Open Rashkae2 opened 19 hours ago
After a bit more testing, it appears to me as though accelerated DHT was modified so it works without the long, complete scan of the network it used to do hourly. And that would be really great. I can Start up and republish a large amounts of data very quickly and with no harmful impact to my network/internet connection.
If I start a second node on the internet with Accelerated DHT turned off, I can retrieve the content I published from the accelerated node quickly and reliably. However, if I also enable accelerated DHT on the 2nd node, it will not be able to find most CID's I try to retrieve.
Checklist
Installation method
dist.ipfs.tech or ipfs-update
Version
Config
Description
When I start ipfs daemon with Accelerated DHT client, the initial network scan before the routing table populates lasts less than 1 minute. (You can check with ipfs stats dht)
However, rather than the previous over 10,000 peers, the dht routing table now shows less than 1000 peers. Also, ipfs routing findprovs does not find many CIDs, but those same CID's can be found very quickly if accelerated DHT is turned off.