Current issues:
- oplog_rollover_test: server crashes due to insert of document exceeding capped max size; server code says the client should retry the insert, but the client is the internal replication sync workers
- write_concern (import and restore): appears that the _mongos variant of these tests are overwriting the export/dump output on the FS mid-execution
- dump_server_ko_test: server is brought down before actual dump begins, causing mongodump to output a different error message
- mongofiles_search: test fails due to missing file shortly before it appears in the output
- parse_grace: db.foo.drop() followed by assert.eq(db.foo.count()) fails on PPC and must be retried (potential server bug)
Relliant on sleeps:
- bson/all_types tests
- force_table_scan_tests
- oplog_flag_tests
- import/options
- type_case
- stat_rowcount
Vulnerable to SERVER-25076:
- bad_files
- bsondump_options
- dump/version_test
- export/stdout
- mongofiles_get
- mongofiles_list
- All mongooplog tests
- stat_header
- stat_mixed_storage_engine
- stat_rowcount
- mongotop_json
Miscellaneous:
- read_preference_and_tags: re-prints shell output unnecessarily
- bson/output_file: uses sleeps after completed FS writes