keith-turner commented on issue #2667:
URL: https://github.com/apache/accumulo/issues/2667#issuecomment-1271312449

   I was able to reproduce this yesterday and the situation seemed similar to 
the one above.  It seems like the same situation that was speculated as a cause 
in an earlier comment, a compaction of an empty tablet followed by a close.
   
   Saw the following in the tserver logs.
   
   ```
   2022-10-06T19:41:09,701 [tablet.location] DEBUG: Loading 3s<< on 
localhost:9997[1000325bae30004]
   2022-10-06T19:41:09,728 [tablet.location] DEBUG: Loaded 3s<< on 
localhost:9997[1000325bae30004]
   2022-10-06T19:41:09,830 [tablet.Tablet] DEBUG: Tablet 3s<< had no dir, 
creating hdfs://localhost:8020/accumulo/tables/3s/default_tablet
   2022-10-06T19:41:09,837 [tablet.Tablet] ERROR: Closed tablet 3s<< 
lastCompactID is inconsistent with metadata : 2 != -1
   2022-10-06T19:41:09,839 [tablet.Tablet] ERROR: Failed to do close 
consistency check for tablet 3s<<
   java.lang.RuntimeException: Closed tablet 3s<< lastCompactID is inconsistent 
with metadata : 2 != -1
   java.lang.RuntimeException: Failed to do close consistency check for tablet 
3s<<
   ```
   
   Saw the following the manager logs
   
   ```
   2022-10-06T19:41:09,520 [security.SecurityOperation] INFO : Granted table 
permission READ for user root on the table 3s at the request of user !SYSTEM
   2022-10-06T19:41:09,523 [security.SecurityOperation] INFO : Granted table 
permission WRITE for user root on the table 3s at the request of user !SYSTEM
   2022-10-06T19:41:09,526 [security.SecurityOperation] INFO : Granted table 
permission BULK_IMPORT for user root on the table 3s at the request of user 
!SYSTEM
   2022-10-06T19:41:09,528 [security.SecurityOperation] INFO : Granted table 
permission ALTER_TABLE for user root on the table 3s at the request of user 
!SYSTEM
   2022-10-06T19:41:09,530 [security.SecurityOperation] INFO : Granted table 
permission GRANT for user root on the table 3s at the request of user !SYSTEM
   2022-10-06T19:41:09,533 [security.SecurityOperation] INFO : Granted table 
permission DROP_TABLE for user root on the table 3s at the request of user 
!SYSTEM
   2022-10-06T19:41:09,537 [security.SecurityOperation] INFO : Granted table 
permission GET_SUMMARIES for user root on the table 3s at the request of user 
!SYSTEM
   2022-10-06T19:41:09,550 [tableOps.Utils] INFO : table 3s 
FATE[3d5c0cbb1211e4a9] locked for write operation: CREATE
   2022-10-06T19:41:09,550 [tables.TableManager] DEBUG: Creating ZooKeeper 
entries for new table bulk_dev_uk_2030907_1665085269455 (ID: 3s) in namespace 
(ID: +default)
   2022-10-06T19:41:09,560 [manager.EventCoordinator] INFO : Table state in 
zookeeper changed for 3s to NEW
   2022-10-06T19:41:09,581 [balancer.TableLoadBalancer] INFO : Loaded class 
org.apache.accumulo.core.spi.balancer.SimpleLoadBalancer for table 3s
   2022-10-06T19:41:09,603 [tables.TableManager] DEBUG: Transitioning state for 
table 3s from NEW to ONLINE
   2022-10-06T19:41:09,604 [manager.EventCoordinator] INFO : Table state in 
zookeeper changed for 3s to ONLINE
   2022-10-06T19:41:09,613 [tableOps.Utils] INFO : table 3s 
FATE[3d5c0cbb1211e4a9] unlocked for write
   2022-10-06T19:41:09,680 [accumulo.audit] INFO : operation: permitted; user: 
root; client: 127.0.0.1:53720; action: compactTable; targetTable: 3s; 
targetNamespace: +default;
   2022-10-06T19:41:09,685 [accumulo.audit] INFO : operation: permitted; user: 
root; client: 127.0.0.1:53712; action: compactTable; targetTable: 3s; 
targetNamespace: +default;
   2022-10-06T19:41:09,700 [tablet.location] DEBUG: Assigned 3s<< to 
localhost:9997[1000325bae30004]
   2022-10-06T19:41:09,730 [manager.EventCoordinator] INFO : tablet 3s<< was 
loaded on localhost:9997
   2022-10-06T19:41:09,731 [tableOps.Utils] INFO : table 3s 
FATE[68a7eab3013b5aaa] locked for read operation: COMPACT
   2022-10-06T19:41:09,751 [tableOps.Utils] INFO : table 3s 
FATE[71a8f902cf66da2a] locked for read operation: COMPACT
   2022-10-06T19:41:11,854 [schema.LinkingIterator] DEBUG: First tablet seen 
provides evidence of earlier tablet in range, retrying 3s;r0e773 [3s; : [] 
9223372036854775807 false,3s<%00; : [] 9223372036854775807 false) 
   2022-10-06T19:41:12,124 [manager.EventCoordinator] INFO : localhost:9997 
reported split 3s;r0e773<, 3s<;r0e773
   2022-10-06T19:41:13,897 [schema.LinkingIterator] DEBUG: First tablet seen 
provides evidence of earlier tablet in range, retrying 3s;r057b6 [3s; : [] 
9223372036854775807 false,3s<%00; : [] 9223372036854775807 false) 
   2022-10-06T19:41:14,123 [manager.EventCoordinator] INFO : localhost:9997 
reported split 3s;r057b6<, 3s;r0e773;r057b6
   2022-10-06T19:41:14,905 [manager.EventCoordinator] INFO : localhost:9997 
reported split 3s;r11bf7;r0e773, 3s<;r11bf7
   2022-10-06T19:41:16,423 [manager.EventCoordinator] INFO : localhost:9997 
reported split 3s;r0af73;r057b6, 3s;r0e773;r0af73
   2022-10-06T19:41:16,902 [schema.LinkingIterator] DEBUG: Tablets end row and 
prev end row not equals 3s;r11bf7;r0e773 3s<;r1391e 
   2022-10-06T19:41:17,425 [manager.EventCoordinator] INFO : localhost:9997 
reported split 3s;r1391e;r11bf7, 3s<;r1391e
   2022-10-06T19:41:18,168 [accumulo.audit] INFO : operation: permitted; user: 
root; client: 127.0.0.1:53712; action: mergeTable; targetTable: 3s; 
targetNamespace: +default;
   2022-10-06T19:41:18,168 [manager.Manager] DEBUG: Creating merge op: 3s from 
startRow: -inf to endRow: +inf
   2022-10-06T19:41:18,237 [tableOps.Utils] INFO : table 3s 
FATE[57d4471744875c60] locked for write operation: MERGE
   2022-10-06T19:41:18,242 [manager.EventCoordinator] INFO : Merge state of 
3s<< set to STARTED
   2022-10-06T19:41:18,382 [manager.Manager] DEBUG: mergeInfo overlaps: 
3s;r057b6< true
   ```
   
   I added some logging to the tserver that will help verify the hypothesis and 
kicked the test off again.  Didn't see the problem again, but did see #3006.  
Also noticed some of the bulk import random walk test are stuck this morning i 
have not had a chance to track that down. Going to keep trying to reproduce the 
problem with improved logging.


-- 
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]

Reply via email to