Commit 317b9a53 authored Sep 27, 2018 by Seth Howell Committed by Jim Harris Oct 02, 2018

test/nvmf: watiforblk in shutdown test

In newer versions of the kernel, there is a race condition that causes
the kernel block layer to mess up if an nvme-oF subsystem is
disconnected before it is fully removed. This can cause the shutdown
test to hang during a sync call aftr nvmf tgt shutdown. see:
https://ci.spdk.io/spdk/builds/review/d1ecb02290a63c7e17cbe9becd7d48e3d0f245d6.1537982067/fedora-06/build.log
for details.

This problem has been observed and fixed in the RPC tests also.
https://review.gerrithub.io/c/spdk/spdk/+/426772



Change-Id: I9ec8517e067448be323e69979adf5d3915195c80
Signed-off-by: Seth Howell <seth.howell@intel.com>
Reviewed-on: https://review.gerrithub.io/427152


Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>

parent 5efc71b5

test/nvmf/shutdown/shutdown.sh

+2 −0

Original line number	Diff line number	Diff line
		@@ -63,6 +63,8 @@ for i in `seq 1 $num_subsystems`; do
		nvme connect -t rdma -n "nqn.2016-06.io.spdk:cnode${i}" -a "$NVMF_FIRST_TARGET_IP" -s "$NVMF_PORT"
		done

		waitforblk "nvme0n1"

		# Kill nvmf tgt without removing any subsystem to check whether it can shutdown correctly
		rm -f ./local-job0-0-verify.state