Details
Description
I ran continuous ingest on a small cluster, stopped ingest, and then restarted Accumulo. After restart I looked last and current locations of tablets in the metadata table. There were 2k tablets and I just spot checked a few random places, but nothing seemed to be assigned to the last loc.
root@accumulo> scan -t accumulo.metadata -b 2 -c loc,last 2;001015b last:142f897d61d0004 [] ip-10-1-3-16:9997 2;001015b loc:342f897d5cd000d [] ip-10-1-3-27:9997 2;002028 last:242f897d5dd0002 [] ip-10-1-3-28:9997 2;002028 loc:142f897d61d01ae [] ip-10-1-3-20:9997 2;00303c last:342f897d5cd0004 [] ip-10-1-3-25:9997 2;00303c loc:142f897d61d01b1 [] ip-10-1-3-23:9997 2;00404a last:342f897d5cd0004 [] ip-10-1-3-25:9997 2;00404a loc:142f897d61d01b5 [] ip-10-1-3-26:9997 2;00505b last:142f897d61d0004 [] ip-10-1-3-16:9997 2;00505b loc:342f897d5cd000d [] ip-10-1-3-27:9997 2;006066 last:142f897d61d0004 [] ip-10-1-3-16:9997 2;006066 loc:342f897d5cd000c [] ip-10-1-3-29:9997 2;007073a last:142f897d61d0004 [] ip-10-1-3-16:9997 2;007073a loc:342f897d5cd000f [] ip-10-1-3-21:9997 . . .
Attachments
Issue Links
- relates to
-
ACCUMULO-2185 determine how rfile block location impacts performance
- Resolved