The backend storage ES index could not be found[Bug] #12390
Replies: 2 comments 1 reply
-
AFAIK, no, we don't use that. Because in the large scale, it is slow. I noticed you set TTL=15, and Day_step is large 10. I don't know whether this works. TTL and DAY STEP should not be so close. TTL should at least 3 times larger of DAY_STEP. |
Beta Was this translation helpful? Give feedback.
-
Another phenomenon that arises when multiple OAP instances share the same backend Elasticsearch (ES) cluster account and rely solely on "different index name prefixes" to differentiate between OAP services is the issue of "one OAP service querying another OAP service's indexes." For instance, if: OAP-A is configured with SW_NAMESPACE=test, |
Beta Was this translation helpful? Give feedback.
-
Search before asking
Apache SkyWalking Component
OAP server (apache/skywalking)
What happened
The backend storage ES index could not be found
OAP version:9.7.0
configs:
SW_CORE_RECORD_DATA_TTL=15
SW_CORE_METRICS_DATA_TTL=15
What you expected to happen
The index data configuration retains data for 15 days. After 15 days, Elasticsearch creates a new index and clears the old index data. The indices
sw-sys00008-cgpt-prod_metrics-all-20240623
andsw-sys00008-cgpt-prod_metrics-all-20240613
are created and the old index data is purged, but OAP (OpenAgent Platform or similar) is still requesting data from the old index. Shouldn't OAP normally use index aliases to query data instead of directly referencing the index names?How to reproduce
Use OAP version 9.7.0, es version 7.16.2 and use the following configuration
Anything else
Specify the retention period for index data
Are you willing to submit a pull request to fix on your own?
Code of Conduct
Beta Was this translation helpful? Give feedback.
All reactions