Commit df38292e authored Aug 03, 2022 by Changpeng Liu Committed by Tomasz Zawadzki Aug 05, 2022

test/nvmf/timeout.sh: use `randread` as the test case



The reconnect timeout test case expects a socket READ error
to trigger the reconnect poller in bdev_nvme module, but
we use the `verify` IO workload to send IO requests which
may cause the client doesn't receive any error, because
it will send WRITE commands first.

Fix issue #2595.

Change-Id: Ibc6b534d157552cdb2b55a03c869e254ec4c462e
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/13840


Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Dong Yi <dongx.yi@intel.com>
Reviewed-by: Michal Berger <michal.berger@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>

parent e8580900

test/nvmf/host/timeout.sh

+1 −1

Original line number	Diff line number	Diff line
		@@ -108,7 +108,7 @@ rm -f $testdir/try.txt

		# Case 3 test reconnect_delay_sec
		# Time to delay a reconnect trial
		$rootdir/test/bdev/bdevperf/bdevperf -m 0x4 -z -r $bdevperf_rpc_sock -q 128 -o 4096 -w verify -t 20 -f &> $testdir/try.txt &
		$rootdir/test/bdev/bdevperf/bdevperf -m 0x4 -z -r $bdevperf_rpc_sock -q 128 -o 4096 -w randread -t 20 -f &> $testdir/try.txt &
		bdevperf_pid=$!

		trap 'process_shm --id $NVMF_APP_SHM_ID; rm -f $testdir/try.txt; killprocess $bdevperf_pid; nvmftestfini; exit 1' SIGINT SIGTERM EXIT