[
https://issues.apache.org/jira/browse/CASSANDRA-17062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17747468#comment-17747468
]
Sam Tunnicliffe commented on CASSANDRA-17062:
---------------------------------------------
{quote}I was trying to model how one table with both cache types would look
like and there are two columns which are not in unweighted (capacity_bytes and
size_bytes) and one in unweighted which is not in weighted (capacity) so we
would need to come up with some "n/a" values if we combine that and that looks
even worse.
{quote}
Right, this is why I proposed distinct tables for these.
{quote}Also, if mutation is going to be introduced and one e.g. do TRUNCATE, I
do not think that an operator truly wants to invalidate both caches at once
{quote}
However many tables we have, unless it's one per cache then TRUNCATE is never
likely to be the right verb for mutating. I would think operators will want to
invalidate on a per-cache basis, so would need to mutate individual rows in the
table(s).
I'm a bit confused what you mean by your final sentence:
{quote}I think we just need to go with the original proposal - to name in
"unweighted_caches" and so to dedicate a separate table for that.
{quote}
Do you mean you prefer to implement this from the Aleksei's first comment
above:
{quote}1. Introduce "weighted" column to the VT and add corresponding notes to
the documentation on how to treat capacity and size for weighted/non-weighted
caches. But we can mention that without having the new column.
{quote}
One of the reasons I'm not keen on this approach is that it introduces
dependencies between columns in a single row and consumers of the data have to
understand those implicit dependencies in order to
interpret/transform/normalise those values. e.g. the semantics/type of a column
should be the same for all rows in the table, not varying on a row-by-row basis.
> Expose Auth Caches metrics
> --------------------------
>
> Key: CASSANDRA-17062
> URL: https://issues.apache.org/jira/browse/CASSANDRA-17062
> Project: Cassandra
> Issue Type: Improvement
> Components: Feature/Virtual Tables, Observability/Metrics,
> Tool/nodetool
> Reporter: Aleksei Zotov
> Assignee: Stefan Miklosovic
> Priority: Normal
> Fix For: 5.x
>
>
> Unlike to other caches (row, key, counter), Auth Caches lack some monitoring
> capabilities. Here are a few particular changes to get this inequity fixed:
> # Add auth caches to _system_views.caches_ VT
> # Expose auth caches metrics via JMX
> # Add auth caches details to _nodetool info_
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
