| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Update the list of ethtool commands to include:
ethtool -e (EEPROM dump)
ethtool -P (permanent MAC address)
ethtool -l (channel/queue settings)
ethtool --phy-statistics
ethtool --show-priv-flags
ethtool --show-eee
All of the above are helpful in understanding the state of modern NICs.
And -P is nice to have as otherwise there is no reliable way to see the
permanent MAC of team ports.
Resolves: #1503
Signed-off-by: Patrick Talbert <ptalbert@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Adds a new plugin for the storage event monitoring daemon peripety.
Resolves: #1502
Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In containerized OSP env we now do not have no mysql packages
installed outside the containers. Therefore the mysql plugin does
not auto run.
This adds the openstack-selinux package to the RH package list of
the mysql plugin like on the OSP plugins.
Resolves: #1501
Signed-off-by: Martin Schuppert <mschuppert@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Adds a new plugin for the stratis storage manager
Resolves: #1495
Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The 'describe' option logic was incorrectly nested behind the 'all'
option logic - so users couldn't successfully use the 'describe' option
unless they also specified 'all'.
Fixes the alignment so that 'describe' can be used properly and
separately.
Additionally fixes an issue where the syntax for getting resources for a
specified namespace only worked for pods, and otherwise generated an
error.
Resolves: #1494
Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We nest ThreadPoolExecutors to run plugins to enable whole-plugin
timeouts, with the timeout being handled by the nested threadpool.
However, calling shutdown() on the nested thread is incorrect in the
event of a plugin timeout, as it only cancels pending jobs, not
currently running jobs that the timed-out plugin is.
Instead, call _threads.clear() to forcibly cancel the running job, thus
allowing the thread the return to the higher level threadpool and allow
other plugins to run via that thread.
This still allows data captured up to the point of the timeout to be
written to the sosreport.
Fixes: #1446
Resolves: #1492
Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Since the temporary directories have already been set up by the
time we process the --list-* handling (plugins, profiles, presets),
we still need to follow the _cleanup() path at exit.
To avoid a duplicate _cleanup() and resulting File Not Found
exception, raise SystemExit out of these paths, which takes us
through the normal exception handling in SoSReport.execute() and
is consistent with all other paths through the command.
Resolves: #1594.
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
| |
The configs member is now a tuple: we can't just slam it into a
add_copy_specs() list or a do_file_sub() call and have it work.
Split out the a_c_s() and pass the tuple alone (it looks less ugly
than having a cons inside the list), and loop over the elements in
postproc to handle the possible file locations.
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Present in the /sys/class/ directory is a fourth fc directory
called fc_vports. Collect attribute info for ports in this
directory as they are useful in troubleshoting fabric config
issues.
Resolves: #1490
Signed-off-by: John Pittman <jpittman@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
| |
The scsi_host attributes can be very important when identifying
LLD/shost level issues. Add host value collection via udevadm.
Signed-off-by: John Pittman <jpittman@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
| |
The queue values for block devices are very important for
performance, alignment, and I/O size issues. Add these values
by default.
Signed-off-by: John Pittman <jpittman@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
| |
Resolves: #1426
Signed-off-by: Ashish Humbe <ahumbe@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Adds syspurpose output collection which is a new tool to assist with
subscription manager on RHEL systems.
Also adds rct output for default product certificates.
Resolves: #1489
Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Direct Rules are an increasingly common way to control the firewall,
but current sos plugin output does not get these.
Update the plugin to grab firwall-cmd output about Direct Rules, and
also pick up the 'state' output and newer logging settings.
Resolves: #1486
Signed-off-by: Patrick Talbert <ptalbert@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
While running some tests using sosreport in a snap, I ran into the
following issue:
Creating compressed archive...
[archive:TarFileArchive] An error occurred compressing the archive:
unorderable types: int() >= str()
This is because the `xz_version` variable might be set to a list
containing one integer, but it was later compared to a list of strings,
raising an exception.
Resolves: #1485
Signed-off-by: Pierre Equoy <pierre.equoy@canonical.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
Remove /sbin/ for service command so it runs on ubuntu.
Move initctl command to upstart plugin
Resolves: #1484
Signed-off-by: Bryan Quigley <bryan.quigley@canonical.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
| |
The outer loop was also double-indented (my local version of
pycodestyle did not pick this up, nor did my eyes...).
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
Collect further ipvsadm details and always do so with the -n numeric
option.
Resolves: #1483
Signed-off-by: Patrick Talbert <ptalbert@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
| |
Re-format the indentation and structure of the innermost loop in
the failed deployments capture.
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
| |
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
| |
The OpenStack undercloud can support > 1 stack. This patch will iterate
through all the possible stacks deployed on the undercloud.
Signed-off-by: Joe Talerico <rook@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
| |
When attempting to add a preset when /var/lib/sos/presets does not
exist, the attempt will fail. Now, if the directory does not exist, sos
will attempt to create it.
Resolves: #1480
Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The manpage for sosreport was previously missing documentation for the
new preset feature in 3.6.
This updates the manpage to include entries for add-preset, del-preset,
list-presets, and the --desc options.
Resolves: #1479
Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Inspect smart card reader information using opensc command
pkcs11-tool and collect its configuration.
Resolves: #1478
Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Conventional DEB package extract juju binary in /usr/bin/juju,
but SNAP extract juju binary in /snap/bin/juju.
Fixes: #1475
Resolves: #1476
Signed-off-by: Eric Desrochers eric.desrochers@canonical.com
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
| |
Resolves: #1474
Signed-off-by: Juan-Luis de Sousa-Valadas <jdesousa@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
| |
In case do_file_sub fails to apply a substitution on a non-existing file,
don't log it as an error log to stdout.
Resolves: #1471
Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
Capture additional datapath flows for the OVS
netdev datapath with: ovs-appctl dpctl/dump-flows
Resolves: #1468
Signed-off-by: Andreas Karis <akaris@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Red Hat distributions have not used the ipsec-tools and racoon packages
since RHEL5, and all of this information is collected in the OpenSwan
module via the same paths.
Similarly, Debian distributions have not provided this since wheezy, and
the same data is collected in OpenSwan.
Resolves: #1465
Signed-off-by: Robb Manes <rmanes@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
| |
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When collecting IPSec data, currently nothing prevents the
collection of keys or the cert.db files. Ensure we don't collect
this information as it is private.
Thanks to Robert Bost (@bostrt) for discovering this and Matt Rogers
(@mrogers950) for double checking what we should ignore.
Resolves: #1463
Signed-off-by: Robb Manes <rmanes@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch removes an unnecessary comparison on package's __path__
attribute with an empty string in get_modules method.
When we pass an empty path to _find_plugins_from_list function it
returns None. The NoneType is not an iterable entity, so this
will leads to runtime error when we extend the plugins list with
None value.
Resolves: #1463
Signed-off-by: Sourabh Jain <sourabhjain@linux.ibm.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
Use append mode when writing to plugin exception files: otherwise,
if multiple exceptions occur within one plugin (e.g. in separate
execution phases), the later ones will overwrite the earlier.
Resolves: #1461
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Add mode (e.g. to allow append) support to add_string() function.
Related: #1461
Signed-off-by: Filip Krska <fkrska@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
Newer versions of ovirt-engine can produce a short and readable
diff from the "factory defaults".
Resolves: #1460
Signed-off-by: Dan Kenigsberg <danken@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
| |
Resolves: #1459
Signed-off-by: David Vallee Delisle <dvd@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
Dont rely on device names from "ip -o link" output where the names
contain also physical device name. Rather get the list of device names
from listing /sys/class/net/ directory.
Resolves: #1458
Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When running `sosreport -l`, or the more specific --list-* options,
sos still creates a temp directory and log files, however we were
not cleaning those up - so over time we could have a multitude of
sos temp directories littering the filesystem.
Now always cleanup the temp directory after execution. Only
exception to this would be if we hit an OSError during execution
_and_ the debug option is specified.
Resolves: #1456
Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
dumpe2fs takes one parameter indicating the device to be dumped.
Closes: #1452
Resolves: #1453
Signed-off-by: Xuewei Zhang <xueweiz@google.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
Dump also any offloaded hardware flows by colling:
ovs-dpctl dump-flows type=offloaded -m
Resolves: #1450
Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This partially fixes issue #1264 by handling the CentOS and
CentOS Atomic hosts, partially based on the CentOS packaging
patch [1]. Without this patch, if you try to develop sosreports
on a CentOS host, you will get an empty list of plugins.
[1] https://git.centos.org/blob/rpms!sos.git/c7/SOURCES!sos-3.4-centos-branding.patch
Resolves: #1447
Signed-off-by: Miguel Angel Ajo <majopela@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The plugin previously collected the CIB, which may contain sensitive
information, without scrubbing it. While the plugin could scrub it, it will
already be collected (scrubbed) by the crm_report command run in setup(), so
the plugin's collection is redundant and can be removed.
Resolves: #1444
Signed-off-by: Ken Gaillot <kgaillot@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
pcs commands should be run in pcs_setup() rather than setup(),
so they are only called on distributions that ship pcs
Related: #1444
Signed-off-by: Ken Gaillot <kgaillot@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Handle (theoretical) cases when "kubectl get namespace" returns a line with
white characters only, and prevent IndexError to be raised.
Resolves: #1442
Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The current command in the openstack_nova.py plugin
$ nova list
Returns an incomplete list of instances when called as
administrator, which results in incomplete sosreports.
The right command to use to get a complete list of instances is:
$ nova list --all-tenants
Closes: #1439
Resolves: #1440
Signed-off-by: Miguel Angel Ajo Pelayo <majopela@redhat.com>
Signed-off-by: Assaf Muller <amuller@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Collect emergency dumps of the lustre ring buffer if all_logs is set.
These files are dumps of the Lustre ring buffer, that are dumped
under certain conditions
Resolves: #1436
Signed-off-by: Nathaniel Clark <Nathaniel.Clark@misrule.us>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
| |
This collects few configuration files and cifs data.
Resolves: #1434
Signed-off-by: Pierguido Lambri <plambri@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Add the ovs-appctl fdb/stats command for each bridge.
Resolves: #1431
Signed-off-by: Eelco Chaudron <echaudron@gmail.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch adds the output of 'systemctl list-jobs', providing a list
of jobs that are in progress. This is sometimes useful when we try to
analyze troubles such as some units are unexpectedly not
started/stopped.
For example, consider there is the inquiry from some customer that
multi-user.target on our system is configured as default.target but is
still inactive after system boot, why?
# LANG=C systemctl get-default
multi-user.target
# LANG=C systemctl status multi-user.target
* multi-user.target - Multi-User System
Loaded: loaded (/usr/lib/systemd/system/multi-user.target; \
enabled; vendor preset: disabled)
Active: inactive (dead)
Docs: man:systemd.special(7)
In such case, systemctl list-jobs gives us the clear suggestion that
there is a still running start job for foobar.service and
multi-user.target and the other units are waiting for it to finish.
# LANG=C systemctl --no-pager list-jobs
JOB UNIT TYPE STATE
112 multi-user.target start waiting
226 systemd-update-utmp-runlevel.service start waiting
258 foobar.service start running
236 systemd-readahead-done.timer start waiting
4 jobs listed.
Resolves: #1427
Signed-off-by: HATAYAMA Daisuke <d.hatayama@jp.fujitsu.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Adds several resources for 'kubectl get' output.
Resolves: #1424
Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
|