See: http://build.wiredtiger.com:8080/job/wiredtiger-test-unit/5201/console
Outputs:
==========
cache dump
file:inmem01.wt(<live>):
internal: 11 pages, 0MB, 0/11 clean/dirty pages, 0/0 clean/dirty MB, 0MB max page, 0MB max dirty page
leaf: 596 pages, 4MB, 409/187 clean/dirty pages, 0/4 clean/dirty MB, 0MB max page, 0MB max dirty page
file:WiredTiger.wt(<live>):
internal: 1 pages, 0MB, 1/0 clean/dirty pages, 0/0 clean/dirty MB, 0MB max page, 0MB max dirty page
leaf: 1 pages, 0MB, 0/1 clean/dirty pages, 0/0 clean/dirty MB, 0MB max page, 0MB max dirty page
cache dump: total found = 5MB vs tracked inuse 5MB
total dirty bytes = 4MB
==========
test_inmem01.test_inmem01.test_insert_over_delete_replace(col) (subunit.RemotedTestCase)
test_inmem01.test_inmem01.test_insert_over_delete_replace(col) ... ERROR
======================================================================
ERROR: test_inmem01.test_inmem01.test_insert_over_delete_replace(col) (subunit.RemotedTestCase)
test_inmem01.test_inmem01.test_insert_over_delete_replace(col)
----------------------------------------------------------------------
_StringException: lost connection during test 'test_inmem01.test_inmem01.test_insert_over_delete_replace(col)'
----------------------------------------------------------------------
Ran 7224 tests in 393.984s
I have a reproducer which can cause the abort due to cache full on inmem, but I'm uncertain if it is the same reproduer - as I have been abortin with only 1MB dirty, rather than 4.
My reproducer involves setting memory_page_max=4k down from memory_page_max=32k
- related to
-
WT-2233 Investigate changing when the eviction server switches to aggressive mode.
- Closed