From sueloverso, split out from WT-3888.
Here's a CONFIG that hit a similar prev issue with a slightly different error message. This hit while I was testing something else and hit after 25 iterations. A second attempt on the same binary is still running after 1100 iterations (sadness). I was running this on my WT-3854 branch so it does not contain any of the lookaside changes from WT-3766. I did pull in the latest develop changes this morning however.
The error:
t: FAILED: nextprev/1193: incrementing || cmp > 0 || (cmp == 0 && tinfo->key->size > key.size): WT_CURSOR.prev returned {0000007725.13/opqrstuvwxyzabcd} then {0000009572.00/opqrstu}
The CONFIG:
############################################ # RUN PARAMETERS ############################################ abort=0 alter=0 auto_throttle=1 backups=1 bitcnt=4 bloom=1 bloom_bit_count=45 bloom_hash_count=15 bloom_oldest=0 cache=60 cache_minimum=0 checkpoints=wiredtiger checkpoint_log_size=46 checkpoint_wait=64 checksum=uncompressed chunk_size=2 compaction=0 compression=none data_extend=0 data_source=lsm delete_pct=20 dictionary=0 direct_io=0 encryption=none evict_max=1 file_type=row-store firstfit=0 huffman_key=0 huffman_value=0 independent_thread_rng=1 in_memory=0 insert_pct=23 internal_key_truncation=1 internal_page_max=12 isolation=random key_gap=14 key_max=57 key_min=17 leaf_page_max=15 leak_memory=0 logging=0 logging_archive=1 logging_compression=none logging_file_max=112418 logging_prealloc=1 long_running_txn=0 lsm_worker_threads=4 merge_max=13 mmap=1 modify_pct=16 ops=50000 prefix_compression=1 prefix_compression_min=3 quiet=1 read_pct=1 rebalance=1 repeat_data_pct=78 reverse=0 rows=10000 runs=1 salvage=1 split_pct=52 statistics=0 statistics_server=0 threads=4 timer=360 transaction_timestamps=1 transaction-frequency=45 value_max=1978 value_min=16 verify=1 wiredtiger_config= write_pct=40 ############################################