DEBUG_BIG_RO_TX_KB - print logs with info about large read-only transactions
DEBUG_BIG_RW_TX_KB - print logs with info about large read-write transactions
DEBUG_SLOW_COMMIT_MS - print logs with commit timing details if commit is slower than this threshold
* Only insert hard-coded tips if both DB and files recovery failed
* Prevent deadlock
* Report some efficiency numbers
* Count properly unrequested bodies
* Initialise allRequests
* Increase timeout
* Fixes to scheduling
* Small fix
* Simplified scheduling
* Remove separate bodyLoop goroutine
* Update from DB at the beginnig of block bodies forward
* Timeout for repeating the request cycle
* Fix timeout
* Fix
* Increase timeout
* Increase timeout
* Try to make flow if possible
* Fix flow
* Lower timeout
* timeout for each blockNum
* Adjustable timeout
* Better log timing
* Track peers
* copy peerID
* fix scheduling
* Too much logging
* Print delivery speed
* Print committed blocks
* Fix race
* Sentry to only reset back-off timer when response to a request is received
* Print bytes/sec and wasted traffic
* Fix bandwidth accounting
* Less logging
* not to wake up on deliveries
* Spam every second
* Print peer map
* Fix npe, print requests
* Timestamps
* Improved logging
* Penalty for peers
* Log penalties only when disconnecting
* Try with smaller window
* window parameter
* Dealing with partially delivered requests
* Init bodyReq
* Fix array index
* More fix for NPE
* More NPE checks
* Print out body progress and header progress
* Fix ending condition
* Bring back waking up and penalties
* Fix duplicate tip
* Duplicate segment
* Fix lint
* Fix lint
* fix lint
* Fix lint
* Hard coded headers in the source files
* Fix lint
* Replace hix-sized header serialisation with rlp (to support clique)
* Remove anchor records from the files
* Fixeds for DB recovery
* Fix compilation
* Fix compile errors
* Fix formatting
* Fix lint
* Fix comments
* Remove headerLoop
* Properly terminate body download
* Support for StatusData p2p proto
* Fix forkid test
* Fix test
* Fix lint
Co-authored-by: Alexey Sharp <alexeysharp@Alexeys-iMac.local>
* Remove hashed changesets
* try this
* oops
* simplify
* better diagnostics
* don't hash account key and use bitmap 64 to check history index
* update docs of bucket
Co-authored-by: alex.sharov <AskAlexSharov@gmail.com>
* move experiments to new branch&reorganise kv_snapshot
* walk&modify tests
* added delete from snapshot tests
* fmt
* state snapshot debug
* snapshot validation passed. copy state snapshot
* debug
* snapshot cursor.Prev test
* Prev works correct. Added Current check
* add err check
* added walk forward and backward test
* before refactoring
* refactoring
* execution with snapshot debug
* fix
* remove useless test
* before dupcursor implimentation
* tests with prev and delete works
* execution based on state snapshot passed
* remove useless tests
* blocks to 1140000 passed
* clean verifier
* cleanup state generation
* clean verify && seeder
* remove debug code
* tests passed
* fix lint
* save state
* test passed
* fix lint
* add state hash
* fix lint
* rpcdaemon performance improve - less reading blocks, less allocs on server
* rpcdaemon performance improve - less reading blocks, less allocs on server
* rpcdaemon performance improve - less reading blocks, less allocs on server
* rpcdaemon performance improve - less reading blocks, less allocs on server
* don't use global variables
* df
* avoid use non-transactional db instance to reduce amount of cursors
