Pulp: Issueshttps://pulp.plan.io/https://pulp.plan.io/favicon.ico2021-06-04T12:09:35ZPulp
Planio RPM Support - Issue #8864 (CLOSED - CURRENTRELEASE): Workers go OOM while trying to sync RHEL 7https://pulp.plan.io/issues/88642021-06-04T12:09:35Zehelms@redhat.comehelms@redhat.com
<p>Setup:</p>
<ul>
<li>CentOS 7 Vagrant box</li>
<li>Memory: 8GB</li>
<li>Running Katello w/ Pulp 3.11</li>
<li>2 workers present</li>
</ul>
<p>When I attempt to sync RHEL 7 Server x86_64, it fails every time with Pulp workers going OOM. At the time when I initiate the sync there is 3.5GB of available memory on the VM.</p> CertGuard - Issue #8785 (CLOSED - DUPLICATE): CONTRIBUTING docs are out of date for how to test p...https://pulp.plan.io/issues/87852021-05-19T19:45:14Zehelms@redhat.comehelms@redhat.com
<p><strong>Ticket moved to GitHub</strong>: "pulp/pulp-certguard/144":<a href="https://github.com/pulp/pulp-certguard/issues/144" class="external">https://github.com/pulp/pulp-certguard/issues/144</a></p> CertGuard - Issue #8783 (CLOSED - CURRENTRELEASE): CA certificate does not handle a CA bundle or ...https://pulp.plan.io/issues/87832021-05-19T15:38:18Zehelms@redhat.comehelms@redhat.com
<p>If I set the value of ca_certificate (e.g. update certguard_rhsmcertguard SET ca_certificate = :'content' ;) where that content is a bundle of CA certificates, clients will receive a 403 unless the CA that signed the client certificate is the first CA in the file.</p> Pulp - Task #4554 (CLOSED - CURRENTRELEASE): Change naming of Pulp 3 services to differentiate th...https://pulp.plan.io/issues/45542019-03-20T14:09:20Zehelms@redhat.comehelms@redhat.com
<p>When Pulp 2 and Pulp 3 are ran on the same system differentiating the services is difficult which affects the ability to debug, restart the proper services and overall manage both applications during transition. Based on some pulp-dev email conversations one of two options should be taken:</p>
<p><a class="issue tracker-3 status-11 priority-6 priority-default closed child" title="Story: As a user, I can have Pulp attempt use auto_retry application wide using the 'unsafe_autoretry' p... (CLOSED - CURRENTRELEASE)" href="https://pulp.plan.io/issues/1">#1</a>: Rename Pulp 3 services to all new names to clearly differentiate</p>
<p>For example:</p>
<pre><code>a) pulp-task-manager
b) pulp-tasking-manager
c) pulp-queue-manager
d) pulp-arborist
e) arborist
f) pulp-task-runner
g) pulp-task-worker
h) pulp-async-worker
i) pulp-tasking-worker
j) pulp-seedling
</code></pre>
<p><del><a class="issue tracker-3 status-11 priority-7 priority-high2 closed" title="Story: As a user, my rpm sync finishes quickly when upstream metadata hasn't changed (CLOSED - CURRENTRELEASE)" href="https://pulp.plan.io/issues/2">#2</a>: Rename Pulp 2 services to include version</del></p>
<pre><code>-a) pulp2-resource-manager
b) pulp2-worker-
</code></pre>
<p><del>As per <a href="https://www.redhat.com/archives/pulp-dev/2019-March/msg00085.html" class="external">this thread</a>, it's decided to go with option 2 - rename pulp2 services.</del></p>
<p>As per <a href="https://www.redhat.com/archives/pulp-list/2019-June/msg00008.html" class="external">this thread</a>, it's decided to go with option 1 a rename pulp3 services to pulpcore-*</p> Pulp - Issue #4508 (CLOSED - CURRENTRELEASE): Remove '3' from Pulp installer and roleshttps://pulp.plan.io/issues/45082019-03-06T18:04:33Zehelms@redhat.comehelms@redhat.com
<p>The current installer and roles for Pulp 3 are tied to the pulp3 version disallowing flexibility for semantic versioning of Pulp and the installer going forward. I would recommend dropping the '3',e.g. `ansible-pulp` and `pulp` and `pulp-redis` for role names.</p>
<p>Further, current variables used inside the roles are not following best practices for variables being namespaced by the role name. Dropping the '3' would align these variables to best practices.</p> Pulp - Task #4495 (CLOSED - CURRENTRELEASE): Rename pulp-wsgi to be more descriptivehttps://pulp.plan.io/issues/44952019-03-04T15:59:35Zehelms@redhat.comehelms@redhat.com
<p>The rest of Pulp services describe their function while pulp-wsgi describes the interface used to serve the application. I think it would be more useful to rename this service to describe the functionality. Some ideas:</p>
<ul>
<li>pulp-core</li>
<li>pulp-api</li>
<li>pulp-api-server</li>
</ul> Pulp - Issue #3883 (CLOSED - CURRENTRELEASE): Switch default logging to consolehttps://pulp.plan.io/issues/38832018-07-25T15:22:28Zehelms@redhat.comehelms@redhat.com
<p>A proposal was submitted [1] to pulp-dev and given the traction I am filing an issue to switch from syslog to console by default for Pulp 3.</p>
<p>[1] <a href="https://www.redhat.com/archives/pulp-dev/2018-July/msg00095.html" class="external">https://www.redhat.com/archives/pulp-dev/2018-July/msg00095.html</a></p> Pulp - Issue #3879 (CLOSED - CURRENTRELEASE): Cannot opt out of settings defaults due to merging ...https://pulp.plan.io/issues/38792018-07-23T16:50:08Zehelms@redhat.comehelms@redhat.com
<a name="Issue"></a>
<h2 >Issue<a href="#Issue" class="wiki-anchor">¶</a></h2>
<p>The pointed issue is that if I want to configure pure console logging (for container environments) I cannot. The syslog handler gets configured and attempts to connect failing. If I try to set my logging configuration like the following, the syslog handler I explicitly try to exclude is added:</p>
<pre><code>logging:
formatters:
simple:
format: "pulp: %(name)s:%(levelname)s: %(message)s"
handlers:
console:
class: logging.StreamHandler
formatter: simple
loggers:
'':
handlers: ["console"]
level: INFO
</code></pre>
<p>I have had to hack around this for now via -- <a href="https://github.com/ehelms/pulp/commit/8e82d27a6294c2db139ce2ffc41002cc70309777" class="external">https://github.com/ehelms/pulp/commit/8e82d27a6294c2db139ce2ffc41002cc70309777</a></p>
<p>How can I only use my settings and not "merge" with the defaults.</p>
<a name="Solution"></a>
<h2 >Solution<a href="#Solution" class="wiki-anchor">¶</a></h2>
<p>Update the behavior to be an 'overwrite' from the top level of each. The merge could would only replace user values for default values for the keys: DATABASES, logging, ..., PROFILING</p>
<p><a href="https://github.com/pulp/pulp/blob/12919c2040b3de1a061e647f1f59e7494294d600/pulpcore/pulpcore/app/settings.py#L219-L240" class="external">https://github.com/pulp/pulp/blob/12919c2040b3de1a061e647f1f59e7494294d600/pulpcore/pulpcore/app/settings.py#L219-L240</a></p> Pulp - Issue #3507 (CLOSED - WONTFIX): Pulp streamer should be able to log to consolehttps://pulp.plan.io/issues/35072018-03-21T00:07:44Zehelms@redhat.comehelms@redhat.comRPM Support - Story #3091 (CLOSED - CURRENTRELEASE): As a user, I can create a manifest for the f...https://pulp.plan.io/issues/30912017-10-20T18:39:47Zehelms@redhat.comehelms@redhat.com
<p>This is a request to provide a supported command line tool, installable as an RPM, to generate PULP_MANIFEST files in the right format for a directory structure. This should handle recursive generation as well as being able to re-run it and generate an updated PULP_MANIFEST.</p>
<p>There are a few examples of this in use today:</p>
<p>1. <a href="http://docs.pulpproject.org/plugins/pulp_rpm/tech-reference/iso-plugins.html#iso-importer" class="external">http://docs.pulpproject.org/plugins/pulp_rpm/tech-reference/iso-plugins.html#iso-importer</a><br>
2. <a href="https://gist.github.com/jlsherrill/321cf092fce13d1c4f3a" class="external">https://gist.github.com/jlsherrill/321cf092fce13d1c4f3a</a><br>
3. <a href="https://gist.github.com/jlsherrill/e7c72e1ed82379955c2208ac472b0be7" class="external">https://gist.github.com/jlsherrill/e7c72e1ed82379955c2208ac472b0be7</a></p>
<p>Rather than have floating around tools, users would benefit from a simple, supported tool from the Pulp project that properly generates what Pulp requires to sync a repository.</p>
<p>Requirements for the tool:<br>
- name = 'pulp-manifest'<br>
- it takes 1 positional argument that specifies a path to the directory where the PULP_MANIFEST file should be created<br>
- it creates PULP_MANIFEST file by walking the directory structure recursively and recording information about files in the CSV format: file_name,checksum,file size<br>
- if a PULP_MANIFEST already exists, it should be regenerated<br>
- pulp-rpm.spec file should be updated to have this tool installable as a separate RPM.</p> Pulp - Issue #2586 (CLOSED - NOTABUG): Pulp workers/beat/resource_manager go missinghttps://pulp.plan.io/issues/25862017-02-13T20:52:22Zehelms@redhat.comehelms@redhat.com
<p>Running on Pulp 2.10.3, I see the following error messages running this setup:</p>
<p>clone <a href="https://github.com/Katello/forklift" class="external">https://github.com/Katello/forklift</a><br>
vagrant up centos7-katello-nightly</p>
<p>If you do not see the error messages initially:</p>
<pre><code>vagrant ssh centos7-katello-nightly
cd /
/vagrant/forklift/bats/bootstrap_vagrant.sh
cd /vagrant/forklift/bats
bats fb-install-katello.bats fb-content-katello.bats
</code></pre>
<pre><code>Feb 13 20:45:22 centos7 pulp: pulp.server.async.scheduler:ERROR: There are 0 pulp_resource_manager processes running. Pulp will not operate correctly without at least one pulp_resource_mananger process running.
Feb 13 20:46:22 centos7 pulp: pulp.server.async.scheduler:ERROR: There are 0 pulp_resource_manager processes running. Pulp will not operate correctly without at least one pulp_resource_mananger process running.
Feb 13 20:47:22 centos7 pulp: pulp.server.async.scheduler:ERROR: There are 0 pulp_resource_manager processes running. Pulp will not operate correctly without at least one pulp_resource_mananger process running.
Feb 13 20:48:22 centos7 pulp: pulp.server.async.scheduler:ERROR: There are 0 pulp_resource_manager processes running. Pulp will not operate correctly without at least one pulp_resource_mananger process running.
Feb 13 20:49:22 centos7 pulp: pulp.server.async.scheduler:ERROR: There are 0 pulp_resource_manager processes running. Pulp will not operate correctly without at least one pulp_resource_mananger process running.
Feb 13 20:50:22 centos7 pulp: pulp.server.async.scheduler:ERROR: There are 0 pulp_resource_manager processes running. Pulp will not operate correctly without at least one pulp_resource_mananger process running.
</code></pre> Pulp - Issue #2433 (CLOSED - WONTFIX): Re-uploading results in old file still being servedhttps://pulp.plan.io/issues/24332016-11-21T18:21:55Zehelms@redhat.comehelms@redhat.com
<p>Using the iso type of repository, a user can upload the same file multiple times. However, the 'last' one to be uploaded is the not the current one being served by the repository. The generated PULP_MANIFEST indicates the various versions of this file that have been uploaded but they appear in semi-random order. An example:</p>
<pre><code>foreman-1.13.1.1.tar.bz2,c2a1985f371a8152b96c8927511e7402d2bbf64e2acb084470ec462df97b1bcc,11582712
foreman-1.13.1.1.tar.bz2,aa7bdcc0c30e2a843a1f74672292d04231749d5b8859e41b597129826ecc8f96,29676401
foreman-1.13.1.1.tar.bz2,f217c027bde37e426ccd8912b48c2f817621b7586359317cd338d5476c2f8a58,11586229
foreman-1.13.1.1.tar.bz2,2b5fc81f743de034eb38232a3d60467b3b71e819c84729a2fdcfedbfd1e3ae86,11583659
foreman-1.13.1.1.tar.bz2,4cde275645bafe5b46448925dec1c6b295119bcd4b4b4ab76cc04f38109f2398,29859413
foreman_theme_satellite-1.0.0.gem,03f8ac79b51cb68f9308249396881a827d5b81840bed58bf466133cef6c37120,158720
</code></pre>
<p>Further, this behavior appears with the YUM repository type as well. A user can:</p>
<pre><code>1) sync a repository
2) upload an RPM that was synced down with a different checksum (maybe the underlying source changed)
3) Pulp will accept the upload with success
4) check the yum metadata and note both checksums in it but only one copy of the RPM in the repo (this will lead to yum errors)
</code></pre> Pulp - Issue #1948 (CLOSED - CURRENTRELEASE): Upgrading RPMs on EL6 sometimes fails during pre sc...https://pulp.plan.io/issues/19482016-05-26T13:52:14Zehelms@redhat.comehelms@redhat.com
<pre><code>Steps to Reproduce:
1. # yum update
...
Updating : python-nectar-1.5.1-3.el6sat.noarch 227/420
error: %pre(pulp-server-2.8.1.3-1.el6sat.noarch) scriptlet failed, exit status 1
Error in PREIN scriptlet in rpm package pulp-server-2.8.1.3-1.el6sat.noarch
error: install: %pre scriptlet failed (2), skipping pulp-server-2.8.1.3-1.el6sat
Updating : pulp-docker-plugins-2.0.0.2-1.el6sat.noarch 229/420
...
pulp-server-2.6.0.20-1.el6.noarch was supposed to be removed but is not!
</code></pre> RPM Support - Issue #1775 (CLOSED - CURRENTRELEASE): Content removed from a repository never returnshttps://pulp.plan.io/issues/17752016-03-15T18:12:21Zehelms@redhat.comehelms@redhat.com
<p>I am seeing the following:</p>
<p>1. Create a repository with a feed URL<br>
2. Sync the repository<br>
3. Check the content counts for the repository<br>
4. Remove 1 or more packages from the repository<br>
5. Re-sync the repository</p>
<p>The packages that were removed from the repository do not re-appear in the repository. Meaning that if I remove all packages from the repository and then re-sync it the repository stays empty and I am unable to recover.</p> Python Support - Issue #1131 (CLOSED - CURRENTRELEASE): Documentation references 'packages_names'...https://pulp.plan.io/issues/11312015-07-09T19:02:59Zehelms@redhat.comehelms@redhat.com
<p>Should be 'package_names'</p>