Tor retries download for unparseable descriptors
15:14 <@diseasearma> hm. moria1 is starting to complain 15:14 <@diseasearma> Mar 18 15:00:15.261 [warn] Invalid uptime "-29903" 15:14 <@diseasearma> so is moria2, and my bridge relay. 15:15 <@diseasearma> that's not a very useful warning message. 15:15 <@diseasearma> Mar 18 15:15:35.895 [info] connection_dir_client_reached_eof(): Received server 15:15 <@diseasearma> info (size 2471) from server '194.109.206.212:80' 15:16 <@diseasearma> is what prefaces it. 15:16 < nickm> seeing that on peacetime too 15:16 <@diseasearma> looks like my relays are going to dizum to get a descriptor they don't have, 15:16 <@diseasearma> and it's a descriptor they don't want, 15:16 <@diseasearma> but then they go back again and get it again. 15:17 <@diseasearma> they go back once a minute 15:17 < nickm> looks like they're counting it as a download failure. 15:17 < nickm> there should be some way to mark it as perma-failed. 15:18 * nickm restarts peacetime on the 0.2.0 branch. 15:19 <@diseasearma> ok. we're going to see this come up on or-talk shortly. 15:20 <@diseasearma> i guess i could preemptively post saying it's a known problem and will be solved 'later'. 15:20 < nickm> s/later/soon/ 15:20 < nickm> we can even add a flyspray entry. 15:20 < nickm> s/we/I/
In summary, when we download a descriptor and then can't parse it, we need to remember to never try to download it again. Also, our log message could be better.
[Automatically added by flyspray2trac: Operating System: All]