Splice 0.5.8
Note: 0.5.7 introduced a significant performance regression related to the processing of topology transactions on participants, mediators, and sequencers. Please skip 0.5.7 and upgrade directly to 0.5.8. Fix performance…
- A newer release in the 0.5 line: Splice 0.5.18
- Newest Splice: Splice 0.9.4
Open the release notes (opens in a new tab)docs.canton.network
Note: 0.5.7 introduced a significant performance regression related to the processing of topology transactions on participants, mediators, and sequencers. Please skip 0.5.7 and upgrade directly to 0.5.8.
- Canton
- Fix performance regression related to the processing of topology transactions.
- Improve performance of some queries that are used in participant pruning. The fix includes a database migration which can take up to 2min but should be faster on most participants.
- Scan
- deprecated
/v0/total-amulet-balanceand/v0/wallet-balanceendpoints have been removed in favor of using /registry/metadata/v1/instruments/ and /v0/holdings/summary, respectively.
- deprecated
- Deployments
-
The default logger has been switched to use an asynchronous appender, for all the nodes, for better performance. The behavior can be switched back to synchronous logging by setting the environment variable
LOG_IMMEDIATE_FLUSH=true. This now includes helm deployments as well, in 0.5.7 the default was changed only for docker-compose deployments. -
This version breaks backwards compatibility with migration dumps taken on 0.4.x versions. Please make sure that you are deploying with
migrating: false(helm) / without-M(docker-compose). For helm-based validator deployments: In some cases, helm might not properly update the state after you removed themigratingflag. You can check before the upgrade if it got properly applied throughkubectl describe deployment -n validator validator-appand look for this env var:If you see it, the deployment still hasmigrating: trueactivated. You can clear that flag by, for example, uninstalling and reinstalling the validator helm release (but not participant and postgres). You can either directly reinstall the new version or first do the reinstall on the old version and then upgrade.
-
The default logger has been switched to use an asynchronous appender, for all the nodes, for better performance. The behavior can be switched back to synchronous logging by setting the environment variable