Advertisement
Guest User

Untitled

a guest
Oct 25th, 2014
123
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 7.50 KB | None | 0 0
  1. 10-21
  2.  
  3.  
  4. 07:03 gnu-sense has joined (~gnu-sense@c-98-223-179-205.hsd1.in.comcast.net)
  5. 07:03 delusional: this is getting silly
  6. 07:04 delusional: tell me again why we can't just ban his ass, i mean, the attack has to be coming from somewhere
  7. 08:13 blast007: it's coming from the Internet
  8. 08:13 blast007: better ban it all
  9. 08:14 blast007: unplug the server from the network! solved!
  10. 08:14 delusional: what, he's got a botnet?
  11. 08:16 blast007: proxiese
  12. 08:16 blast007: proxies*
  13. 08:16 delusional: cant we ban the proxies?
  14. 08:16 blast007: there's tens of thousands of them, so, no
  15. 08:17 delusional: well, if we did ban the ones he's using right now, we'd at least get a temporary repreive?
  16. 08:17 blast007: last time I tried adding a huge list of proxies to my firewall I made my server very unhappy
  17. 08:17 blast007: yeah, for a few minutes until he refreshes the proxy page and gets 10 more
  18. 08:18 blast007: sounds like you think it's not easy to get a list of several hundred proxies...
  19. 08:18 delusional: why cant we detect them quicker?
  20. 08:18 blast007: and then there's also Tor (which is actually easy to block) and VPNs
  21. 08:19 delusional: why cant we detect, and ban them quicker?
  22. 08:19 blast007: or I should say that Tor is easy to block completely, but a bit more challenging if you only want to block Tor exit nodes that have an exit policy that allows them to talk to your ports that are under attack
  23. 08:20 blast007: how would you detect a proxy?
  24. 08:20 delusional: by detecting the attack
  25. 08:20 delusional: i know,,, thats what we're trying to do
  26. 08:21 blast007: so really then the detecting the proxy is irrelevant - you just want to detect the attack and react to that
  27. 08:21 blast007: I'd imagine it's possible to write a Snort ruleset for the attack
  28. 08:21 delusional: so... when he joins, we never get the full IP?
  29. 08:21 blast007: but I've never used Snort before
  30. 08:21 blast007: what?
  31. 08:22 delusional: we see the multiple attempts to join.. or whatever it is he's doing... don't we get the IP address from that?
  32. 08:27 blast007: of course
  33. 08:27 blast007 has changed mode: -o blast007
  34. 08:28 delusional: why are we having such a problem detecting that?
  35. 08:28 blast007: because nobody has tried to detect it?
  36. 08:30 delusional: I don't know, but it seems to me it should be relatively simple to get the address, and move it over to the firewall banlist. Why are we having such a probem getting the address?
  37. 08:33 delusional: I'm getting really fucking pissed off.
  38. 08:35 delusional: I guess if he knows his way around with proxies, he can continue to run rings around us forever.
  39. 09:57 blast007: again, I don't think anyone is trying to do that yet
  40. 15:32 blast007: also, if you do get flooded with connections and you firewall them out, the connections will still remain there for a LONG time, though that may depend on how you firewall them.
  41. 15:54 blast007: I've used a tool called killcx before to forcefully kill connections
  42. 15:56 blast007: I had ended up using a command like this:
  43. 15:56 blast007: netstat -nut | grep '186.150.130.118' | awk '{ print $5 }' | xargs -l1 -t -I {} ./killcx {} 2>&1 | less
  44. 15:56 blast007: where 186.150.130.118 was the attacker IP address
  45. 15:58 blast007: might actually want to replace the order of that awk and the grep though just to be safe..
  46. 15:59 blast007: netstat -nut | awk '{ print $5 }' | grep '186.150.130.118' | xargs -l1 -t -I {} ./killcx {} 2>&1 | less
  47. 15:59 blast007: of course, you'd want to do that *after* you firewall, and only if the connections hang around
  48. 16:00 blast007: I used that when DR had flooded my server with connections, and I wanted to see if I could fix it without simply restarting bzfs
  49. 16:00 blast007: http://killcx.sourceforge.net/
  50. 16:01 blast007: needs a few perl modules as well, and needs to be run as root (I think)
  51. 18:28 delusional: it's makin me sick
  52. 19:56 You have joined the channel
  53. 19:56 delusional has joined (~delusiona@pool-173-71-173-74.pitbpa.fios.verizon.net)
  54. 19:56 Topic: BZFlag Anti-Cheater Task Force
  55. 19:56 blast007 set the topic at: Oct 15, 2014, 7:38 PM
  56. 19:56 Mode: +Cnpst
  57. 19:56 Created at: Oct 15, 2014, 1:29 PM
  58.  
  59.  
  60.  
  61.  
  62. 10-20
  63.  
  64. 03:46 Flash: On my machine, I can join/quit/join ... I wonder what is different
  65. 03:46 Flash: unfortunately, I am completely out of time for tonight
  66. 08:13 gnu-sense has joined (~gnu-sense@c-98-223-179-205.hsd1.in.comcast.net)
  67. 08:49 delusional: I'm having no problems right now
  68. 08:55 delusional: i could join rejoin at will, until a second player got there
  69. 08:56 delusional: and you can just wail away on the enter key, trying to connect, easy way to recreate the hack
  70. 10:32 You have joined the channel
  71. 10:32 delusion_ has joined (~delusiona@pool-173-71-173-74.pitbpa.fios.verizon.net)
  72. 10:32 Topic: BZFlag Anti-Cheater Task Force
  73. 10:32 blast007 set the topic at: Oct 15, 2014, 7:38 PM
  74. 10:32 Mode: +Cnpst
  75. 10:32 Created at: Oct 15, 2014, 1:29 PM
  76. 10:50 You have joined the channel
  77. 10:50 delusional has joined (~delusiona@pool-173-71-173-74.pitbpa.fios.verizon.net)
  78. 10:50 Topic: BZFlag Anti-Cheater Task Force
  79. 10:50 blast007 set the topic at: Oct 15, 2014, 7:38 PM
  80. 10:50 Mode: +Cnpst
  81. 10:50 Created at: Oct 15, 2014, 1:29 PM
  82. 11:11 You have joined the channel
  83. 11:11 delusion_ has joined (~delusiona@pool-173-71-173-74.pitbpa.fios.verizon.net)
  84. 11:11 Topic: BZFlag Anti-Cheater Task Force
  85. 11:11 blast007 set the topic at: Oct 15, 2014, 7:38 PM
  86. 11:11 delusional has left IRC (Read error: Connection reset by peer)
  87. 11:11 Mode: +Cnpst
  88. 11:11 Created at: Oct 15, 2014, 1:29 PM
  89. 14:25 delusion_: Help Me Obi-Flash-Kenobi, you're my only hope.
  90. 17:08 Flash: ok, that should fix it
  91. 17:08 Flash: I even think I understand why :)
  92. 17:08 Flash: but all the cheater limiting is gone
  93. 17:09 Flash: allejo should be able to add it back in, if he is so inclined
  94. 17:29 allejo: wait, so are we back at square 1? lol
  95. 17:45 Flash: sort of
  96. 17:45 Flash: I took out most all of what I did, mostly because it was never supposed to be committed like that
  97. 17:46 Flash: However, the first thing I did (eliminate random memory read) broke it (init time of last action to 0)
  98. 17:46 Flash: This last change correctly fixes the problem noted in the first change
  99. 17:47 Flash: https://github.com/BZFlag-Dev/bzflag-import-3/commit/2ac2df538edc28969fcf69703cf91238fbacf33b
  100. 17:47 Flash: put the bzfs change back in if you like
  101. 17:48 Flash: it should be better now
  102. 18:01 allejo: so reimplement that patch with the current trunk?
  103. 18:01 Flash: yes, just the bzfs file
  104. 18:14 allejo: makes changes to bzfs.cxx
  105. 18:15 Flash: crosses fingers
  106. 18:16 allejo: compiles and i can rejoin freely :D
  107. 18:16 allejo: i think
  108. 18:16 allejo: delusion was saying he could rejoin freely until more people joined
  109. 18:18 Flash: it was actually until 10 mins after the first join
  110. 18:19 Flash: I see libcommon.a as part of the build.
  111. 18:19 allejo: oh
  112. 18:19 Flash: wrong window
  113. 18:21 apeman has joined (~vtw15@132.181.52.166)
  114. 18:30 Flash: no more than 3 joined so far, but seems better
  115. 18:36 Flash: had up to 5; saw some jitter issues
  116. 18:36 Flash: no lag to speak of
  117. 19:29 delusion_: i don't have a working client
  118. 19:43 Flash: still tough. Are you seeing any denied?
  119. 20:03 apeman: yes, clients are definitely being denied
  120. 21:52 gnu-sense has left IRC (Quit: gnu-sense)
  121. 23:30 Flash: it occurs to me that rather than close the bad sockets, we should hold them open.
  122. 23:43 delusion_: he cant come back?
  123. 23:48 Flash: well, if we close the socket, it's an immediate indication to his script to open a new one. but if we keep it open, he can keep sending data until his TCP buffer fills up
  124. 23:48 Flash: we just wouldn't read it
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement