feat: start random walk and allow configuration for disabling #42
Conversation
ghost
assigned 
src/index.js
Outdated
| if (err) { | ||
| return callback(err) | ||
| } | ||
| this.randomWalk.start() |
There was a problem hiding this comment.
This should be started just like any other Peer Discovery Service, that means, start it on js-libp2p instead.
There was a problem hiding this comment.
I will change the PR and create a new one for js-libp2p! Thanks
vasco-santos
changed the title
vasco-santos
force-pushed
the
fix/start-random-walk
branch
2 times, most recently
from
3f1abd9 to
45dfb33
Compare
vasco-santos
force-pushed
the
fix/start-random-walk
branch
from
f0bd76c to
f6577f0
Compare
There was a problem hiding this comment.
@vasco-santos the code looks good but it looks like there is an issue with libp2p-mplex that we should fix. Since this is going to be running periodically and by default we need to avoid any uncaught errors.
Looking at the CI error we need to be checking drained here https://github.com/libp2p/js-libp2p-mplex/blob/v0.8.0/src/internals/index.js#L167-L169. If L160 sets drained to false, pushing again will error, which is what CI is not happy about.
|
I will look into that problem @jacobheun . Thanks for finding it out |
|
@jacobheun it is green now! @diasdavid can we have your feedback to move this forward? |
vasco-santos
changed the title
vasco-santos
changed the title
In the context of the integration of the DHT to
js-ipfsjs-ipfs#856.I changed the implementation of
randomWalk. It was using asetInterval, which was causing this function to run the_walkfunction after five minutes. On the contrary,go-libp2p-kad-dhtruns immediately after the node starts and then, each 5 minutes, as we can see in go-libp2p-kad-dht/dht_bootstrap.go#L81-L85Worklist:
random-walkonce the DHT is started