When I launch my IPFS daemon, content I have published is remotely available for a few hours. After a while, nevertheless, the network forgets about it. If I do “ipfs dht provide” to refresh the network, the IPFS network is able to locate my content again.
I have not changed the default configuration for the reprovider:
The reason you might be facing this problem is that provider runs take longer than 24 hours which is also the expiration period for provider records.
As mentioned above, enabling the accelerated DHT Client should help at the cost of higher resource consumption.
Another option is to change the provider strategy to pinned or roots which would reduce the number of records published (they would still be available via Bitswap) .
You can find more information about dealing with this problem in the following blog post:
After enabling the accelerated DHT client, the reported publishing time is less than 9 minutes. I don’t know the old publishing time, but I very much doubt it was > 24 hours.
Changing provider strategy to “pinned” will not help because all my storage is pinned (not cached foreign content) and “roots” is flaky if whoever is requesting my data is not a peer of me. Burned already trying that.
The provided blog link was useful and interesting. Thanks.
I currently enabled the accelerated DHT client. Lets see. If not enough, I will decrease the publishing period. Advertising blocks is the current fundamental Achilles heel of IPFS.