Bee v1.1.0 release notes

Hello bees!
These last couple of months have been paved with a lot of learnings for the bee team. We've been meticulously implementing your recommendations, trying to improve the bee developer experience so that you could build faster on Swarm, as well as inspecting the production network behavior, ironing out all sorts of kinks, bugs and performance issues.

IMPORTANT NOTE: the xDai chain is going to implement the London hardfork. Please stay informed and update your infrastructure accordingly as OpenEthereum clients are still not guaranteed to be supported after the hardfork!

Breaking changes

• We renamed log field traceid to traceID (#2324)
• Standalone mode has been phased out (in favor of dev mode, see below) (#2331)

Deprecations

• PATCH under the /bzz endpoint has been moved to the stewardship endpoint. Please use that instead (#2356)
• The stamps API has moved to the debug API (#2171)

New features

• Add estimation of expected lifetime of batches (#2336)
• Add possibility to cancel transactions (#2212)
• Allow node to start in dev mode which does not require a blockchain backend. All storage is done in memory and backend behavior is mocked, so that you could develop your apps on Swarm even faster! Run ./bee dev and give it a try (#2347)
• Add chunk upload endpoint that allows to stream chunks via websockets (#2230)

Hardening

• Improve pushsync skiplist logic (#2348)
• Improve pushsync replication in neighborhood logic, as it would sometimes replicate to peers which are not in the neighborhood (#2237)
• Replicate with peers further from the chunk (#2293)
• Discovery announce to lightnodes was not in place, which could under certain conditions impair their connectivity (#2351)
• Disable bootnode active dialing (#2169)
• Update dependency versions (libp2p, prometheus, among others) (#2316)
• Peers are now checked for reachability before being persisted in the addressbook (#2319)
• Bump base docker image (#2315)
• Continue processing pullsync batch on unverifiable stamp. Previously, the whole batch would error and other chunks potentially discarded (#2228)
• Prevent sending transaction if we don't have sufficient balance (#2337)
• Cache on forwarding in retrievals (#2234)
• Rate limit hive broadcast calls (#2235)
• Disable forwarding node stamp verification, this is to mitigate nodes which are not properly synced to the chain from interrupting the syncing process (#2233)
• Verify signer result chain ID conform to node configuration (#2214)
• Postage stamp usable flag now indicates also if the batch was already expired (#2390)
• When uploading content which is stamped using an already expired postage stamp, the node will not attempt to sync the content. You are advised to use longer-lived postage stamps and encrypt your content to work around this. It is not possible to reupload unencrypted content which was stamped using an expired postage stamp. We're working on improving on this (#2392)

Bugfixes

• Fixed wrong timestamp usage in time settlements (#2358)
• Fixed correct blocktime for xdai (#2340)
• Fixed possible gcSizeChange inconsistency (#2296)
• Fixed deadlocks on context done in pushsync and retrieval protocols. This caused goroutine leaks and elevated memory consumption (#2327)
• Fixed a mantaray lookup bug where files that have shared prefix aren't found correctly (#2380)
• Limit concurrent access to some parts of the api resources to improve API consistency, since bee does not support parallel on-chain operations (#2379)
• Query header instead of block (resolves london hardfork related issues) (#2372)
• Fixed a bug in kademlia that caused choosing a sanctioned peer closest to a chunk. Oops! (#2374)
• Fixed a bug in the cancelled transactions nonce check (#2395)
• Fixed excruciatingly long shutdown by allowing metrics to flush dirty (#2428)
• Fixed unnecessary flushing of metrics that haven't changed in kademlia (#2418)

Performance improvements

• Improve PSS efficiency (#2140, #2287)

Misc

• Updated go-ethereum version (#2342, #2408)
• Bump clef in dockercompose (#2366)
• Added batchstore checksums (#2227)
• Expose the transaction API early (#2210)
• Add endpoint that emits information about collisions within all buckets of a batch (#2321)
• Kademlia metrics flush in a separate goroutine in a way that doesn't affect connection management due to long flush times (#2424)

Docker images

• docker pull ethersphere/bee:1
• docker pull ethersphere/bee:1.1
• docker pull ethersphere/bee:1.1.0
• docker pull ethersphere/bee:stable
• docker pull quay.io/ethersphere/bee:1
• docker pull quay.io/ethersphere/bee:1.1
• docker pull quay.io/ethersphere/bee:1.1.0
• docker pull quay.io/ethersphere/bee:stable
• docker pull ethersphere/bee:1-slim
• docker pull ethersphere/bee:1.1-slim
• docker pull ethersphere/bee:1.1.0-slim
• docker pull ethersphere/bee:latest-slim
• docker pull ethersphere/bee:stable-slim
• docker pull quay.io/ethersphere/bee:1-slim
• docker pull quay.io/ethersphere/bee:1.1-slim
• docker pull quay.io/ethersphere/bee:1.1.0-slim
• docker pull quay.io/ethersphere/bee:latest-slim
• docker pull quay.io/ethersphere/bee:stable-slim