seawinde commented on code in PR #66255:
URL: https://github.com/apache/doris/pull/66255#discussion_r3709988143
##########
fe/fe-core/src/main/java/org/apache/doris/datasource/InternalCatalog.java:
##########
@@ -3771,9 +3771,11 @@ private List<Partition> truncateTableInternal(OlapTable
olapTable, List<Partitio
olapTable.dropPartitionForTruncate(olapTable.getDatabase().getId(),
isforceDrop, pair.getValue());
}
- // Reset table-level visibleVersion to TABLE_INIT_VERSION so it stays
consistent
- // with the newly created partitions (which also start at
PARTITION_INIT_VERSION).
- olapTable.resetVisibleVersion();
+ // Truncate changes visible data without changing the table ID. Keep
the table version
+ // monotonic so snapshot consumers cannot mistake new data for an old
version.
+ if (Config.isNotCloudMode()) {
Review Comment:
Fixed in commit 4744410e13c. Cloud now takes the legacy/reset branch after
the local partition swap, so visibleVersionTime changes synchronously and
fences stale local SimpleAgg cache entries. The authoritative Cloud table
version remains Meta Service-managed; only non-Cloud applies the persisted
positive version/time.
##########
fe/fe-core/src/main/java/org/apache/doris/datasource/InternalCatalog.java:
##########
@@ -3771,9 +3771,11 @@ private List<Partition> truncateTableInternal(OlapTable
olapTable, List<Partitio
olapTable.dropPartitionForTruncate(olapTable.getDatabase().getId(),
isforceDrop, pair.getValue());
}
- // Reset table-level visibleVersion to TABLE_INIT_VERSION so it stays
consistent
- // with the newly created partitions (which also start at
PARTITION_INIT_VERSION).
- olapTable.resetVisibleVersion();
+ // Truncate changes visible data without changing the table ID. Keep
the table version
+ // monotonic so snapshot consumers cannot mistake new data for an old
version.
+ if (Config.isNotCloudMode()) {
+ olapTable.updateVisibleVersionAndTime(olapTable.getNextVersion(),
System.currentTimeMillis());
Review Comment:
Addressed the journal compatibility pieces in commit 4744410e13c: the
non-Cloud version/time are selected while holding the table lock, persisted in
TruncateTableInfo, and replayed exactly. Missing fields deserialize to 0, so a
new FE replaying an old record preserves the legacy reset behavior. I did not
add a new electable-FE capability protocol because there is no existing journal
capability gate that can reliably prove every electable FE understands these
fields. An old binary will ignore the new fields, so mixed-version safety still
relies on the supported upgrade/election ordering: an old FE must not be
promoted after a new master starts emitting the new record. Adding distributed
capability negotiation would be a broader upgrade-framework change rather than
part of this table-version fix.
##########
regression-test/suites/mtmv_p0/test_truncate_table_mtmv.groovy:
##########
@@ -71,6 +86,18 @@ suite("test_truncate_table_mtmv","mtmv") {
REFRESH MATERIALIZED VIEW ${mvName} AUTO
"""
waitingMTMVTaskFinishedByMvName(mvName)
+
+ // Save a non-PCT table snapshot at version 3 while the MV does not
contain k2=2.
Review Comment:
Updated in commit 4744410e13c: order_qt_truncate_partition is restored
immediately after the first AUTO refresh wait, and the non-PCT ABA scenario now
has a distinct order_qt_non_pct_truncate assertion. The corresponding .out
entry still needs to be generated by the regression runner; repository policy
forbids hand-writing it, and the owner-managed FE restart required by the
regression gate has not been confirmed, so I am leaving this thread open until
that artifact is generated.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
