aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Merge initrd and bootloader plug-ins into a single boot moduleBryn M. Reeves2013-04-292-33/+15
| | | | | | | | | | | | | | | The existing initrd plug-in is default-disabled so it rarely collects anything useful. The command used to extract the init script from the initramfs image is also broken for modern systems as the pattern ("initrd*") does not match and the file /init on systems using systemd is actually a symbolic link to the systemd binary. Replace the collection of this file with a listing of the current initramfs image (via lsinitrd). Optionally collect a listing from each non-kdmump image present in /boot. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Rename internationalization plug-inBryn M. Reeves2013-04-291-3/+3
| | | | | | | This partially reverts commit fbbb6ac. Renaming the module was a mistake; internationalization is just too long. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Add grub2 support and make bootloader more modularBryn M. Reeves2013-04-294-6/+93
| | | | | | | | | Add a new grub2 plug-in and separate the grub and lilo support from the bootloader module which now collects legacy and special purpose or arch bootloader configuration and a directory listing of /boot. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Move all block device data from filesys to block moduleBryn M. Reeves2013-04-292-43/+14
| | | | | | | | | | | The filesys module has long had a lot of (very weird) code for detecting and processing a list of whole-disk devices in the system. Apart from being garbage this doesn't really belong in the file systems module - move it to block and replace the hokey /proc-and-hdparm-or-regex thing with the existing directory walk over /sys/block. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Remove bogus fdisk invocation from cluster plug-inBryn M. Reeves2013-04-291-1/+0
| | | | | | | | Disk partition data does not belong in the cluster code (or the file systems module for that matter but that's a topic for the next commit). Remove it. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Remove redundant check_enabled() and fix formatting in abrt.pyBryn M. Reeves2013-04-291-8/+9
| | | | | | | | | | The abrt plug-in's check_enabled() just checks for one package and one file; add .packages and .files members to the plug-in class and delete the method. Also tidies up some long lines and string formatting. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Merge pull request #146 from battlemidget/patch-remove-java-i18n-methodsBryn M. Reeves2013-04-291-41/+1
|\ | | | | Remove methods related to java and internationalization.
| * Remove methods related to java and internationalization.Adam Stokes2013-04-271-41/+1
|/ | | | | | | - Old methods not needed anymore, set_i18n and _get_classloader both only useful if we were still supporting jython. Signed-off-by: Adam Stokes <adam.stokes@ubuntu.com>
* Fix lies, damn lies & misnaming in anacron plug-inBryn M. Reeves2013-04-261-1/+1
| | | | | | | The anacron plugin falsely declares itself to be named 'anaconda' (it started life as a copy of it). Fix it. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Fix long lines in veritas plug-inBryn M. Reeves2013-04-261-1/+2
| | | | Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Add system udev rules directory to udev plug-inBryn M. Reeves2013-04-261-1/+5
| | | | Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Add default log size limits to squid plug-inBryn M. Reeves2013-04-261-2/+4
| | | | | | | The squid plug-in can potentially collect large volumes of logs. Limit it to 15M to be consistent with similar modules. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Merge pull request #145 from battlemidget/patch-improve-makefile-debianBryn M. Reeves2013-04-265-59/+6
|\ | | | | Patch improve makefile debian
| * Remove make gpgkey in travis build scriptAdam Stokes2013-04-261-1/+0
| | | | | | | | Signed-off-by: Adam Stokes <adam.stokes@ubuntu.com>
| * Update readme to reflect the removal of jboss/application server build andAdam Stokes2013-04-261-2/+1
| | | | | | | | | | | | reflect the new build task name for debian builds. Signed-off-by: Adam Stokes <adam.stokes@ubuntu.com>
| * Remove uneeded build tasks in MakefileAdam Stokes2013-04-263-56/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | - Since removing the jboss/application server bits we no longer need those build tasks in the Makefile. - Remove building gpgkeys entirely since it is not referenced anywhere in the code. - Update the creation of tarballs to exclude-vcs directories such as .git, .bzr, etc. - Move build tasks from deb-unsign to deb because its less to type and still properly explains what the task is. Signed-off-by: Adam Stokes <adam.stokes@ubuntu.com>
* | Remove duplicate code in system plug-inBryn M. Reeves2013-04-261-18/+3
| | | | | | | | | | | | | | The Red Hat and Debian/Ubuntu system plug-ins are identical. Get rid of them and have a comon System class. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* | Remove ntp data from system plug-inBryn M. Reeves2013-04-262-5/+6
| | | | | | | | | | | | | | Ntp information belongs in the ntp plug-in.. Duh. Move the few files collected in the system module to ntp.py. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* | Remove cron related information from system plug-inBryn M. Reeves2013-04-262-11/+11
| | | | | | | | | | | | | | | | | | The cron data does not belong in the system plug-in and it seems pointless to have a separate module just for crontabs (but not for cron!). Move the cron logs and tabs into a new plug-in named cron. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* | Remove user crontabs from crontab plug-inBryn M. Reeves2013-04-261-3/+0
| | | | | | | | | | | | | | | | | | It's unwise to collect user crontabs (especially "for all accounts with a directory in /home"..) since they could contain confidential information. The icky shell call-out also has not worked in some time - remove it all. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* | Remove anacron data from system plug-inBryn M. Reeves2013-04-262-1/+26
| | | | | | | | | | The system plug-in is a dumping ground of all sorts of random bits and pieces. Move the anacron stuff to its own file.
* | Enable ppp plug-in on more distributionsBryn M. Reeves2013-04-261-3/+2
| | | | | | | | Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* | Rename printing plug-in and add check for cups packageBryn M. Reeves2013-04-261-1/+3
| | | | | | | | | | | | | | | | | | The 'printing' plug-in is almost entirely cups-specific (the lpstat invocations would work with another lp implementation but that's fairly unlikely..). Add a check for the package and rename the module to cups.py. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* | Make printing default log size limit consistentBryn M. Reeves2013-04-261-1/+1
| | | | | | | | | | | | | | The default log size limit is 15M for most plug-ins. The printing plug-in uses 50M. Set it to 15M to avoid very large reports. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* | Enable Openswan plug-in on more distributionsBryn M. Reeves2013-04-261-2/+2
| | | | | | | | Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* | Fix nfsstat invocation in nfsserver plug-inBryn M. Reeves2013-04-261-1/+2
| | | | | | | | | | | | | | | | | | | | The '-a' option is an undocumented option to nfsstat that only reports that ACL data is not yet implemented. The options should be '-o all' to retrieve all available NFS stats. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* | Improve code formatting in nfsserver plug-inBryn M. Reeves2013-04-261-3/+7
| | | | | | | | Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* | Fix indentation in nfsserver.pyBryn M. Reeves2013-04-261-8/+8
|/ | | | Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Remove obsolete netdump plug-inBryn M. Reeves2013-04-261-27/+0
| | | | | | Netdump has not been shipped since RHEL4 and is a Red Hatism. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Remove top output from kvm plug-inBryn M. Reeves2013-04-261-4/+0
| | | | | | | | Subsystem plug-ins have no business collecting general performance data. Remove the top output collection and topOutput command line option. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Fix detection of ext file systems in filesys plug-inBryn M. Reeves2013-04-261-3/+5
| | | | | | | | | | | | The code to collect dumpe2fs output does not work on modern systems due to changes in the output of the mount command. Have the plug-in use /proc/mounts instead and simplify and improve the regex used to find devices containing ext[234] file systems and switch to 'dumpe2fs -h' collection rather than getting the full group lists. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Fix broken option_list in filesys plug-inBryn M. Reeves2013-04-261-2/+2
| | | | | | | | The filesys plugin has the "option_list defined twice" bug. Fix it and ensure that both the 'lsof' and 'dumpe2fs' options are usable. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Fix UnboundLocalError during command output substitutionBryn M. Reeves2013-04-261-0/+4
| | | | | | | | | | | | | | | | | | | | | Plugin.do_cmd_output_sub() assumes that the list of executed commands is non-empty if the list of commands requested is non-empty. This can be false (e.g. when a command is run speculatively and the binary is not present) leading to an unbound variable exception as the code never sets the variable used as the return value: Traceback (most recent call last): File "/usr/sbin/sosreport", line 23, in <module> main(sys.argv[1:]) File "/usr/lib/python2.7/dist-packages/sos/sosreport.py", line 1154, in main sos.execute() UnboundLocalError: local variable 'replacements' referenced before assignment > /usr/lib/python2.7/dist-packages/sos/plugins/__init__.py(197)do_cmd_output_sub() -> return replacements Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Clean up plug-in naming and set plugin_name for all plugin classesBryn M. Reeves2013-04-2688-93/+290
| | | | | | | | | | | | | | | | Clean up some plug-ins with non-conforming or undescriptive names, set plugin_name in all base plug-in classes to avoid <dist><plug> names appearing in reports and remove the obsolete amd plug-in. Rename several plug-ins: - rhn -> satellite (also spacewalk) - i18n -> internationalization - qpidd -> qpid - ftp -> vsftpd Enforce pep8 style class names for all plug-in classes. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Kill the automount daemon plug-in (amd)Bryn M. Reeves2013-04-261-28/+0
| | | | | | | Nothing current ships the amd automounter any more. Remove the plug-in from the tree. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Fix running of RHN hardware.py scriptBryn M. Reeves2013-04-251-1/+1
| | | | | | | The hardware.py script is no longer executable. Run it via the python interpreter. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Fix cpu frequency data collection in hardware.pyBryn M. Reeves2013-04-251-4/+10
| | | | | | | | | | The cpu frequency reporting on modern systems has moved from cpufrequtils and the cpufreq command to kernel-tools and the cpupower command. Check for the respective packages and run the correct command. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Break up the device-mapper trainwreckBryn M. Reeves2013-04-256-45/+171
| | | | | | | | | | | | | | | | | | | | | Break the device-mapper plug-in up into separate plugins for dm itself and each of the current device-mapper applications: - devicemapper.py - dmraid.py - block.py - lvm2.py - md.py The data collection for each subsystem is largely unchanged other than to fix bugs and remove obsolete items: - support for multipath bindings in /var dropped (only relevant to RHEL5 and earlier) - collect multipath -l separately from -ll - fix 'udevinfo' -> 'udevadm info' Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Add RHEL7 support to RedHatPolicy.rhel_version()Bryn M. Reeves2013-04-251-0/+2
| | | | Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Merge pull request #138 from battlemidget/include-authors-fileBryn M. Reeves2013-04-243-23/+4
|\ | | | | Point sosreport manpage to see the AUTHORS file for updated contributors
| * Point sosreport manpage to see the AUTHORS file for updated contributorsAdam Stokes2013-04-233-23/+4
|/ | | | | | | - Include AUTHORS file in the tarball. - Consolidated where list of Authors may be found. Signed-off-by: Adam Stokes <adam.stokes@ubuntu.com>
* Fix scoping bug in named.pyBryn M. Reeves2013-04-231-1/+1
| | | | | | | | The path "/etc/named.conf" was made into a class variable to avoid repeating the string through the file. Make sure all uses qualify the variable appropriately. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Move command line parsing out of SoSReport classBryn M. Reeves2013-04-231-70/+316
| | | | | | | | | | | | | | | | | | | | Separate the command line logic from the rest of the SoSReport class and add a new SoSOptions class to contain it. SoSOptions may either be initialised from a command line by passing in an argument list to the initializer or by creating an empty object and setting values for individual properties. For now a SoSOptions object is initialized in the SoSReport initializer (from the argument list passed from __main__). Eventually this will move into __main__ allowing SoSOptions and SoSReport to become an interface for external users. The next step is to move all input validation into the SoSOptions class (it's currently spread around the code that uses the options). Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Rename listPlugins to list_pluginsBryn M. Reeves2013-04-231-2/+2
| | | | | | | Rename the option variable listPlugins to list_plugins to conform with pep8 and other uses in the options set. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Clean up SoSReport.execute()Bryn M. Reeves2013-04-231-15/+22
| | | | | | | | | | | | | | | | | | | | Reorganise the SoSReport object's execute() method to facilitate separation of the command line and report generation logic: - Rename ensure_plugins() as verify_plugins() and have it return a boolean to the execute() method. - Push logging down from execute into the individual action methods that carry out the tasks being logged. - Have execute() return a boolean indicating whether the requested operation completed successfully or not. Longer term it may be better to define an enumeration class for sos return statuses. This would allow more expression than simple 'pass' or 'fail' and would help the move toward an API. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Fix scoping bug in kernel.pyBryn M. Reeves2013-04-231-2/+2
| | | | | | | | The path "/sys/module" was made into a class variable to avoid repeating the string through the file. Make sure all uses qualify the variable appropriately. Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
* Merge branch 'new_ldap' of https://github.com/karibou/sosreportBryn M. Reeves2013-04-231-3/+20
|\ | | | | | | | | Conflicts: sos/plugins/ldap.py
| * ldap.py : improved data collection on Ubuntu/DebianLouis Bouchard2013-04-231-4/+20
| |
* | Merge pull request #136 from karibou/masterBryn M. Reeves2013-04-234-6/+7
|\ \ | | | | | | Minor fixes and reformat
| * | fix small misses following refactoringLouis Bouchard2013-04-222-3/+5
| | |
generated by cgit (git 2.34.1) at 2024-09-23 12:21:40 +0000