Advertisement
Guest User

Untitled

a guest
Nov 28th, 2014
170
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 5.13 KB | None | 0 0
  1. [9:06:10] Mac: it is your network configuration
  2. [9:06:19] Mac: there's no other immediate use for the local 10.0.2.2
  3. [9:06:28] Mac: we're almost certain something must have changed
  4. [9:06:40] Mac: theres no logical outcome when it doe sit suddenly.
  5. [9:06:45] Mac: short of removing amped and removed l2d
  6. [9:06:52] Mac: and loggin in with a good ammount of testers.
  7. [9:07:45] Siqma L2Gold: well
  8. [9:07:54] Siqma L2Gold: there were no problems almost all day yesterday
  9. [9:08:11] Siqma L2Gold: until after 12h of being up, the ort report happened again
  10. [9:08:21] Mac: ort report happens because of referneces
  11. [9:08:23] Mac: normally it's harmless
  12. [9:08:39] Mac: but it suggests tha tsome connection with players\server\npc gets corked
  13. [9:08:45] Mac: and all info starts going nuts since there's no communication.
  14. [9:09:41] Siqma L2Gold: problem is that we cant get like more than 25 testers
  15. [9:09:46] Siqma L2Gold: which is not even close to accurate
  16. [9:09:55] Mac: then boot up without amped
  17. [9:09:56] Mac: on live
  18. [9:09:59] Mac: see what happens
  19. [9:10:07] Mac: boot up without l2d see what happens
  20. [9:10:08] Mac: etc
  21. [9:10:20] Mac: either way, if it's a player initiated issue
  22. [9:10:25] Mac: amped or some other thing woudl log it
  23. [9:10:28] Mac: and what cha\account\ip etc
  24. [9:10:48] Mac: but given that amped + lserver logs start pointing to local\10.0.2.2 it looks like a virtualbox issue
  25. [9:11:06] Mac: when shit goes nuts that is.
  26. [9:11:42] Mac: from what it looks like in the logs, once this issue starts then random players start appearing to be sending bad packet data
  27. which sounds to me a lot like fucked up network stack from vm
  28. it thinks its reading from socket x but its really reading from socket y
  29. [9:12:10] Mac: thereby causing all sorts of crap due to differentiated encryption per socket.
  30. [9:12:15] Mac: ergo it's decrypting data wrongly.
  31. [9:13:52] Siqma L2Gold: alright so if the problem keeps appearing like every day, we will try what you said
  32. [9:14:04] Mac: we can also start ruling out code sections
  33. [9:14:04] Siqma L2Gold: cause until yesterday it was happening every 4-5 days
  34. [9:14:06] Mac: native, amped, us.
  35. [9:14:13] Mac: only start*
  36. [9:14:41] Mac: if it happens in more than 1 combination of missing amped or us
  37. [9:14:44] Mac: its not code
  38. [9:14:49] Mac: i refuse to believe it's l2server itself.
  39. [9:15:00] Mac: such an issue must have happened before in general.
  40. [9:15:20] Mac: so if its not added code, then it points back to the network stack
  41. [9:15:25] Mac: which when im VM, would be very pointy towards VM.
  42. [9:15:31] Mac: im=in
  43. [9:15:56] Mac: If veljko is still using VB id suggest him to move.
  44. [9:16:09] Mac: its not really awesome for high bandwidth virtualizing.
  45. [9:16:22] Mac: or atleast someone that's super into VB needs to config the VM
  46. [9:16:33] Mac: to not have VB internally throttle or mess up stuff.
  47. [9:16:48] Siqma L2Gold: alright i will tell him
  48. [9:16:58] Mac: we use vb for test servers
  49. [9:16:59] Mac: and its fine
  50. [9:17:09] Mac: but ive heard people whine about VB and live servers alot.
  51. [9:17:41] Siqma L2Gold: well there are some ppl complaining about lagging
  52. [9:17:49] Siqma L2Gold: dunno if it has to do anything with vb
  53. [9:19:40] Siqma L2Gold: Also, people are saying that when they login, some weird windows start to pop up
  54. [9:19:45] Siqma L2Gold: we applied new patch recently
  55. [9:19:50] Siqma L2Gold: and that wasnt happening before
  56. [9:19:51] Siqma L2Gold: http://s24.postimg.org/v9tz1m1rp/Shot00143.png
  57. [9:20:02] Mac: new patch?
  58. [9:20:05] Siqma L2Gold: although the patch contained only some refining part and implementing new armor
  59. [9:20:14] Mac: if it really only contained dat
  60. [9:20:17] Mac: should not matter
  61. [9:20:18] Mac: those windows
  62. [9:20:20] Mac: are because of packet shit
  63. [9:20:26] Mac: its like client sends Open Inventory
  64. [9:20:40] Mac: but server reads OPEN WAREHOUSE§!
  65. [9:20:41] Mac: and shit goes nuts
  66. [9:20:44] Mac: and no one can login
  67. [9:21:00] Mac: because when it starts happening, server reads the starting login handshakes that determines encryption
  68. [9:21:11] Mac: and sees disallowed keys\data comparing to normal protocol.
  69. [9:21:27] Siqma L2Gold: so what should we do?
  70. [9:21:30] Mac: communication between client\server goes nuts.
  71. [9:21:43] Mac: we're like 99% pointing finger towards VM.
  72. [9:22:12] Siqma L2Gold: alright
  73. [9:22:22] Mac: because its at that level it happens
  74. [9:23:06] Mac: server->data send > data encrypt > raw socket send\windows <> raw socket receive > client receive > client decrypt > bad shit.
  75. [9:23:08] Mac: same reversed.
  76. [9:23:17] Mac: so somwhere in the socket reading\sending shit goes nuts.
  77. [9:23:43] Mac: when leaving server or leaving client at that application level it's fine
  78. [9:23:50] Mac: its when it goes further down shit goes nuts.
  79. [9:24:25] Siqma L2Gold: ok i will copy him what you said
  80. [9:25:03] Mac: you can see it happening if you do l2phx
  81. [9:25:08] Mac: and wait for it happen.
  82. [9:25:21] Mac: because clientside it sending correct packets based on actions fine
  83. [9:25:31] Mac: but server side you see weird ass logs about packets being wrong
  84. [9:25:33] Mac: for warehouse or etc.
  85. [9:25:46] Mac: and you can see, that server packets start going nuts.
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement