Was this page helpful?
Caution
You're viewing documentation for an unstable version of ScyllaDB Open Source. Switch to the latest stable version.
Due to its thread-per-core architecture, many things within ScyllaDB can be better understood when you look at it on a per-CPU basis. There are Linux tools such as top and perf that can give information about what is happening within a CPU, given a CPU number.
A common mistake users make is to assume that there is a direct and predictable relationship between the ScyllaDB Shard ID and the CPU ID, which is not true.
Starting in version 3.0, ScyllaDB ships with a script to let users know about the mapping between CPUs and ScyllaDB Shards. For users of older versions, a copy of the script can be downloaded from the Seastar git tree.
To list the mapping of a specific shard:
seastar-cpu-map.sh -n scylla -s 0
Output:
shard: 0, cpu: 1
To list the mapping of all shards:
seastar-cpu-map.sh -n scylla
Output:
shard: 0, cpu: 1
shard: 1, cpu: 2
shard: 3, cpu: 4
shard: 2, cpu: 3
Was this page helpful?