View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0005645||GNUnet||DNS service||public||2019-03-13 10:46||2019-05-02 14:28|
|Product Version||SVN HEAD|
|Target Version||Fixed in Version|
|Summary||0005645: Issuing request to a local DNS service breaks node|
|Description||I found there is a local resolver service running that gets DNS traffic from iptables and decided to try it.|
The dig query timed out, but I found gnunet-arm went unresponsive and there's a constant stream of ERROR/MQ messages in transport and ats logs:
Mar 13 08:21:00-682310 transport-25692 ERROR MQ with 10000 entries extended by message of type 364 (FC broken?)
Mar 13 09:35:54-724788 ats-28266 ERROR MQ with 10381 entries extended by message of type 345 (FC broken?)
The spew persists after restarts and seems to die out after a while.
|Steps To Reproduce||1) gnunet-arm -i dns|
2) dig www.W4GVVA8FXH1QPTHS533MZTGASTSKK3MWGGTBH9T0WMFDR482CE4G @169.254.1.1
4) Watch transport and ats go haywire.
|Tags||No tags attached.|
I tried to reproduce it a with and without UDP transport enabled and it is much more stable. The dig still silent, but at least the logs aren't getting hammered.
Additionally, issuing `gnunet-transport -in` dumps a limited amount of MQ errors [in ats logs] with UDP enabled, but silent otherwise.
||I'm not sure this is related to the local DNS request, usually those errors appear after a while just in general AFAIK. And getting this mess under control is why I'm rewriting TRANSPORT/ATS in the "TNG" redesign these days...|