Wiki access blocked

[cpicanco]

Indeed, I am being blocked regardless of the browser. So, what should I do?

[Martin_fr]

Anyone getting persistent connection refused errors from our wiki server, please contact Jonas and provide your IP.


He is the only one (that I know of) who may be able to answer. See also https://forum.lazarus.freepascal.org/index.php/topic,71256.0.html


All I know is from what I read from various sources, including his mails/posts.

Underlaying problem is that there are companies selling DDOS behaviour to scrape any webcontent. That is they will (for their clients, e.g AI companies) load all pages from server, using IP addresses from ranges used for end-user home internet. They disguise as well as possible to be no different to a normal user. And they will use thousands or tens of thousands of IP simultaneously to get as much content in as short a time as possible => i.e. they will use 100% of the targets server load only for themself.

To any open source site, or even medium size company that has the same effect as a full DDOS attack.

Anubis means each IP must do a time expensive puzzle (you will notice your browser will take a few seconds). That means, they must do that for each IP they use, which increases their cost. But apparently (I am guessing) that is not enough. So we still have to block entire ranges of IP while we get exorbitantly high traffic from them. And due to the sheer amount of IP, this only works on big ranges.

[PascalDragon]

Yeah, that sums it up rather well...

Maybe we should provide a pinned post for users that might encounter this problem to quickly find a way to contact e.g. Jonas.

[Martin_fr]

I added a link to my post on the right panel.

[dsiders]

Anyone getting persistent connection refused errors from our wiki server, please contact Jonas and provide your IP.


He is the only one (that I know of) who may be able to answer. See also https://forum.lazarus.freepascal.org/index.php/topic,71256.0.html


All I know is from what I read from various sources, including his mails/posts.

Underlaying problem is that there are companies selling DDOS behaviour to scrape any webcontent. That is they will (for their clients, e.g AI companies) load all pages from server, using IP addresses from ranges used for end-user home internet. They disguise as well as possible to be no different to a normal user. And they will use thousands or tens of thousands of IP simultaneously to get as much content in as short a time as possible => i.e. they will use 100% of the targets server load only for themself.

To any open source site, or even medium size company that has the same effect as a full DDOS attack.

Anubis means each IP must do a time expensive puzzle (you will notice your browser will take a few seconds). That means, they must do that for each IP they use, which increases their cost. But apparently (I am guessing) that is not enough. So we still have to block entire ranges of IP while we get exorbitantly high traffic from them. And due to the sheer amount of IP, this only works on big ranges.

Thank you.

[Seenkao]

Почему тогда для Wiki не сделать Captcha? Которую человек пройдёт легко, а машине потребуется много ресурсов для её прохождения (желательно такую, чтоб машина вообще не могла распознать её).

Суть в том, что новичок, который захочет узнать что-то на вики и не сможет туда попасть, не будет больше пытаться туда попасть. Лишь некоторые люди будут искать решение проблемы.


Google translate:
Why not make a Captcha for Wiki then? Which a human can easily pass, but a machine will need a lot of resources to pass (preferably one that a machine cannot recognize at all).

The point is that a newbie who wants to learn something on a wiki and cannot get there will not try to get there anymore. Only a few people will look for a solution to the problem.

[arnoldpa]

Почему тогда для Wiki не сделать Captcha? Которую человек пройдёт легко, а машине потребуется много ресурсов для её прохождения (желательно такую, чтоб машина вообще не могла распознать её).

Суть в том, что новичок, который захочет узнать что-то на вики и не сможет туда попасть, не будет больше пытаться туда попасть. Лишь некоторые люди будут искать решение проблемы.


Google translate:
Why not make a Captcha for Wiki then? Which a human can easily pass, but a machine will need a lot of resources to pass (preferably one that a machine cannot recognize at all).

The point is that a newbie who wants to learn something on a wiki and cannot get there will not try to get there anymore. Only a few people will look for a solution to the problem.

This is exactly how I feel as someone who is new to Pascal and wanted to learn through the wiki. I almost gave up. But then again, I found some kind of a work around to access the wiki. I use the "Brave" browser's Tor functionality. As of this time, that is the one I am using to access the wiki while waiting for the issue, at least for me, to be resolved.

[Martin_fr]

Depends on the kind of captcha. Many captchas are machine solvable nowadays.

Also, Anubis is exactly that. It is solvable but really expensive to bots like that. Anubis costs them a lot more than solving a captcha. And yet, they keep going to "attack". They make a ton of money by "attacking".

The problem is, that even if they just try to solve it => it still attacks us. If the would fail 9 out of 10 times on a captcha, that would mean the would make 10 times more request to our server. Or killing our server 10 times faster. So with a captcha too, they would still bring the wiki down.

[circular]

I am thinking, maybe allow only access to the login page / create account page as long as the user is not logged in (redirect to login page otherwise). This could be done for suspected IP ranges / when there are too many connections.

[duralast]

The FPC/Lazarus wiki uses MediaWiki, which is what Wikipedia uses. I've never known Wikipedia to be down/blocked. Perhaps because they are much, much larger.

But the FreeBSD wiki has never had this repeated problem to my knowledge; granted they don't use MediaWiki, but the software used doesn't seem to be the problem. FreeBSD isn't large like Wikipedia.

[regs]

Working for me now. I can see and pass some kind of DDoS protection.

[Tony_v_B]

I also can't access the Wiki.
I'm using ExpressVPN so my IP changes. The Anubis filter makes me wait a bit for an animation to finish then something too
fast to read happens and it says invalid response and the Oh noes graphic appears. I usually use Firefox and selectively enable
Javascript using No Script; freepascal.org is marked trusted.
Too bad. Duckduckgo finds some interesting articles I wanted to read. I'll try the Brave-using-Tor suggestion.

Regards,
Tony

[Martin_fr]

I usually use Firefox and selectively enable
Javascript using No Script;

I am using FF with Noscript too, and get past anubis. So it may be some other part... But that would have to be answered by Jonas.

[PascalDragon]

Why not make a Captcha for Wiki then? Which a human can easily pass, but a machine will need a lot of resources to pass (preferably one that a machine cannot recognize at all).

Because we don't want the scrapers to reach the Wiki. Anubis is a reverse proxy in front of the Wiki and accessing it is much less costly on the server than accessing the Wiki itself is.

The FPC/Lazarus wiki uses MediaWiki, which is what Wikipedia uses. I've never known Wikipedia to be down/blocked. Perhaps because they are much, much larger.

While the infrastructure of Wikipedia is bigger, they nevertheless feel the strain as well.

[d2010]

I also can't access the Wiki.
I'm using ExpressVPN so my IP changes. The Anubis filter makes me wait a bit for an animation to finish then something too
Regards,
Tony

Yes , you have right, i try my self from smartphone, Samsung, and I can not access to the wiki.com.  Perhaps "lazarus forum. com" work  slow on smartphone?