Influxdata
/
influxdb
mirror of https://github.com/influxdata/influxdb.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

docs: Update size calculations in the description of PartitionCache

pull/24376/head
Carol (Nichols || Goulding) 2023-06-21 13:50:55 -04:00
parent 0d9f89ae48
commit 1912840c25
No known key found for this signature in database
GPG Key ID: E907EE5A736F87D4
1 changed files with 7 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
ingester/src/buffer_tree/partition/resolver/cache.rs
View File

@ -20,10 +20,10 @@ use crate::{
};
/// A read-through cache mapping `(table_id, partition_key)` tuples to
/// `(partition_id, max_sequence_number)`.
/// `(partition_id, optional_partition_hash_id)`.
///
/// This data is safe to cache as only one ingester is ever responsible for a
/// given table partition, and this amortises partition persist marker discovery
/// given table partition, and this amortises partition discovery
/// queries during startup, eliminating them from the ingest hot path in the
/// common startup case (an ingester restart with no new partitions to add).
///
@ -32,11 +32,12 @@ use crate::{
/// Excluding map overhead, and assuming partition keys in the form
/// "YYYY-MM-DD", each entry takes:
///
/// - Partition key: String (8 len + 8 cap + 8 ptr + data len) = 34 bytes
/// - TableId: 8 bytes
/// - PartitionId: 8 bytes
/// - `PartitionKey`: String (8 len + 8 cap + 8 ptr + data len) = 34 bytes
/// - `TableId`: 8 bytes
/// - `PartitionId`: 8 bytes
/// - `Option<PartitionHashId>`: 8 bytes
///
/// For a total of 50 bytes per entry - approx 20,971 entries can be held in
/// For a total of 58 bytes per entry - approx 18.078 entries can be held in
/// 1MiB of memory.
///
/// Each cache hit _removes_ the entry from the cache - this eliminates the