Closed robinbb closed 2 days ago
Probably related to lack of database flush. See #1243
2024-07-03T14:40:51.431041+00:00 - INFO Using default mainnet genesis ledger
2024-07-03T14:40:51.433502+00:00 - DEBUG Creating a new IndexerStore in /mnt/mina-indexer-prod/db/0.7.0-340000
2024-07-03T15:24:46.008149+00:00 - DEBUG INDEXER_VERSION file exists. Checking for compatability
2024-07-03T15:24:46.010012+00:00 - INFO Starting the web server listening on 0.0.0.0:8080
2024-07-03T15:24:46.010107+00:00 - INFO Starting mina-indexer server
The above shows the duration of the log recovery between lines 3 and 4 - it is 44 minutes.
This is solved sufficiently well that it should be de-prioritized. (The snapshot-after-ingestion fix helped.)
Actually, I think it better to actually close this issue, and simply re-open if the problem re-emerges.
When many database .log files exist at the time of the shutdown, then the subsequent invocation of the indexer must recover the logs, resulting in a long startup time. The database LOG file fills with thousands of entries like so: