CASSANALYTICS-194: Parse <replicas>/<transient> replication factor fo… by mansikhara · Pull Request #238 · apache/cassandra-analytics

mansikhara · 2026-08-28T22:04:21Z

…r witness-enabled keyspaces

Cassandra accepts a replication factor of the form /, which witness replicas under mutation tracking reuse, so a witness-enabled keyspace declares 'datacenter1': '3/1' meaning three replicas of which one is a witness.

CqlUtils.extractReplicationFactor() passed each datacenter value to Integer.parseInt, so a bulk read against any such keyspace failed during job setup with an uncaught NumberFormatException.

Transient counts are now tracked per datacenter alongside the existing totals and exposed through getFullReplicationFactor(), getTransientReplicationFactor(), getFullReplicas(dc), getTransientReplicas(dc) and hasTransientReplicas(). getTotalReplicationFactor() keeps its existing meaning of all replicas including witnesses, so behaviour for untracked keyspaces is unchanged. Parsing applies the same constraints Cassandra enforces in locator.ReplicationFactor.validate. A new parseStrict factory reports an unparseable or empty replication map at parse time rather than dropping the datacenter and failing later with a misleading "DC not found in replication factor"; the lenient constructor is retained unchanged for CDC callers. The Kryo serializer and CassandraRing's hand-rolled JDK readObject/writeObject are updated so the new field survives serialization to Spark executors.

Prerequisite for CASSANALYTICS-164.

…r witness-enabled keyspaces Cassandra accepts a replication factor of the form <replicas>/<transient>, which witness replicas under mutation tracking reuse, so a witness-enabled keyspace declares 'datacenter1': '3/1' meaning three replicas of which one is a witness. CqlUtils.extractReplicationFactor() passed each datacenter value to Integer.parseInt, so a bulk read against any such keyspace failed during job setup with an uncaught NumberFormatException. Transient counts are now tracked per datacenter alongside the existing totals and exposed through getFullReplicationFactor(), getTransientReplicationFactor(), getFullReplicas(dc), getTransientReplicas(dc) and hasTransientReplicas(). getTotalReplicationFactor() keeps its existing meaning of all replicas including witnesses, so behaviour for untracked keyspaces is unchanged. Parsing applies the same constraints Cassandra enforces in locator.ReplicationFactor.validate. A new parseStrict factory reports an unparseable or empty replication map at parse time rather than dropping the datacenter and failing later with a misleading "DC not found in replication factor"; the lenient constructor is retained unchanged for CDC callers. The Kryo serializer and CassandraRing's hand-rolled JDK readObject/writeObject are updated so the new field survives serialization to Spark executors. CassandraRing gains a pinned serialVersionUID and a format-version byte, because the class signature did not change and an older stream would otherwise be silently misread rather than rejected. Prerequisite for CASSANALYTICS-164. patch by Mansi Khara; reviewed by TBD for CASSANALYTICS-194

sarankk

Thanks Mansi, left few comments

sarankk · 2026-09-15T16:03:37Z

 *     }
+ * <p>
+ * Replica counts may also use the {@code <replicas>/<transient>} form, e.g. {@code "DC1" : "3/1"}, meaning three
+ * replicas of which one is transient. Witness replicas under mutation tracking (CEP-45/CEP-46) reuse this form, so


Nit: Transient replication was added before MT, it was updated to work on top of MT later. Also why do we use term reuse here?

Suggested change

* replicas of which one is transient. Witness replicas under mutation tracking (CEP-45/CEP-46) reuse this form, so

* replicas of which one is transient. Witness replicas (CEP-46) reuse this form, so

Good catch, you are right that transient replication came first. Applied your suggestion, thanks!

sarankk · 2026-09-15T16:28:03Z

+    private final Map<String, Integer> transientOptions;

+    /**
+     * Lenient parse: a replication value that cannot be parsed is logged and its datacenter omitted. Retained for


In what scenario do we anticipate using lenient parse?

None, as it turns out. I checked the callers and there are only two, ReplicationFactorSupplier and CdcOptions, and both build hardcoded maps that cannot fail to parse. So I have removed the lenient path entirely and made strict the only behaviour! Thanks for the find

sarankk · 2026-09-15T17:17:03Z

+     * plain {@code <replicas>} form, which keeps behaviour identical for those keyspaces.
+     */
+    @NotNull
+    private final Map<String, Integer> transientOptions;


Options hold dc : 3 and transient options hold dc : 1 for transient replication dc: 3/1. The maps could diverge. Shall we combine the 2 maps into 1 using ReplicaCounts?

Agreed and done, thanks!

sarankk · 2026-09-15T17:19:20Z

+    private ReplicationFactor(@NotNull Map<String, String> options, boolean strict)
    {
        this.replicationStrategy = ReplicationFactor.ReplicationStrategy.getEnum(options.get("class"));
        this.options = new LinkedHashMap<>(options.size());


If we plan to merge the maps, shall we make this immutable ?

sarankk · 2026-09-15T17:21:43Z

+        ReplicationFactor replicationFactor = new ReplicationFactor(ImmutableMap.of(
+        "class", "NetworkTopologyStrategy",
+        "datacenter1", "3/3"));
+        assertThat(replicationFactor.getOptions()).doesNotContainKey("datacenter1");


We should throw in this scenario instead, if Cassandra does not allow it, Sidecar should fail as well.

Great suggestions! Done.

sarankk · 2026-09-15T17:25:17Z

+        ReplicationFactor replicationFactor = new ReplicationFactor(ImmutableMap.of(
+        "class", "NetworkTopologyStrategy",
+        "datacenter1", "-3"));
+        assertThat(replicationFactor.getOptions()).doesNotContainKey("datacenter1");


Should we throw in this scenario as well?

Done, thanks!

sarankk · 2026-09-15T17:29:26Z

+        // Witness-enabled keyspace as created by Cassandra's WitnessAlwaysReadsFullReplicaTest on the
+        // cep-45-mutation-tracking branch: the <replicas>/<transient> form plus replication_type = 'tracked'
+        String schema = "CREATE KEYSPACE witnessks WITH REPLICATION = {'class': 'NetworkTopologyStrategy', "
+                        + "'datacenter1': '3/1', 'datacenter2': '3/1'} AND replication_type = 'tracked' "


replication_type = 'tracked' this field is not yet supported in Sidecar schema response, shall we remove this from example?

You are right, I confirmed it. Removed it from these fixtures since it is irrelevant to replication factor parsing anyway.

…ictly Combines the per-datacenter total and transient counts into a single Map<String, ReplicaCounts> instead of two parallel maps. The maps could previously drift apart, since nothing tied options={dc1=3} to transientOptions={dc1=1}; holding them together makes that unrepresentable. ReplicaCounts becomes a public nested type exposing allReplicas, fullReplicas and transientReplicas. getOptions() and getTransientOptions() are retained, now derived and unmodifiable, so CassandraRing, ConsistencyLevel and PartitionedDataLayer are unaffected. Removes the lenient parse. Its only two callers built hardcoded maps that cannot fail to parse, so it was never needed, and silently dropping a datacenter that Cassandra itself would reject is the wrong default: 3/3 and a negative replication factor now raise rather than being skipped. parseStrict therefore disappears as a separate factory, since strict is the only behaviour. Corrects the javadoc: transient replication predates mutation tracking, and it is witness replicas (CEP-46) that reuse the form. Drops replication_type from the replication factor test fixtures, since Sidecar builds its schema response from the driver's exportAsString() which does not emit that property, and it is irrelevant to replication factor parsing.

…Strict factory

sarankk reviewed Sep 15, 2026

View reviewed changes

mansikhara force-pushed the mutation-tracking-support branch from e2d75d0 to 3a07399 Compare September 15, 2026 20:45

CASSANALYTICS-194: Remove stale comment referencing the deleted parse…

9d7b04f

…Strict factory

	* replicas of which one is transient. Witness replicas under mutation tracking (CEP-45/CEP-46) reuse this form, so
	* replicas of which one is transient. Witness replicas (CEP-46) reuse this form, so

Uh oh!

Conversation

mansikhara commented Aug 28, 2026

Uh oh!

sarankk left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sarankk Sep 15, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

sarankk Sep 15, 2026 •

edited

Loading