this is really the "native" format for
this information, as far as Linux is concerned. The dmidecode
program (which historically at least read the data directly
from firmware tables, as no proc or sysfs interface existed) is
a legacy formatting of this information which we are
unfortunately obliged to keep around a little longer yet, as
many users have grown dependencies on its format.
The data included in this directory is small (about 500 bytes
on my system) and will not have an impact on the time or size
of the report.
Original discussion at:
https://github.com/sosreport/sos/pull/1034
Signed-off-by: David Coronel <david.coronel@canonical.com>
Signed-off-by: Adam Stokes <battlemidget@users.noreply.github.com>
|
|
|
|
|
|
|
|
| |
The binary name changed from docker.io to docker.
Update config and log file locations.
Signed-off-by: Karsten Graul <kgraul@linux.vnet.ibm.com>
Signed-off-by: Christy Perez <christy@linux.vnet.ibm.com>
Signed-off-by: Adam Stokes <battlemidget@users.noreply.github.com>
|
|
|
|
|
|
| |
Fixes patch collection on an incorrect samba path.
Signed-off-by: Guenther Deschner <gd@samba.org>
Signed-off-by: Adam Stokes <battlemidget@users.noreply.github.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
On Red Hat systems, we use the version of the installed filesystem'
package to enable handling of UsrMove[1] in the PATH used when
running commands from sos.
If the package is not present (e.g. because of a broken system
configuration, rpm timeout, corrupt database or other cause),
previous versions would fail immediately. In this situation
there was no way for the user to attempt a complete run (short
of fixing the problem related to the 'filesystem' package).
Now that the timeout has been lengthened for this command, make
the UsrMove determination a bit more robust to prevent users from
being unable to run sos at all:
- If RPM fails entirely, exit (plugin enablement tests will fail)
- If RPM succeeds but 'filesystem' is missing assume UsrMove
(this is true for all supported Red Hat distros today)
- Otherwise apply the normal PATH logic.
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
Do not collect temporary mount paths at /var/lib/ceph/tmp/mnt.XXXX.
Resolves #1006
Replaces #1005
Signed-off-by: Tomas Petr <tpetr@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Finish off core set of Neutron resources:
* Networks
* Security groups
* Floating IPs
As well as the API extensions exposed by the server:
* Extensions
Resolves: #1011.
Signed-off-by: Assaf Muller <amuller@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
| |
Resolve issue #886
Change-Id: I0496cf4f4dbee33a8a6c52b2b6a1a3271834fb6f
Signed-off-by: Sandro Bonazzola <sbonazzo@redhat.com>
|
|
|
|
|
|
|
| |
The per-pid cgroup data is in a pseudofile named 'cgroup', and not
'cgroups' as in commit 2523ad5.
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The pstore data should be collected upon sosreport run for all
architecture (currently only enabled in powerpc.py) because the
data there will include any past oops event and/or panic events.
This data would assist diagnosing panics with little other data
available.
Also, the data included is a few kb in size, so it will not add
much size to the output of the report.
I just moved the /sys/fs/pstore collection from powerpc.py
to kernel.py in order to have it collect for all architectures.
Signed-off-by: Chris Newcomer <chris@thenewcomers.org>
Signed-off-by: Adam Stokes <battlemidget@users.noreply.github.com>
|
|
|
|
|
|
|
| |
When the stat is done on a broken symlink, it will return either OSError
or FileNotFoundError. Handle these and report accordingly.
Signed-off-by: Louis Bouchard <louis.bouchard@canonical.com>
Signed-off-by: Adam Stokes <battlemidget@users.noreply.github.com>
|