Project

General

Profile

Actions

Task #4117

closed

Remove check-for-self in ChronoSync updates

Added by Nicholas Gordon over 7 years ago. Updated over 7 years ago.

Status:
Closed
Priority:
Low
Target version:
Start date:
06/02/2017
Due date:
% Done:

0%

Estimated time:

Description

Currently the sync-logic-handler checks whether the received update is from itself. In theory this shouldn't be possible, because:

  1. Broadcast strategy does not send the interest back on the Face it received it on. This does not protect from loops of 3 or more, except
  2. Nonces in the Interest should cause NFD to drop the interest before it gets to this point.

It's a small change, but this code appears to be redundant and removing it would slightly improve performance and readability.

Actions #1

Updated by Nicholas Gordon over 7 years ago

  • Status changed from New to Closed
  • Assignee set to Nicholas Gordon
Actions #2

Updated by Ashlesh Gawande over 7 years ago

I am seeing NLSR getting its own updates on the testbed so check might be necessary

Logs from titan:

1497532235.104 DEBUG: [Lsdb] Received data for LSA(name): /localhop/ndn/NLSR/LSA/pt/uminho/%C1.Router/vnetlab/name/%C47
1497532235.107 DEBUG: [Lsdb] Received data for LSA(name): /localhop/ndn/NLSR/LSA/pt/uminho/%C1.Router/vnetlab/coordinate/E%D8
1497532235.110 DEBUG: [Lsdb] Received data for LSA(name): /localhop/ndn/NLSR/LSA/pt/ulusofona/copelabs/%C1.Router/ndn/name/%00%01%0FS
1497532235.114 DEBUG: [Lsdb] Received data for LSA(name): /localhop/ndn/NLSR/LSA/pt/ulusofona/copelabs/%C1.Router/ndn/coordinate/%80%CB
1497532235.136 DEBUG: [SyncLogicHandler] Received Nsync update event
1497532235.136 DEBUG: [SyncLogicHandler] Update Name: /localhop/ndn/NLSR/LSA/edu/memphis/%C1.Router/titan Seq no: 75080220140371968
1497532235.161 FATAL: [NlsrRunner] ERROR: Data size exceeds maximum limit
1497532235.161 DEBUG: [Fib] Fib::clean called
Actions #3

Updated by Nicholas Gordon over 7 years ago

Isn't that the greatest? Code without an obvious purpose, that should be unnecessary, yet is required by production evidence? Further, there's no mention even of the anomalous necessity of it.

I'll undo the change, then.

Actions #4

Updated by Ashlesh Gawande over 7 years ago

Nicholas Gordon wrote:

Isn't that the greatest? Code without an obvious purpose, that should be unnecessary, yet is required by production evidence? Further, there's no mention even of the anomalous necessity of it.

I'll undo the change, then.

I think we should try to think and reproduce how it might have happened - it might not happen with different sequence numbers.

Actions

Also available in: Atom PDF