python / planet

Configuration for Python planets (e.g. http://planetpython.org)
133 stars 173 forks source link

Posts not displaying at planetpython.org #519

Open dsuch opened 2 years ago

dsuch commented 2 years ago

Hello everyone,

I post articles to our Zato Blog - I link to it below - and I noticed that Planet Python did not publish the latest one.

The RSS feed is valid, I can see the Planet's robot access it, but somehow the post does not display over at https://planetpython.org.

Can you please check it in the Planet's logs and help me out in resolving it? Is there perhaps anything that I can do about it?

https://zato.io/blog/ https://zato.io/blog/rss.xml https://zato.io/blog/posts/ssh-api-service.html

Thank you.

dsuch commented 2 years ago

Hello again,

I am very much interested in resolving this situation.

Can any of the Planet's maintainers advise what I can do to help you out with it?

Thank you.

matrixise commented 2 years ago

I will check asap. Thank you for your feedback.

dsuch commented 2 years ago

Hello @malemburg,

per your messages from the Python forum, can you please help me and @matrixise in resolving the situation from this ticket?

Thank you.

malemburg commented 2 years ago

I guess I would need to have a look at the system log files for this, but don't have access. The posts listed on the website are only shown until Aug 2, so your latest post on the blog does not show up anymore due to the cut-off logic.

dsuch commented 2 years ago

Thanks @malemburg - I changed the date of the latest post to today, let's see in a couple of hours if that works.

Since I have your attention, I would like to add that what we are seeing here - particularly in the #521 spam links ticket - is a maintenance problem.

About a year ago I contacted planet@python.org with a similar issue, which was ultimately resolved by Tres Seaver. I offered some of the time of one of my employees to help clean everything up but there was no follow up from any of the Planet's maintainers.

Now, some of the tickets wait for weeks or months without any reply, which is unrelated to the fact that it is summertime now. The spammer from #521 kept the entire communications channel blocked for one week.

When it came to removing it, no single person really knows what to do, where to click, where the server really is (I only heard a while ago that it is somewhere in the Digital Ocean cloud), who has access or who can grant access.

This is what we can easily call a maintenance problem. No single person who owns the space, the distributed, community-based model notwithstanding.

All of this is perfectly understandable, people come and go, they have time or they do not, nothing unusual.

At the same time, in a situation like this one, it became obvious that there are other people, like @The-Judge, @sandrotosi or @pamoroso who would like to help.

I really suggest that new people be given access to the server and GitHub - I am not part of any core team but obviously you are (@malemburg) and I am sure that your voice will be heard.

I am also sure that @The-Judge and others would have more ideas about what to do and how.

Thank you.

malemburg commented 2 years ago

I only stopped by since I saw the topic on Discourse. I'm not part of the Planet team, nor do I know who owns the project. From looking at the code history, I would assume Stéphane (@matrixise) does. I also don't have admin access to the repo, just merge permissions and can't even see the list of other people having similar rights (which is a Github thing, not something someone at the PSF intended).

I can ping Stéphane and see what he thinks to perhaps get the ball rolling to extend the team.

That said, my experience with such projects is that when something needs attention, many people jump forward offering help, but this rarely helps projects in the longer term. After a short initial run, things usually go back to mostly the same people who had maintained the project before, since maintaining such projects is not necessarily fun, but mostly just plain work. You need people with a passion for the project to sustain it.

malemburg commented 2 years ago

Back on topic: The posts don't appear to show up on the blog for some reason, but in order to find out why, we'd need access to logs (I don't have access to the VM running Planet).

malemburg commented 2 years ago

I ran Planet locally (in Docker) and this is the output from your blog:

planet_1  | INFO:planet:Updating feed <https://zato.io/blog/rss.xml>
planet_1  | DEBUG:planet:E-Tag: "62f13ffc-50160"
planet_1  | DEBUG:planet:Last Modified: 2022-08-08T16:55:24+00:00
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/wordpress-webhooks.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/confluence-python-api.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/web-scraping-api-integrations.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/jira-python-api.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/salesforce-credentials.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/salesforce-python.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/zato-architecture-primer.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/vscode-remote-api-debugging.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/zato-dev-workflow.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/windows-api-integrations.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/scalable-api-ai-architectures.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/howto-rest-api-client-cert.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/soap-wsdl.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/zato-version-cli.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/websocket-timeouts.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/pubsub-api-service.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/sql-config-details.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/config-env-variables.html> as hidden (new feed)
planet_1  | DEBUG:planet:Marked <https://zato.io/blog/posts/ibm-mq-python-api-integrations.html> as hidden (new feed)
planet_1  | DEBUG:planet:Items in Feed: 20

The SSH post is not marked as hidden and does show up on the local webserver, so things do appear to work fine when run locally.

matrixise commented 2 years ago

@ewdurbin I think you have the access to the planet server, what's the log about this issue? I just use my commit flag for the PRs, but I can't work on the server.

matrixise commented 2 years ago

Already tested locally, and @malemburg has right, except the SSH post, the other ones are hidden.

to test locally

# remove the unused RSS feeds in config/config.ini
docker compose build
docker compose up
dsuch commented 2 years ago

Thank you both @matrixise and @malemburg although this still persists in the live environment.

Just to make sure that it is not anything with that one particular post about SSH, about 24 hours ago I changed the date of one of the older posts that some time ago was republished by the Planet to the yesterday's.

Sadly, it was not republished yesterday either.

https://zato.io/blog/posts/api-rate-limiting-intro.html

ewdurbin commented 2 years ago

@matrixise and @malemburg, glad to see someone is interested in keeping planet's code working. PSF-Infra really just monitors to keep the host online and running :)

The only information in log of running sudo -u planet /srv/run-planet.sh

is

INFO:planet:Feed <https://zato.io/blog/rss.xml> unchanged

@matrixise please open a PR to add yourself to https://github.com/python/psf-salt/blob/main/pillar/base/users.sls using the syntax noted, the only modifier is that you'll want

      planet:
        sudo: True
        allowed: True

as part of your entry, this will grant you access to planet.nyc1.psf.io


@malemburg your access has been added with https://github.com/python/psf-salt/pull/280

ewdurbin commented 2 years ago

after running

$ sudo rm /srv/cache/zato.io,blog,rss.xml
$ sudo -u planet /srv/run-planet.sh

I see output for zato as:

INFO:planet:Updating feed <https://zato.io/blog/rss.xml>
DEBUG:planet:E-Tag: "62f4daaa-4d623"
DEBUG:planet:Last Modified: 2022-08-11T10:32:10+00:00
DEBUG:planet:Marked <https://zato.io/blog/posts/ssh-api-service.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/wordpress-webhooks.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/confluence-python-api.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/web-scraping-api-integrations.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/jira-python-api.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/salesforce-credentials.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/salesforce-python.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/zato-architecture-primer.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/vscode-remote-api-debugging.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/zato-dev-workflow.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/windows-api-integrations.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/scalable-api-ai-architectures.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/howto-rest-api-client-cert.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/soap-wsdl.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/zato-version-cli.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/websocket-timeouts.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/pubsub-api-service.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/sql-config-details.html> as hidden (new feed)
DEBUG:planet:Marked <https://zato.io/blog/posts/config-env-variables.html> as hidden (new feed)
DEBUG:planet:Items in Feed: 20

This indicates that for whatever reason the e-tag or last-modified date is not being updated appropriately by the zato.io server, or recognized correctly by planet.

ewdurbin commented 2 years ago

I do see Understanding API rate-limiting techniques on the feed now though.