Increase default difficulty Part 1
With this release we increase the default difficulty from 30 to 36. This change is needed to protect the network against different attacks. This change only effects the identity generation and signing. You can still use an old identity with a lower difficulty as long as it is already signed.
In part 1 (this release) we change the default for the identity binary but not for the certificate authority. You will still be able to sign identities with difficulty 30 or higher. If you haven’t signed your identity you should do it now. Last call
In part 2 (next release) we change the certificate authority as well. At that point you can only sign an identity with difficulty 36 or higher.
Remove default concurrency limit
By default a storage node had a limit of 6 concurrent transfers and will reject additional uploads until one of the ongoing transfers is finished. We believe with dRPC the limit is not needed anymore. By default the storage nodes will now take as many uploads as they can. You can still set the limit in your config file like before. Removing it from the config will reset it to the new default value. Changelog v0.24.5
Last release only halve of that change was released. This time the change is real. In addition to that we also increase the audit timeout from 25s to 5 minutes to avoid disqualifications for responding too slow. If you haven’t already please remove any limit you might have set in your config file. We want to test the limits of dRPC.
Graceful exit
Doing our internal tests we found some edge cases that we want to cover first. Graceful exit is still under construction. Please don’t call it.
Storj-sim requires redis
dRPC by default
The gateway and uplink binaries are compiled with dRPC by default but storagenode, satellite and all the other binaries are still compiled with gRPC. If you want to compile one of them from source you have to append -tags=grpc. If you are using one of our binaries, auto updater or docker containers you don’t need to do anything.
I think limit is still persisting (I removed from config storage2.max-concurrent-requests) and after storagenode upgraded to latest version see this on log:
docker logs storagenode 2>&1 | grep rejected
2019-11-07T20:01:27.060Z ERROR piecestore upload rejected, too many requests {“live requests”: 9}
2019-11-07T20:01:28.599Z ERROR piecestore upload rejected, too many requests {“live requests”: 9}
2019-11-07T20:01:30.996Z ERROR piecestore upload rejected, too many requests {“live requests”: 8}
2019-11-07T20:02:54.721Z ERROR piecestore upload rejected, too many requests {“live requests”: 8}
2019-11-07T20:12:28.007Z ERROR piecestore upload rejected, too many requests {“live requests”: 8}
2019-11-07T20:12:28.461Z ERROR piecestore upload rejected, too many requests {“live requests”: 8}
That iz is not what it means. Watchtower will not immediately update all nodes on the network at once. This is intentional in order to avoid too many nodes going offline at once. Please just wait until it is your turn.
[quote=“littleskunk, post:1, topic:2517”]
concurrency
sorry ment edge, still not working though. do i need a difficulty of 36… I believe my token is signed, that means that you ran the authorize storagenode command right???
