aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* [docs] Add general sos manpage, rename sosreport.1, add symlinksJake Hunsaker2020-04-224-334/+418
| | | | | | | | | | | | Adds a general manpage for the sos command that then points to the available components. Renames sosreport.1 to sos-report.1 to match the component style, and then symlinks sosreport.1 to it. Adds a symlink from sos-collector.1 to sos-collect.1. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [docs] Add manpage for sos collectJake Hunsaker2020-04-221-0/+312
| | | | | | | | | | Adds the manpage from sos-collector to the tree as 'sos-collect', and updates the instances of 'sos-collector' to 'sos collect' within the manpage. Additionally updates the man page to include newly available options. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [collector] Properly set desc attribute for parserJake Hunsaker2020-04-221-1/+8
| | | | Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [linters] PEP8, PEP257, flake8 fixesJake Hunsaker2020-04-225-140/+136
| | | | | | | Various fixups for PEP8, PEP257, and unused imports as highlighted by flake8. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [Hosts|Policies|Collect] Merge SoSHost into Policy, drop SoSHostJake Hunsaker2020-04-228-264/+108
| | | | | | | | | | | | | | | | | | Moves the functions from `SoSHost` into `Policy`, and adjusts the calls in `SoSNode` appropriately. Adds a `pkg_version` method to `PackageManager` to satisfy the version checking requirements that had previously been using `SoSHost`. Finally drops the sos/collector/hosts directory from the tree as it is no longer used. At this point, `sos collect` should be able to be run on any distribution that has a defined `Policy`. Note that new policies will need to specify 2 new class attrs, `sos_pkg_name` and `sos_bin_path` in order for them to be supported as well. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [collector|Policies] Load a Policy rather than SoSHost within SoSNodeJake Hunsaker2020-04-222-14/+15
| | | | | | | | | | | | | | | Rather than load a `SoSHost` instance, `SoSNode` will now load a `Policy` that is determined by the content of `/etc/os-release` on the remote node, as this is a standardized file for Linux-based systems and allows for reliable identification of the host. Each per-node policy will load the host-side package information for use by collector in determining cluster profile and sos capabilities. This per-node policy will not however load an `InitSystem()` or a `ContainerRuntime()`. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [PackageManager] Allow for remote execution of package commandsJake Hunsaker2020-04-227-21/+50
| | | | | | | | | | When loading a policy, there is now support for a `remote_exec` parameter, which should be an ssh command, to which will be appended `PackageManager` related commands so that collector can get package data off of the remote nodes, without having to reimplement a subset of `PackageManager` itself. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [policies] Expose init and probe_runtime to all policiesJake Hunsaker2020-04-226-22/+33
| | | | | | | | | Exposes the `init` and `probe_runtime` parameters of `Policy` to all existing vendor-defined policies. This will allow collector to replace `SoSHost` with `Policy` while not loading the local system's init system and container runtime information. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [Policy] Allow check() to inspect remote contentJake Hunsaker2020-04-228-15/+68
| | | | | | | | | | | Updates the `check()` method within `Policy` to allow consumers to specify a `remote` parameter that can be used in place of the locally available files. This is the first step in merging `SoSHost` from collector with the standard `Policy` which is far better defined. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [collect] Make batch actually completely non-interactiveJake Hunsaker2020-04-221-3/+10
| | | | | | | | | We should not have any exceptions to --batch being non-interactive, so in the event that a user would need a password but fails to provide one alongside the --batch option, it should be expected that the run will fail, rather than stop to collect the password. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [collect] PEP8 fixupsJake Hunsaker2020-04-221-21/+19
| | | | Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [component|collector] Fix collector --list-optionsJake Hunsaker2020-04-222-13/+27
| | | | | | | | | | | Fixes `sos collect --list-options` to actually display known cluster options, as this was initially lost/overlooked in the move from a standalone utility to sos-4.0 integration. Additionally, only setup a log file if none of the known `list*` options are set for any component. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [clusters|node] Fix setting node SSH key based on clusterJake Hunsaker2020-04-223-2/+18
| | | | | | | | | | | | | | Previously, the ovirt (and family) cluster profile would attempt to set a global SSH key if the ovirt-generated key was locally available by changing the options entry for ssh_key. This no longer worked since decoupling the shared dict, so instead create a new mechanism by which clusters may set this. It may prove necessary in the future to extend this functionality to arbitrary settings, however for now I do not see that being common place. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [sosnode] Move modification of sos command into SoSNode()Jake Hunsaker2020-04-223-27/+35
| | | | | | | | Moves the update of options for a node based on a cluster profile out of `Cluster()` and into `SoSNode()`, since the commons dict is not a shared instance. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [collector] Adjust order of prompts and setupJake Hunsaker2020-04-222-34/+41
| | | | | | | | | Since moving away from a shared dict configuration, we need to do the initial steps of displaying the disclaimer, prompting for input, and connecting to the master to determine a cluster type in a slightly different order. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [archive|component] Move archive setup and cleanup into SoSComponentJake Hunsaker2020-04-224-93/+86
| | | | | | | | | | | | | | | | | | | | Moves archive setup and cleanup directly into `SoSComponent()`, then has `SoSCollector` utilize the class rather than manually creating the archive at the end of a run. `SoSReport()` has been updated where needed to leverage the new location of `cleanup()`. This also adds encryption options to `SoSCollector()` to encrypt the final archive. It may prove later than this option should also (or instead) pass the encryption option to the sosreport runs that occur on the hosts. Note that this commit marks the point from which functional, complete, runs of `sos collect` are expected to run without error and result in a tarball of one or more collected sosreports. Resolves: #1988 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [collector] Align with SoSComponent.execute() callsJake Hunsaker2020-04-222-6/+12
| | | | | | | | | Moves the actual workings of SoSCollector() into an execute() method to be called by the sos binary. The initialization of SoSCollector now just handles policy and logging setup, and option validation. All connections to nodes and further command execution is now gated behind execute() Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [collector] Drop Configuration class and use SoSOptionsJake Hunsaker2020-04-225-447/+267
| | | | | | | | | | | | | | | Moves the `collect` component away from using the Configuration class and merges relevant options into the SoSOptions framework, whereas other options are made into either class attrs or part of a 'commons' dict similar to what we do in `report`. At this point, `sos collect` has basic discovery functionality, meaning that a user can run it locally or point it to a master node via `--master` and get a returned list of nodes that would be collected from. However, collection is likely to fail even if the remote command execution works. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [collector] Transition logging to logger setup by SoSComponentJake Hunsaker2020-04-223-49/+39
| | | | | | | Rather than create multiple loggers, transition the sos-collector code to instead use the logger created by SoSComponent() initialization. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [collector] Add cluster and host loading into SoSCollector() directlyJake Hunsaker2020-04-221-2/+99
| | | | | | | | | | | | | | As part of moving away from the Configuration() approach, move the loading of support host and cluster types directly into SoSCollector(). Again, `sos collect` is not expected to be functional as of this commit, these are iterative updates to move towards that goal. Similarly, it is still desired that by the end of this patch series, "host types" are really just `Policy()` instances. Related: #1988 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [collector] Initial import of sos-collector into sosJake Hunsaker2020-04-2214-1/+2868
| | | | | | | | | | | | | | | | | | | | | | | | This commit represents the first import of the sos-collector project into sos natively for 4.0. It is not expected to be functional as of this commit. A minimum integration has been done at this point - namely tying SoSCollector() as a SoSComponent() subclass, and hooking up the parser functions. SoSCollector will load a policy in the same way as 'sos report', and should use the same logging methods, which will be done in a future commit in this series. As a divergence from the standalone sos-collector project, this integration aims to hook in host detection with the existing Policy model provided by sos. Additionally, we should be able to drop the Configuration dict-subclass approach by the time this series is ready for merge. Related: #1988 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Fix unittestsJake Hunsaker2020-04-213-4/+4
| | | | | | | | | | Fixes a unittest failure due to an attempt to pass print() a `file` keyword for Red Hat, which is no longer accepted in python3 Additionally fixes 2 string related errors in the test suite since changing from python2 to python3. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [insights] collect insights-client dumpPavel Moravec2020-04-211-0/+4
| | | | | | | | | | Collect data of "insights-client --offline" into sos_commands/insights/insights-client-dump directory. Resolves: #2030 Signed-off-by: Pavel Moravec <pmoravec@redhat.com> Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [systemd] update trigger with better generic file checkHeather Lemon2020-04-211-4/+1
| | | | | | | | | | | | | | | | Currently it checks for the existance of systemd using two different directories depending on the distro used. Updating systemd trigger to handle the detection of checking for systemd more reliably and generically. Reference: https://github.com/systemd/systemd/blob/master/src/libsystemd/sd-daemon/sd-daemon.c#L598-L604 Closes: #2013 Resolves: #2029 Signed-off-by: Heather Lemon <heather.lemon@canonical.com> Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [landscape] Substitute oidc conf in service fileEric Desrochers2020-04-211-0/+20
| | | | | | | | | | | | | | | | | | | | Substitute sensitive informations about oidc found in landscape service configuration file. From release 19.10, Landscape can use OpenID-Connect (OIDC) to authenticate users. To enable OpenID-Connect support, please add oidc-issuer, oidc-client-id and oidc-client-secret to /etc/landscape/service.conf in the [landscape] section. Reference: https://docs.ubuntu.com/landscape/en/onprem-auth#openid-connect-support Closes: #2023 Resolves: #2025 Signed-off-by: Eric Desrochers <eric.desrochers@canonical.com> Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [navicli] replace interactive prompt by plugin optionPavel Moravec2020-04-161-25/+5
| | | | | | | | | | | List of ClariiOn IP addresses should be provided via -k navicli.ipaddrs plugopt, as a space separated list. Closes: #405 Resolves: #2020 Signed-off-by: Pavel Moravec <pmoravec@redhat.com> Signed-off-by: Bryan Quigley <bryan.quigley@canonical.com>
* [freeipmi] Add freeipmi supportArif Ali2020-04-161-0/+33
| | | | | | | Resolves: #2022 Signed-off-by: Arif Ali <arif.ali@canonical.com> Signed-off-by: Bryan Quigley <bryan.quigley@canonical.com>
* [pmem] Adds a new 'pmem' (Persistent Memory) plugin to collect dataSteve Scargall2020-04-161-0/+91
| | | | | | | | | | Uses ndctl, daxctl, and ipmctl utilities Closes: #1398 Resolves: #2021 Signed-off-by: Steve Scargall <steve.scargall@intel.com> Signed-off-by: Bryan Quigley <bryan.quigley@canonical.com>
* [landscape] simplify and remove deprecated log dirEric Desrochers2020-04-141-11/+15
| | | | | | | | | | | | | | * Make add_copy_spec & add_cmd_output method to accept a list. * Log directory "/var/log/landscape-server" is now fully deprecated. Nowadays, all supported versions are logging inside "/var/log/landscape". Resolves: #2012 Signed-off-by: Eric Desrochers <eric.desrochers@canonical.com> Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [global] fix some flake8 issuesBryan Quigley2020-04-144-9/+6
| | | | | | | | | | Mostly unused import or missing variables. Including 1 mistyped variable by me. Resolves: #2011 Signed-off-by: Bryan Quigley <bryan.quigley@canonical.com> Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [global] Drop use of python six and other 2.X bitsBryan Quigley2020-04-0917-120/+40
| | | | | | Resolves: #2006 Signed-off-by: Bryan Quigley <bryan.quigley@canonical.com>
* [global] Drop use of __future__Bryan Quigley2020-04-088-17/+0
| | | | | | | | | | This was used either to * get the print function * get the with function (for py2.5) Resolves: #2005 Signed-off-by: Bryan Quigley <bryan.quigley@canonical.com>
* [build] Update setuptools for new package layout, remove python2 testsJake Hunsaker2020-04-083-12/+4
| | | | | | | | | | | | | Updates setup.py to reflect project changes with 4.0, thus allowing the 'python setup.py install' checks to run normally. Removes futures from requirements.txt, as futures is packaged with python 3.4 and newer, and sos no longer supports python2. In that vein, also removes the pytohn-2.7 test from the travis configuration. Resolves: #2004 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [options] Remove extraneous importJake Hunsaker2020-04-081-1/+1
| | | | | | Removes an extraneous import identified by LGTM Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [travis] Update test script for new binaryJake Hunsaker2020-04-081-9/+9
| | | | | | | Updates the test script `simple.sh` used by Travis to use the new `sos` binary with the `report` subcommand. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [global] PEP8 fixesJake Hunsaker2020-04-084-11/+13
| | | | | | Consolidated PEP8 fixes following reorganization and redeisgn work. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [binary] Allow subcommand alias referencesJake Hunsaker2020-04-081-6/+17
| | | | | | | | | | Adds support to the parser to allow aliases for subcommands in the parser. At the moment, this only applies to `report`, with the 'rep' shorthand. This is however intended for further 4.0 use via the integration of sos-collector and soscleaner. For example, invoking 'sos clean' and 'sos cleaner' would both be possible. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [reporting] Move reporting under sos/reportJake Hunsaker2020-04-083-2/+2
| | | | | | Moves the HTML/JSON reporting bits under sos/report. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [binary] Have legacy sosreport binary redirect to sosJake Hunsaker2020-04-083-6/+16
| | | | | | | | | | | | | | | | | | We need to carry a `sosreport` binary for some time to allow for end users and downstreams to adjust to the new binary. It was hoped originally that the old `sosreport` binary could be maintained in place and simply provide the older set of functionality. This has proven to not be possible givent he overhaul of the options handling that allows us to have multiple subcommands. So while we will still ship an `sosreport` binary, and it will be locked to `report` functionality, it is now a simple redirection script that also makes the user aware of the new `sos` binary. Closes: #1986 Resolves: #1993 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [binary] Set python3 as the interpreterJake Hunsaker2020-04-081-1/+1
| | | | | | Since sos-4.0 is py3 only, set the interpreset of the binary as such Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [report] Add a description for help outputJake Hunsaker2020-04-081-8/+5
| | | | | | | Adds the `desc` class attr to SoSReport() for use by the help text when `sos --help` is run. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [options] Ensure verbosity is an intJake Hunsaker2020-04-081-1/+1
| | | | | | | When loading from a config file, ensure that verbosity is set to and integer rather than a string. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [report] Remove extra 'report' string from log messageJake Hunsaker2020-04-081-1/+1
| | | | Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [report] Don't override self.sys_tmp in __init__()Jake Hunsaker2020-04-081-1/+0
| | | | | | | | Removes an extraneous setting of self.sys_tmp, which is now handled via `SoSComponent`. This was getting overridden to `None` which was causing the final archive creation to fail. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [options] Cast arg_names to list instead of dict_keysJake Hunsaker2020-04-081-1/+1
| | | | | | | For python3, calling `.keys()` on a dict no longer returns a list, so we need to explicitly make it a list of values to behave as expected. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [report] Update SoSReport() to use SoSComponent()Jake Hunsaker2020-04-083-328/+197
| | | | | | | | | | | Moves SoSReport() from a standalone class to a subclass of SoSComponent() in order to take advantage of the new 4.0 redesign. This commit also pulls out methods that were previously transitioned to other places, e.g. logging setup moving from SoSReport() to the base SoSComponent() class. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [redhat] Fix SoSOptions importJake Hunsaker2020-04-081-1/+1
| | | | Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [options] Drop explicit subclass of objectJake Hunsaker2020-04-081-1/+1
| | | | | | | | | For the 3.x line, many classes subclassed `object` directly, in order to use the 'new-style' objects introduced with python3, while still maintaining python2 functionality. As sos-4.0 is python3 only, we no longer need to do this. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [utilities] Move _sos back into sos/__init__.pyJake Hunsaker2020-04-083-13/+11
| | | | | | | | Moving the `_sos_ definition wrapper for gettext into utilities proved to be the wrong direction, so this moves it back into sos/__init__.py in order to preserve the ability to properly import it elsewhere. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [component] Add Policy loading and move to new component.pyJake Hunsaker2020-04-082-144/+172
| | | | | | | | | | | | | Adds Policy loading into `SoSComponent()` so that any subcommand can, if it needs it, have policies loaded consistenly. This is controlled via the `load_policy` class attr, which defaults to True. Future subcommands that do not need to have a policy loaded (E.G. the planned `info` command), can simply set this to False in order to skip this step. Moves `SoSComponent()` out of sos/__init__.py and into sos/component.py to ease import conflicts arising from policy loading. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
generated by cgit (git 2.34.1) at 2024-11-14 23:49:40 +0000