milleruntime commented on issue #1872: URL: https://github.com/apache/accumulo/issues/1872#issuecomment-766961695
I did some testing with Uno running MultiTable RW and have yet to see the log statement I added in #1881. I was able to see the delete FATE `CleanUp` waiting briefly and it was usually for compactions. So I think your theory of waiting for compactions is more likely. From the master log: <pre> 2021-01-25 10:16:20,209 [tableOps.CleanUp] DEBUG: Still waiting for table(1k) to be deleted; Target tablet state: UNASSIGNED, Current tablet state: HOSTED, locationState: 1k<;9@(null,localhost:11000[1000072ccc80003],localhost:11000[1000072ccc80003]) </pre> From tserver log: <pre> 2021-01-25 10:16:19,508 [tablet.MinorCompactor] DEBUG: Begin minor compaction hdfs://localhost:8020/accumulo/tables/1k/default_tablet/F0000265.rf_tmp 1k<;9 2021-01-25 10:16:20,202 [tablet.Tablet] DEBUG: initiateClose(saveState=false queueMinC=false disableWrites=false) 1k<;9 2021-01-25 10:16:20,202 [tablet.Tablet] DEBUG: completeClose(saveState=false completeClose=true) 1k<;9 2021-01-25 10:16:21,375 [tablet.Compactor] DEBUG: Compaction 1k<;9 14,027 read | 14,027 written | 7,889 entries/sec | 1.778 secs | 47,299,377 bytes | 26602574.241 byte/sec 2021-01-25 10:16:21,377 [tablet.Tablet] DEBUG: Logs for memory compacted: 1k<;9 localhost:11000/hdfs://localhost:8020/accumulo/wal/localhost+11000/bbbd0052-911d-4204-990b-ddfef97738bf 2021-01-25 10:16:21,377 [tablet.Tablet] DEBUG: Logs to be destroyed: 1k<;9 localhost:11000/hdfs://localhost:8020/accumulo/wal/localhost+11000/bbbd0052-911d-4204-990b-ddfef97738bf 2021-01-25 10:16:21,386 [tablet.DatafileManager] TABLET_HIST: 1k<;9 MinC [memory] -> hdfs://localhost:8020/accumulo/tables/1k/default_tablet/F0000265.rf 2021-01-25 10:16:21,386 [tablet.DatafileManager] DEBUG: MinC finish lock 0.00 secs 1k<;9 2021-01-25 10:16:21,436 [tablet.Tablet] TABLET_HIST: 1k<;9 closed 2021-01-25 10:16:21,437 [tserver.TabletServer] DEBUG: Unassigning 1k<;9@(null,localhost:11000[1000072ccc80003],null) 2021-01-25 10:16:21,441 [tserver.TabletServer] INFO : unloaded 1k<;9 </pre> While looking at this code path, I also found another point where it could be waiting and that was for active scans during the `Tablet.close()`. https://github.com/apache/accumulo/blob/9c957ab55f28f10dde1321eb613a9ec3e2fa5322/server/tserver/src/main/java/org/apache/accumulo/tserver/tablet/Tablet.java#L1439-L1446 ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected]
