Melvil Stakhanov

[keyplyr]

UA: Melvil Stakhanov/1.0
Protocol: HTTP/1.1
Robots.txt: Yes
Host: ovh.com
91.121.0.0 - 91.121.255.255
91.121.0.0/16

Is this the new log spam... naming a bot after yourself?

[not2easy]

Maybe related to that other OVH Melvil guy - UA: Melvil Rawi/1.0 ? [webmasterworld.com...]

[keyplyr]

Maybe, saw that guy today also. Can't tell behavior since they're both blocked. Almost identical headers though.

Could just be that a lot of Russians that use French server farms to run bad bots are named Melvil.

[lucy24]

Maybe it stands for something? I-forget-which university's library catalog used to be called Melvyl (with a Y). I kinda doubt it was named after anyone.

[keyplyr]

Or Rawi died and Melvil got a new partner named Stakhanov :)

[lucy24]

:: bump ::

I find this suggestive:

91.121.64.34 - - [30/Aug/2016:08:15:09 -0700] "GET /robots.txt HTTP/1.1" 200 689 "-" "Melvil Stakhanov/1.0" 
94.23.152.68 - - [30/Aug/2016:21:06:06 -0700] "GET /ebooks/bourquin/ HTTP/1.1" 403 1801 "-" "Melvil Rawi/1.0" 
46.105.161.110 - - [31/Aug/2016:09:06:07 -0700] "GET /ebooks/bourquin/ HTTP/1.1" 403 1801 "-" "Melvil Rawi/1.0" 
5.135.213.144 - - [31/Aug/2016:21:06:12 -0700] "GET /ebooks/bourquin/ HTTP/1.1" 403 1801 "-" "Melvil Rawi/1.0" 
51.255.122.73 - - [01/Sep/2016:09:06:22 -0700] "GET /ebooks/bourquin/ HTTP/1.1" 403 1801 "-" "Melvil Rawi/1.0" 

When the only occurrence of Melvil S. (in all logs, ever, searching for "Melvil") is on the same day as a visit from Melvil R., you have to wonder.

The page requests included

Accept-Charset: utf-8

while robots.txt didn't. Otherwise identical. (I do have some "badcharset" patterns, but this isn't one of them.)

Note the timestamps, incidentally. That's weird.

[keyplyr]

I find this suggestive

I get that a lot.

Accept-Charset: utf-8

Is that always in a header request for .txt files? I've never paid much attention. I know it's there for html requests.

Anyway, it looks like one cousin goes for robots.txt and the other goes for pages.

[lucy24]

Is that always in a header request for .txt files?

No, in fact one of the problems with logging headers on robots.txt requests is that they don't necessarily send the same headers as for page requests. It's just a rough starting point. If anything, the robots.txt headers are "worse". In this specific case it's even trickier, because it's legitimately possible for humans to send no charset header at all while utf-8 alone isn't currently on my badcharset list.

Hm. A txt file, by its nature, can't contain any information about its character set--a BOM in a plain-text file can lead to problems--so there wouldn't be much point to the header.

Fortunately, Melvil has not yet done anything that would move me to poke a hole for him, so the exact grounds for denial don't matter.