Network Flooded with Queries

YaCy YaCy Peer-to-Peer
21

:latin_cross: It is good to see your public confession of faith. :sparkles:

Your have uncovered some unnatural search queries…

why does my dog dog in the backyard

As for me, I do not know why the dog would do such a thing. :thinking:

Information sharing for our investigation

I am still seeing some of your machines sending out solr queries, and I do not know why… do you?

Example, a useless query for most:

www

solrselect

The IP address that sent that query came from one of your machines.

Impossible
Impossible1021×646 49.2 KB

And this is despite the fact that my machine is in Robinson mode, which is supposed to reject all peer-to-peer activity. So, does that mean this solr query came from your machine as a request independent of Yacy?

22

Thank you Roamn, I had looked at that, but I would like to shoe the pattern over time, with queries and such if I can. I did figure out a way, and I am testing it now. Thanks for the idea, that was a good place to start. :slight_smile:

23

Hmmm… Well, I would think that if someone searched ‘www’ on that node, the search is broadcast out to the peers it is connected to, so that would probably be why it got to you.

As for Robinson mode, I might be misunderstanding that, as I have never used it, but I thought that just applied to DHT activity, and searches would still be handled as normal. I’ll have to look into that. Any search that came to you from any of my nodes that is not marked ‘-dev’ is coming from a stock yacy node. I am not sure what you mean by a request independent of Yacy?

Hard to keep the very best thing that ever happened to me… a secret. :slight_smile: Plus there is Matt 10:32-33.

As for the unusual searches, that is the tip of teh iceberg. Much of what I am seeing is just random noise being searched directly on my nodes.

I did stop the server crashing, but I see you have a separate thread about that, so I will chime in there.

24

Right, and not for you and me only, but for all mankind. John 7:37-39

Yes, there are the two search methods: Using the Yacy search interface (like in the web browser) is one.

yacysearch-form
yacysearch-form1494×934 351 KB

The other is a solr query, which appears to currently have no admin settings that limit its use.

I prefer the solr query, because it is a more powerful way to search. Ensuring all the keywords desired appear in the results is a huge advantage. So, I would not want to take this feature away, but I would like administration controls to:

  • limit the tool’s call rate
  • turn it on/off
  • option for required authentication

More admin control could protect servers from crashing, especially in cases like we have seen the past week with a burden of too-frequent queries.

25

You are very detailed. I love the video, but a link would have probably worked. :slight_smile:

I see what you are talking about, you are talking about bypassing the web interface, and I was thinking you were saying bypassing yacy entirely, as if there was a way to query solr remotely without touching the yacy install/interface layer for solr. I’m up to speed now, thanks.

… indeed.