Details
-
Task
-
Resolution: Fixed
-
Major
-
None
-
4.5.0
-
None
-
None
Description
For initial index build, seeing there is difference in reported memory usage in forestdb. Here are the 2 runs:
Watson (4.5-2016)
http://perf.jenkins.couchbase.com/job/secondary-bharath/676/
Sherlock (4.1.1-5910)
http://perf.jenkins.couchbase.com/job/secondary-bharath/678/
In watson, here are the stats reported after initial build is done. The stat memory_used_storage (foretsdb buffer cache) is 15G. The stat memory_used is combined memory usage by indexer and forestdb.
2016-04-07T01:58:09.646-07:00
"bucket-1:myindex:data_size":5353304353
"bucket-1:myindex:disk_size":15718506496
"bucket-1:myindex:items_count":20000000
"memory_quota":31457280000
"memory_used":17398835200
"memory_used_storage":15718924288
In 4.1.1, here are the stats reported after initial build is done. There is no independent stat for forestdb, but stat memory_used for forestdb + indexer is 9G which is much less than memory_used_storage in watson (15G).
2016-04-07T03:56:12.945-07:00
"bucket-1:myindex:data_size":6457262379
"bucket-1:myindex:disk_size":43244154880
"bucket-1:myindex:items_count":20000000
"memory_quota":31457280000
"memory_used":9009459200