502 Bad Gateway on the Public Bug Tracker

[Air Force One]

Try clearing the cookies for www.virtualbox.org (some months ago there was a change to the SameSite cookie attribute which should've made things more compatible with the browsers out there, maybe you still have the old state in your cookie collection).

I don't think that this is trouble on the client side, like cookies and so on. As you wrote last year, the cause is in the backend, like a load balancer, a thread pool or some issues with the database. Could it be that yesterday we also had something like this for a short time, and afterward it was gone? Maybe somebody could see something in the logs?

[klaus]

Could check the logs... not today, but hopefully in the next days. The load balancer control (which is needed during authentication) is actually relying on cookies, one which is set at the beginning of auth and cleared again when the response has reached the same backend server. It is done to mitigate between the needs of the auth solution (which has to be handled by a single backend server, no matter which one) and spreading the normal application load (which means that every connection goes to a different backend server).

[Air Force One]

The load balancer control (which is needed during authentication) is actually relying on cookies, one which is set at the beginning of auth and cleared again when the response has reached the same backend server.

My Firefox is set up to not have any history and so on, it is in permanent private mode. So there should be no cookies and no troubles with cookies. Am I wrong here?
What I also like to tell - the changes in configuration are working. Previously I got this situation more often; now it came first after like 6 months. I just wanted to inform you.

[Air Force One]

Could check the logs... not today, but hopefully in the next days. The load balancer control (which is needed during authentication) is actually relying on cookies, one which is set at the beginning of auth and cleared again when the response has reached the same backend server. It is done to mitigate between the needs of the auth solution (which has to be handled by a single backend server, no matter which one) and spreading the normal application load (which means that every connection goes to a different backend server).

After the site redesign, I start once again having all kind of 5xx errors on the forum and bug tracker access, like 502 and 504. Can there be something done with it?

[klaus]

Half the world seems to be crawling all pages on virtualbox.org in the dumbest possible way lately (without admitting it's by a bot/crawler, leave alone bothering to look at robots.txt), and the load caused by that is pushing the servers to their limit.

Suspect that countless AI projects are collecting everything they can get hold of.

See the new link at the bottom of www.virtualbox.org - I know that it will not help in any way.

[Air Force One]

Half the world seems to be crawling all pages on virtualbox.org in the dumbest possible way lately (without admitting it's by a bot/crawler, leave alone bothering to look at robots.txt), and the load caused by that is pushing the servers to their limit.

Thanks g%d this one wasn't caused by some refugees from Haiti?

So this time we can't do anything about it, because this time it's a real load, and not some load balancer troubles?

Another trouble: the trac instance is broken after the site redesign. I don't see the attachments section in the ticket. So one can't add some attachments. And if I could do it by researching the URL of the attachment-adding page, only one attachment can be added, because after the addition, I land in Nirvana (not the band ), not the page asking for another attachment.