Long Migrations
Those are applied only when you are upgrading from the previous release.
Receipts RLP -> CBOR
For better eth_getLogs performance.
Duration: about 6 hours on an SSD.
Log Indices Re-Creation:
For better eth_getLogs performance.
This one drops the old index during the migration and re-creates during the sync. So one of the sync stages will take longer.
Duration: 2-3 hours.
--
Total migrations duration: about 8-9 hours.
Changelog
-
Alpine support in Docker is restored
-
Fix a syncing deadlock
-
Fix stability and disk space leaks of the DB when big migrations happen
-
Geth 1.9.22
-
New options to select the db batch size (lower values can improve performance, especially on slow storages like HDD)
-
More documentation at https://turbo-geth.readthedocs.io
-
RPC and main node stability improvements under load
-
RPC improvements:
-
RPC stability under load
-
new methods and
tg_RPC namespace: see more -
fix
debug_getModifiedAccountsByNumber
-
-
Logging improvements
SHA256
| tool | sha256 | validation command |
|---|---|---|
| ./bin/tg | 2317e68f79137bb32e52bead6a695cb56490daf25d59e161f92a3636aaf8aeb4
| make docker && docker run turbo-geth:latest sha256sum /usr/local/bin/tg
|
| ./bin/rpcdaemon | ed076902a023a79ed33b52bcd18beffd6e8888e7cf78032b3319b671839c6783
| make docker && docker run turbo-geth:latest sha256sum /usr/local/bin/rpcdaemon
|
It should be the same if building on Linux with go 1.15.