Open tzach opened 1 month ago
When running a full restore on the clusters, there is no need to limit resource consumption, which is there for backup and partial restore cases.
Examples: batch-size can be increased parallel set to zero etc
batch-size
parallel
Another recommendation is to restore to a bigger cluster and remove nodes later once the cluster with restored data is up and running.
In terms of batch size, its impact on performance is already a part of the docs, but the default value should be adjusted (https://github.com/scylladb/scylla-manager/issues/3809).
When running a full restore on the clusters, there is no need to limit resource consumption, which is there for backup and partial restore cases.
Examples:
batch-size
can be increasedparallel
set to zero etcAnother recommendation is to restore to a bigger cluster and remove nodes later once the cluster with restored data is up and running.