[
https://issues.apache.org/jira/browse/CASSANDRA-21264?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Caleb Rackliffe updated CASSANDRA-21264:
----------------------------------------
Change Category: Operability
Complexity: Normal
Component/s: Feature/2i Index
Feature/SAI
Fix Version/s: 5.x
5.0.x
Status: Open (was: Triage Needed)
Given the new system table, and the fact that it's not that hard to avoid the
problem here (and we could even prevent it with some guardrails?), I'm guessing
this should be a trunk-only fix. That's perhaps complicated by the move to
release 6.0 soon, but we'll see...
> Improve the Scalability of Secondary Index Status Propagation
> -------------------------------------------------------------
>
> Key: CASSANDRA-21264
> URL: https://issues.apache.org/jira/browse/CASSANDRA-21264
> Project: Apache Cassandra
> Issue Type: Improvement
> Components: Feature/2i Index, Feature/SAI
> Reporter: Caleb Rackliffe
> Priority: Normal
> Fix For: 5.x, 5.0.x
>
>
> Secondary indexes gossip their "queryability" status around the cluster to
> assist coordinators in selecting replica sets that are most likely to avoid
> availability gaps. However, the textual format used to gossip this
> information can only fit about 32K characters. If a schema contains several
> hundred secondary indexes, it's not that difficult to breach this limit, and
> the associated serialization failure can even prevent node startup.
> CASSANDRA-20058 gave us more headroom by streamlining the status format, but
> we eventually might need to consider using a distributed system table (or
> some similar mechanism) to bookkeep this information. We need to be able to
> maintain it incrementally, rather than as a monolithic status summary per
> endpoint.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
