metacpan-web
metacpan-web copied to clipboard
metacpan
Recent page appears stuck
Describe the bug The "Recent" page at https://metacpan.org/recent is not showing any new releases since Friday. I know there has been at least one as I made one yesterday.
To Reproduce Steps to reproduce the behavior:
- Go to https://metacpan.org/recent
- See that there are no new releases since Friday when ordinarily there are maybe 20 per day.
Expected behavior I expect to see the releases made since Friday.
Screenshots
Desktop (please complete the following information):
- OS: Linux
- Browser Pale Moon
- Version 33.0.0 (64-bit)
Looking around, it seems that this is a symptom of a wider problem. My dist released yesterday (Password-Policy-Rule-Pwned-0.02_02.tar.gz) is not recorded anywhere within MetaCPAN AFAICS. :-/
PAUSE also had problems accepting new uploads on Friday and Saturday, but this seems to be fixed since Sunday.
It would be helpful for users if a short alert about the issues with PAUSE were added to the page.
MetaCPAN has fully tipped over.
I was checking to see if it caught up to an upload I made after PAUSE came back online and the recent page now returns an 'Internal Server Error' instead. Search (such as https://metacpan.org/search?q=perl) no longer returns results and even direct links like https://metacpan.org/pod/Task::Kensho are failing.
Edit: Finally getting true HTTP errors instead of MetaCPAN pages with 200 responses and inline messages.
I did not see any obvious problems anymore in the last weeks.
@oalders: Was the root cause found and fixed?
Was the root cause found and fixed?
@eserte, there have been a couple of things going on. PAUSE had some issues that were taken care of. I don't know if they'll pop up again, but it looks like we're getting closer to being able to deploy PAUSE somewhere else.
MetaCPAN had some issues with the Elasticsearch cluster and those were taken care of at the time, but I can't rule out further Elasticsearch issues, based on our current setup. I do hope we'll be able to improve that soon, though. We will see what happens in Lisbon.
At any rate, I think we can close this particular issue now. Thanks, everyone!
Had to fix a bunch of disk errors, which I assume will happen again in the near future. We have a plan to move off the disk, though. Thanks for reporting @eserte. The indexer is catching up now.