-
Notifications
You must be signed in to change notification settings - Fork 595
[upload] Fix for anonymous upload duplicate directory issue #4156
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: main
Are you sure you want to change the base?
Conversation
Uploads of files greater than 1Gb switch to go to sftp. The
code incorrectly tried to write to the home directory in sftp,
but should switch into the /{user} directory instead so
proper permissions exist.
Related: RHEL-52919
Signed-off-by: David Wolstromer <dwolstro@redhat.com>
Allow running cleaner concurrently via child processes. They synchronize on the ordering of items added to dataset of each mapper by creating numbered files in a directory specific for each mapper. Together with deterministic generation of obfuscated values, this ensures the individual processes end up with identical mappings. Resolves: sosreport#3097 Closes: sosreport#3988 Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
Capture /etc/services file Related: RHEL-105407 Signed-off-by: Jose Castillo <jcastillo@redhat.com>
This Update allows a specific container runtime to be passed to add_cmd_output so commands that must use a specific runtime do so. Related: SUPDEV-180 Signed-off-by: David Wolstromer <dwolstro@redhat.com>
WHile plugin execution makes sense only when mdadm package is installed. Closes: sosreport#4095 Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
The command apt-mark showhold is useful to show what packages have been marked as 'held' in apt, as this can lead to a variety of issues. Signed-off-by: Daniel Emmons <dan.emmons@canonical.com>
Modify the function add_dir_listing() to accept a new argument, 'sparse', that will enable the use of the option '-s', to list the allocated size of a file in blocks. Related: RHEL-107275 Signed-off-by: Jose Castillo <jcastillo@redhat.com>
Capture the block usage of files inside /var/log to show the size of spare files in that directory. Related: RHEL-107275 Signed-off-by: Jose Castillo <jcastillo@redhat.com>
These are large and unnecessary in most cases: forbid all *.img files in the /boot/boom/cache directory allowing the plugin to collect all the text files under /boot/boom that are actually useful. Related: sosreport#4098 Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
Grab the output of 'boom cache list': it's more readable than the cacheindex.json metadata file and displays the status of cached files. Resolves: sosreport#4098 Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
Add a plugin for the snapm snapshot manager tool: https://github.com/snapshotmanager/snapm Collect: * /etc/snapm configuration directory * snapm snapset list * snapm snapshot list * snapm plugin list * snapm schedule list * snapm -vv --debug=all snapset show N.b: the /etc/snapm directory and the 'snapm schedule list' command are part of the v0.5.0 release currently in testing. Earlier releases lack these files/subcommands: https://github.com/snapshotmanager/snapm/releases/tag/v0.5.0 Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
This commit marks the release of sos-4.10.0. Signed-off-by: Jake Hunsaker <jacob.r.hunsaker@gmail.com>
Improve postproc obfuscation in two ways: - apply postproc also to /var/lib/openstack/config/nova on RedHatNova - obfuscate just password from transport_url, not the whole URL Closes: sosreport#4108 Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
The fix is three-fold: - obfuscate_file must clean file content and not filename - cleaner's main_archive must be populated by parsers first - obfuscate_file dont need short_name as it is always called with implicit value of short_name that cleaner will strip itself Closes: sosreport#4109 Closes: sosreport#4110 Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
Signed-off-by: Mateusz Kulewicz <mateusz.kulewicz@canonical.com> [loki] Collecting Loki logs is optional [loki] Pass labels as option Signed-off-by: Mateusz Kulewicz <mateusz.kulewicz@canonical.com> [loki] Get logs in separate method Signed-off-by: Mateusz Kulewicz <mateusz.kulewicz@canonical.com> [loki] Logs pagination Signed-off-by: Mateusz Kulewicz <mateusz.kulewicz@canonical.com> [loki] Configurable endpoint [loki] Formatting Signed-off-by: Mateusz Kulewicz <mateusz.kulewicz@canonical.com> [loki] Constants, code robustness Signed-off-by: Mateusz Kulewicz <mateusz.kulewicz@canonical.com> [loki] Debugging Signed-off-by: Mateusz Kulewicz <mateusz.kulewicz@canonical.com>
Signed-off-by: Nagoor Shaik <nshaik@redhat.com>
Add lsslot commands to collect information about hot-plug capable PCI and PHB slots and their resources. The output from the added commands is generally only a few KBs. Signed-off-by: Sourabh Jain <sourabhjain@linux.ibm.com>
Collect detailed disk usage of Podman resources, including images, containers, and volumes. Signed-off-by: Sourabh Jain <sourabhjain@linux.ibm.com>
Capture the /etc/crypttab file Related: RHEL-108949 Signed-off-by: Jose Castillo <jcastillo@redhat.com>
Related: RHEL-103783 Signed-off-by: David Wolstromer <dwolstro@redhat.com>
Closes: sosreport#4119 Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
When the command 'rhc status' runs in the plugin rhc, it starts automatically the rhsm service. This commit gates the command to make sure it's not run when the rhssm service is not running. Related: RHEL-112563 Signed-off-by: Jose Castillo <jcastillo@redhat.com>
Spyre chip is IBM’s AI accelerator, designed to handle AI inferencing and workloads. The Spyre plugin collects data about the Spyre card’s VFIO device node tree, configuration files, and more. Signed-off-by: Sourabh Jain <sourabhjain@linux.ibm.com>
We have been seeing some odd failures with the CI around Fedora, particularly with the rpmbuild failing which then cancels the rest of the suite due to dependency resolution. We're working with Cirrus support, but until we have a fix let's disable the Fedora testing to unblock the rest of the suite. Signed-off-by: Jake Hunsaker <jacob.r.hunsaker@gmail.com>
We need to add the netmask to the string to obfuscate, i.e. to the first group of matched r.e. Otherwise, '2620:52:0:25c0::/64' like string is obfuscated to '534f:53ff:fe00:0002::/64/64'. Closes: sosreport#4124 Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
57bbc89 commit set tmpdir to source the dir from Policy. Which means HOST sysroot directory is newly applied already in LinuxPolicy._container_init method. Removed lines mimic the same in a worse way, so let drop them here. Resolves: sosreport#4116 Closes: sosreport#4118 Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
The formatting of the distros column was leaving the string 'Plu' when it found 'PluginOpt'. This fix explicitly deals with both 'PluginOpt' and 'Plugin' strings and removes them. Signed-off-by: Jose Castillo <jcastillo@redhat.com>
* Use latest version of Foreman * Test using Ubuntu 22.04, Debian 12, and CentOS Stream 9 Related: sosreport#4094 Co-authored-by: Pavel Moravec <pmoravec@redhat.com> Signed-off-by: Arif Ali <arif-ali@ubuntu.com>
This will allow containers = (..) to be checked against given runtime, if it is not default one. Related: sosreport#4123 Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
OpenShift OVN has both podman and crio runtimes and sos detects podman as the default one. While the containers triggering openshift_ovn enabledness are under 'crio' runtime. Closes: sosreport#4123 Relevant: RHEL-110512 Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
Signed-off-by: Javier Blanco <javier@jblanco.es>
Evident typo. Closes: sosreport#4136 Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
Signed-off-by: Arif Ali <arif-ali@ubuntu.com>
Resolves: sosreport#4129 Closes: sosreport#4132 Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
The default value for networking.namespaces in the file tests/report_tests/options_tests/options_tests_sos.conf is 100, while the value for this variable in tests/report_tests/options_tests/options_tests_preset.json is 200. To make sure that preset beats the config file we need to change the value from 100 to 200. Related: sosreport#4087 Signed-off-by: Jose Castillo <jcastillo@redhat.com>
RE match does not match /var/log/juju/unit-nova-compute-*.log* to anything. Closes: sosreport#4143 Relevant: sosreport#4129 Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
This is now automated in Launchpad, and not required upstream. The build is done on a daily basis now, rather than on every commit. Signed-off-by: Arif Ali <arif-ali@ubuntu.com>
The LTS releases are the ones that are supported for a long time, so these won't change for a long time. It makes sense to have the versions of these visible, rather than just changing to prior or prior2. This can get confusing when updating or adding new versions. Signed-off-by: Arif Ali <arif-ali@ubuntu.com>
Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
Signed-off-by: Deezzir <yurii.kondrakov@canonical.com>
* Config option messaging_urls was not detected in ceilometer. * Revamp the regex for this plugin to potentially detect most scenarios. * Revamp other plugins to have better regex, to potentially pick up others. Signed-off-by: Arif Ali <arif-ali@ubuntu.com>
Signed-off-by: Pablo Fernández Rodríguez <pafernan@redhat.com>
Signed-off-by: Nagoor Shaik <nshaik@redhat.com> EDA Version in AAP 2.4 uses environment file to store configuration ex: /etc/ansible-automation-platform/eda/environment EDA version AAP 2.5 and above uses a YAML file to store config ex: /etc/ansible-automation-platform/eda/settings.yaml
Substitute the current 'domain-status -o' command with 'sssctl domain-status $domain' which shows list of of available AD/IPA servers as well along with online/offline status for the specified domain. Also include output of command sssctl access-report $domain Related: RHEL-108358 Signed-off-by: Jose Castillo <jcastillo@redhat.com>
The /etc/lnet.conf file can have configured lnet settings for a lustre server or client. Signed-off-by: Nathaniel Clark <nclark@whamcloud.com>
Forgetting to provide mandatory username of the AAP containerized deployment leads to skipping plugin completely and lack of collected data. Let add a heuristic "is there a sole user running some podman container?" to detect the user in such a case. Closes: sosreport#4150 Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
with fix for uploads greater than 1 GB Related: RHEL-52919 Signed-off-by: David Wolstromer <dwolstro@redhat.com>
|
Congratulations! One of the builds has completed. 🍾 You can install the built RPMs by following these steps:
Please note that the RPMs should be used only in a testing environment. |
TurboTurtle
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It's not clear to me what the purported issue this is intended to fix is, or how it actually goes about fixing it. An actual commit message would go a long way here.
Additionally, you'll need to rebase your changed on current main, so that this branch doesn't attempt to pull in 50 other preceding commits again.
Collects dump files on Debian when the following conditions are met: - The dump file is not older than 24 hours. - The dump file size is not greater than 2 GB. - Plugin specific option kdump.get-vm-core=true is passed Additionally, it collects the dmesg files, the kexec_cmd file, and the linux-image file present in the dump target configured in the kdump configuration file /etc/default/kdump-tools. Signed-off-by: Sourabh Jain <sourabhjain@linux.ibm.com>
|
THIS is the key commit: but I agree, you need to rebase your branch. Do I get it right that the 4003238 "Enable sos uploads of files bigger than 1Gb" has some side-effect or missing piece? I miss the description of the error you aim to fix /o. |
|
I do approve the code change (when SFTP-ing to Red Hat SFTP server under anonymous user, a randomly generated on-time username is used that does not have home. In that case, we cant upload to its home but to default dir). Please squash the commits into one, or create a new branch based on current upstream, and apply there just the patch: |
|
This is still not squashed, and still has 55 commits; we ought to be seeing only 1 commit. Once that is done, this can be reviewed and work on |
Please place an 'X' inside each '[]' to confirm you adhere to our Contributor Guidelines