You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/how/updating-datahub.md
+3-1Lines changed: 3 additions & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -170,9 +170,11 @@ Requirements:
170
170
171
171
### Known Issues
172
172
173
+
- **(Operations / Elasticsearch 8) Spurious mappings reindex on every Helm upgrade.** Deployments backed by **Elasticsearch 8.x** may trigger a full index mappings reindex on **every** system-update run (including each Helm upgrade), even when no authored mapping change was shipped. ES8 persists mappings differently than DataHub's authored definitions (for example injecting `type: object` on nested object fields), so the index builder detects spurious drift and schedules reindex work repeatedly. This is most likely when **`ELASTICSEARCH_INDEX_BUILDER_MAPPINGS_REINDEX=true`** (including Docker quickstart, which enables it for initial index builds). **Workaround:** Let the **first** system-update after install or upgrade finish so indices are created, then set **`ELASTICSEARCH_INDEX_BUILDER_MAPPINGS_REINDEX=false`** on **`datahub-gms`** and **`datahub-upgrade`** before later Helm upgrades until a fixed release is available.
174
+
173
175
### Potential Downtime
174
176
175
-
- **System-update / Elasticsearch:** Upgrades that trigger reindexing, incremental index migration, or optional Elasticsearch ZDU (#16887) can run for an extended period depending on catalog size. Reindex duration was improved (#16949) and a perpetual ES8 reindex loop fix was included (#17402). Helm upgrades without `--atomic` may report failure while background reindex jobs continue — allow system-update to finish before retrying. See **Helm Notes** under the [0.10.0](#0100) section below for timeout guidance patterns.
177
+
- **System-update / Elasticsearch:** Upgrades that trigger reindexing, incremental index migration, or optional Elasticsearch ZDU (#16887) can run for an extended period depending on catalog size. Reindex duration was improved (#16949). Elasticsearch 8 deployments with mappings reindex enabled may repeat unnecessary reindex work on every Helm upgrade — see **Known Issues** above. Helm upgrades without `--atomic` may report failure while background reindex jobs continue — allow system-update to finish before retrying. See **Helm Notes** under the [0.10.0](#0100) section below for timeout guidance patterns.
176
178
- **First deploy after bootstrap moves:** Entity Types ingestion moved to system-update (see Breaking Changes). Expect additional system-update work on the first upgrade after this release.
177
179
- **Aspect schema version sweep (#16930):** Large deployments may observe background migration activity during upgrade; plan maintenance windows accordingly.
Copy file name to clipboardExpand all lines: metadata-io/src/main/java/com/linkedin/metadata/search/elasticsearch/client/shim/impl/Es8SearchClientShim.java
Copy file name to clipboardExpand all lines: metadata-io/src/test/java/com/linkedin/metadata/search/elasticsearch/client/shim/Es8SearchClientShimConversionTest.java
0 commit comments