TPC-C test by the tool loadgen fails when the parameter -concurrency=128

Hi:

I deploy my cockroach cluster on three machines. And use the tool loadgen load data into my cluster. Then I do the TPC-C test on my cluster.My cockroach version is 1.1.2.

load data:

[root@ceph4 tpcc]# ./tpcc -concurrency=128 -warehouses=100 -duration=5m -load postgresql://172.16.50.103:26257/tpcc?sslmode=disable

tpc-c test:
-concurrency=32

[root@ceph4 tpcc]# ./tpcc -concurrency=32 -warehouses=100 -duration=5m postgresql://172.16.50.103:26257/tpcc?sslmode=disable

_elapsed___newOrders___newOrder(cum)__avg(ms)__p50(ms)__p95(ms)__p99(ms)_pMax(ms)
300.0s 22031 73.4 263.1 260.0 419.4 637.5 1409.3

-concurrency=64

[root@ceph4 tpcc]# ./tpcc -concurrency=64 -warehouses=100 -duration=5m postgresql://172.16.50.103:26257/tpcc?sslmode=disable

_elapsed___newOrders___newOrder(cum)__avg(ms)__p50(ms)__p95(ms)__p99(ms)_pMax(ms)
300.0s 26609 88.7 434.8 419.4 738.2 1140.9 5637.1

But when the -concurrency was 128,it doesn’t work.

[root@ceph4 tpcc]# ./tpcc -concurrency=128 -warehouses=100 -duration=5m postgresql://172.16.50.103:26257/tpcc?sslmode=disable

2017/11/10 04:47:55 error in payment: select by customer idfail: pq: restart transaction: HandledRetryableTxnError: TransactionAbortedError: txn aborted “sql txn” id=e1191170 key=/Table/51/1/1/0 rw=true pri=0.04576235 iso=SERIALIZABLE stat=ABORTED epo=1 ts=1510307274.138623162,19 orig=1510307274.138623162,19 max=1510307260.919322496,34 wto=false rop=false seq=13

why it happen? And I What shoud I do to finish the tpc-c test with the -concurrency=128? Thanks!

Hi @going. Looks like you found a bug in our tpcc load generator! I filed