[
https://issues.apache.org/jira/browse/CASSANDRA-18111?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Stefan Miklosovic updated CASSANDRA-18111:
------------------------------------------
Status: Patch Available (was: In Progress)
https://github.com/apache/cassandra/pull/3374 passes java 17 pre-commit CI, I
think this is something we can consider to be reviewable.
I tried to take quite a holistic approach so it may seem as I restructured a
lot of stuff. One not-so-obvious change is that we were clearing ephemeral
snapshots as part of the startup sequence as part of startup checks and the
logic was done is such a way that we were not using SnapshotManager because it
was present as a singleton in StorageService and it was way early for that.
That meant that we were trying to identify snapshots "manually" / by other
means as by SnapshotManager. I rewrote that so SnapshotManager is a standalone
"service" and the only source of truth for snapshot management globally. That
also removed a lot of stuff from e.g. Directories which contained methods
dealing with snapshots, they are not there anymore.
I will try to do some perf testing on this but I can't put any timestamp on
that.
SnapshotManager is also integrated into everything where we deal with snapshots
as such.
> Centralize all snapshot operations to SnapshotManager and cache snapshots
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-18111
> URL: https://issues.apache.org/jira/browse/CASSANDRA-18111
> Project: Cassandra
> Issue Type: Improvement
> Components: Local/Snapshots
> Reporter: Paulo Motta
> Assignee: Stefan Miklosovic
> Priority: Normal
> Fix For: 5.x
>
> Time Spent: 1h 10m
> Remaining Estimate: 0h
>
> Everytime {{nodetool listsnapshots}} is called, all data directories are
> scanned to find snapshots, what is inefficient.
> For example, fetching the
> {{org.apache.cassandra.metrics:type=ColumnFamily,name=SnapshotsSize}} metric
> can take half a second (CASSANDRA-13338).
> This improvement will also allow snapshots to be efficiently queried via
> virtual tables (CASSANDRA-18102).
> In order to do this, we should:
> a) load all snapshots from disk during initialization
> b) keep a collection of snapshots on {{SnapshotManager}}
> c) update the snapshots collection anytime a new snapshot is taken or cleared
> d) detect when a snapshot is manually removed from disk.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
