ivampiresp/Spdk
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
8d22cb3e4d
Spdk/test/ocf/management/persistent-metadata.sh

81 lines
2.0 KiB
Bash
Raw Normal View History

test/ocf: add persistent metadata tests We check if get_ocf_bdevs after shutdown is the same as before, which shows that state for all vbdevs wes correctly restored from metadata. Change-Id: I9089642aeb840fc940b07908aa39c8193887432f Signed-off-by: Vitaliy Mysak <vitaliy.mysak@intel.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455424 Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-05-21 21:21:03 +00:00
#!/usr/bin/env bash
test: Shellcheck - correct rule: Expanding an array Correct shellcheck rule SC2128: Expanding an array without an index only gives the first element. Signed-off-by: Maciej Wawryk <maciejx.wawryk@intel.com> Change-Id: I0e7c335af678114dc78dfb12a02369a69158e435 Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/474989 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Seth Howell <seth.howell@intel.com> Reviewed-by: Karol Latecki <karol.latecki@intel.com>
2019-11-18 14:55:26 +00:00
curdir=$(dirname $(readlink -f "${BASH_SOURCE[0]}"))
test/ocf: add persistent metadata tests We check if get_ocf_bdevs after shutdown is the same as before, which shows that state for all vbdevs wes correctly restored from metadata. Change-Id: I9089642aeb840fc940b07908aa39c8193887432f Signed-off-by: Vitaliy Mysak <vitaliy.mysak@intel.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455424 Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-05-21 21:21:03 +00:00
rootdir=$(readlink -f $curdir/../../..)
test/ocf: change backend device in persistent-metadata test This patch is related to issue #815. To avoid hung and re-enable testing this part of ocf, this patch changes backed device from aio to nvme. Hung in this test was caused by implementation of AIO bdev and bdev layer. All AIO bdevs are using the same shared context with limited queue depth, so in some cases AIO can return NOMEM status. It's ok when we have more than one IO from bdev layer perspective, if not we will stuck because nothing will triger retry io procedure. This patch only enables testing and are not fixing root cause of hung. To prevent before unexpected behavior of this and next one test we need to clear ocf metadata on nvme device. This patch also reenables this test in CI. Signed-off-by: Marcin Dziegielewski <marcin.dziegielewski@intel.com> Change-Id: Ibab4aefb9aaf33d725db20345bd5c09c1e5eebdd Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463605 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Vitaliy Mysak <vitaliy.mysak@intel.com>
2019-08-22 15:55:46 +00:00
source $rootdir/scripts/common.sh
test/ocf: add persistent metadata tests We check if get_ocf_bdevs after shutdown is the same as before, which shows that state for all vbdevs wes correctly restored from metadata. Change-Id: I9089642aeb840fc940b07908aa39c8193887432f Signed-off-by: Vitaliy Mysak <vitaliy.mysak@intel.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455424 Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-05-21 21:21:03 +00:00
source $rootdir/test/common/autotest_common.sh
test/ocf: change backend device in persistent-metadata test This patch is related to issue #815. To avoid hung and re-enable testing this part of ocf, this patch changes backed device from aio to nvme. Hung in this test was caused by implementation of AIO bdev and bdev layer. All AIO bdevs are using the same shared context with limited queue depth, so in some cases AIO can return NOMEM status. It's ok when we have more than one IO from bdev layer perspective, if not we will stuck because nothing will triger retry io procedure. This patch only enables testing and are not fixing root cause of hung. To prevent before unexpected behavior of this and next one test we need to clear ocf metadata on nvme device. This patch also reenables this test in CI. Signed-off-by: Marcin Dziegielewski <marcin.dziegielewski@intel.com> Change-Id: Ibab4aefb9aaf33d725db20345bd5c09c1e5eebdd Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463605 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Vitaliy Mysak <vitaliy.mysak@intel.com>
2019-08-22 15:55:46 +00:00
function clear_nvme()
{
# Clear metadata on NVMe device
$rootdir/scripts/setup.sh reset
test: fix errors in test persistent-metadata.sh Test failed both on x86_64 and aarch64 on some cases: 1.nvme_cfg=$($rootdir/scripts/gen_nvme.sh) If the driver haven't been changed to vfio-pci(haven't run scripts/setup.sh), we got no result from gen_nvme.sh. 2."mountpoints=$(lsblk /dev/$blkname --output MOUNTPOINT -n | wc -w)" The global variable "blkname" (defined in get_nvme_name_from_bdf) is always null here. For the scope of shell global variable is current process. However, "name=$(get_nvme_name_from_bdf $1)" will create a new process. We use variable "name" here. 3."name=$(get_nvme_name_from_bdf $1)" If the "$rootdir/scripts/setup.sh reset" above is not ready, $name will be null. For some non-x86 platforms, the waiting time maybe not enough. So I adjust it to 5 seconds according to repeated tests. Signed-off-by: Richael Zhuang <richael.zhuang@arm.com> Change-Id: Ia91460a3116bc0cc7c6c5f5ee471957caf9d0e06 Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/474184 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-by: Vitaliy Mysak <vitaliy.mysak@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2019-11-14 08:48:11 +00:00
sleep 5
test/ocf: change backend device in persistent-metadata test This patch is related to issue #815. To avoid hung and re-enable testing this part of ocf, this patch changes backed device from aio to nvme. Hung in this test was caused by implementation of AIO bdev and bdev layer. All AIO bdevs are using the same shared context with limited queue depth, so in some cases AIO can return NOMEM status. It's ok when we have more than one IO from bdev layer perspective, if not we will stuck because nothing will triger retry io procedure. This patch only enables testing and are not fixing root cause of hung. To prevent before unexpected behavior of this and next one test we need to clear ocf metadata on nvme device. This patch also reenables this test in CI. Signed-off-by: Marcin Dziegielewski <marcin.dziegielewski@intel.com> Change-Id: Ibab4aefb9aaf33d725db20345bd5c09c1e5eebdd Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463605 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Vitaliy Mysak <vitaliy.mysak@intel.com>
2019-08-22 15:55:46 +00:00
name=$(get_nvme_name_from_bdf $1)
test: fix errors in test persistent-metadata.sh Test failed both on x86_64 and aarch64 on some cases: 1.nvme_cfg=$($rootdir/scripts/gen_nvme.sh) If the driver haven't been changed to vfio-pci(haven't run scripts/setup.sh), we got no result from gen_nvme.sh. 2."mountpoints=$(lsblk /dev/$blkname --output MOUNTPOINT -n | wc -w)" The global variable "blkname" (defined in get_nvme_name_from_bdf) is always null here. For the scope of shell global variable is current process. However, "name=$(get_nvme_name_from_bdf $1)" will create a new process. We use variable "name" here. 3."name=$(get_nvme_name_from_bdf $1)" If the "$rootdir/scripts/setup.sh reset" above is not ready, $name will be null. For some non-x86 platforms, the waiting time maybe not enough. So I adjust it to 5 seconds according to repeated tests. Signed-off-by: Richael Zhuang <richael.zhuang@arm.com> Change-Id: Ia91460a3116bc0cc7c6c5f5ee471957caf9d0e06 Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/474184 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-by: Vitaliy Mysak <vitaliy.mysak@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2019-11-14 08:48:11 +00:00
mountpoints=$(lsblk /dev/$name --output MOUNTPOINT -n | wc -w)
test/ocf: change backend device in persistent-metadata test This patch is related to issue #815. To avoid hung and re-enable testing this part of ocf, this patch changes backed device from aio to nvme. Hung in this test was caused by implementation of AIO bdev and bdev layer. All AIO bdevs are using the same shared context with limited queue depth, so in some cases AIO can return NOMEM status. It's ok when we have more than one IO from bdev layer perspective, if not we will stuck because nothing will triger retry io procedure. This patch only enables testing and are not fixing root cause of hung. To prevent before unexpected behavior of this and next one test we need to clear ocf metadata on nvme device. This patch also reenables this test in CI. Signed-off-by: Marcin Dziegielewski <marcin.dziegielewski@intel.com> Change-Id: Ibab4aefb9aaf33d725db20345bd5c09c1e5eebdd Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463605 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Vitaliy Mysak <vitaliy.mysak@intel.com>
2019-08-22 15:55:46 +00:00
if [ "$mountpoints" != "0" ]; then
$rootdir/scripts/setup.sh
exit 1
fi
dd if=/dev/zero of=/dev/$name bs=1M count=1000 oflag=direct
$rootdir/scripts/setup.sh
}
test/ocf: add persistent metadata tests We check if get_ocf_bdevs after shutdown is the same as before, which shows that state for all vbdevs wes correctly restored from metadata. Change-Id: I9089642aeb840fc940b07908aa39c8193887432f Signed-off-by: Vitaliy Mysak <vitaliy.mysak@intel.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455424 Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-05-21 21:21:03 +00:00
rpc_py=$rootdir/scripts/rpc.py
test: fix errors in test persistent-metadata.sh Test failed both on x86_64 and aarch64 on some cases: 1.nvme_cfg=$($rootdir/scripts/gen_nvme.sh) If the driver haven't been changed to vfio-pci(haven't run scripts/setup.sh), we got no result from gen_nvme.sh. 2."mountpoints=$(lsblk /dev/$blkname --output MOUNTPOINT -n | wc -w)" The global variable "blkname" (defined in get_nvme_name_from_bdf) is always null here. For the scope of shell global variable is current process. However, "name=$(get_nvme_name_from_bdf $1)" will create a new process. We use variable "name" here. 3."name=$(get_nvme_name_from_bdf $1)" If the "$rootdir/scripts/setup.sh reset" above is not ready, $name will be null. For some non-x86 platforms, the waiting time maybe not enough. So I adjust it to 5 seconds according to repeated tests. Signed-off-by: Richael Zhuang <richael.zhuang@arm.com> Change-Id: Ia91460a3116bc0cc7c6c5f5ee471957caf9d0e06 Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/474184 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-by: Vitaliy Mysak <vitaliy.mysak@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2019-11-14 08:48:11 +00:00
$rootdir/scripts/setup.sh
test/ocf: change backend device in persistent-metadata test This patch is related to issue #815. To avoid hung and re-enable testing this part of ocf, this patch changes backed device from aio to nvme. Hung in this test was caused by implementation of AIO bdev and bdev layer. All AIO bdevs are using the same shared context with limited queue depth, so in some cases AIO can return NOMEM status. It's ok when we have more than one IO from bdev layer perspective, if not we will stuck because nothing will triger retry io procedure. This patch only enables testing and are not fixing root cause of hung. To prevent before unexpected behavior of this and next one test we need to clear ocf metadata on nvme device. This patch also reenables this test in CI. Signed-off-by: Marcin Dziegielewski <marcin.dziegielewski@intel.com> Change-Id: Ibab4aefb9aaf33d725db20345bd5c09c1e5eebdd Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463605 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Vitaliy Mysak <vitaliy.mysak@intel.com>
2019-08-22 15:55:46 +00:00
nvme_cfg=$($rootdir/scripts/gen_nvme.sh)
config="
$nvme_cfg
[Split]
Split Nvme0n1 7 128
"
echo "$config" > $curdir/config
# Clear only nvme device which we will use in test
bdf=$($rootdir/scripts/gen_nvme.sh --json | jq '.config[0].params.traddr' | sed s/\"//g)
clear_nvme $bdf
test/ocf: add persistent metadata tests We check if get_ocf_bdevs after shutdown is the same as before, which shows that state for all vbdevs wes correctly restored from metadata. Change-Id: I9089642aeb840fc940b07908aa39c8193887432f Signed-off-by: Vitaliy Mysak <vitaliy.mysak@intel.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455424 Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-05-21 21:21:03 +00:00
$rootdir/app/iscsi_tgt/iscsi_tgt -c $curdir/config &
spdk_pid=$!
waitforlisten $spdk_pid
# Create ocf on persistent storage
rpc: Rename ocf create and delete related rpcs Rename construct_ocf_bdev to bdev_ocf_create. Rename delete_ocf_bdev to bdev_ocf_delete. Change-Id: Iec456c2004cc7e6c61d29439e1b7a9b5dfe7dfaa Signed-off-by: Pawel Kaminski <pawelx.kaminski@intel.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/466518 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2019-08-28 08:30:48 +00:00
$rpc_py bdev_ocf_create ocfWT wt Nvme0n1p0 Nvme0n1p1
$rpc_py bdev_ocf_create ocfPT pt Nvme0n1p2 Nvme0n1p3
$rpc_py bdev_ocf_create ocfWB0 wb Nvme0n1p4 Nvme0n1p5
$rpc_py bdev_ocf_create ocfWB1 wb Nvme0n1p4 Nvme0n1p6
test/ocf: add persistent metadata tests We check if get_ocf_bdevs after shutdown is the same as before, which shows that state for all vbdevs wes correctly restored from metadata. Change-Id: I9089642aeb840fc940b07908aa39c8193887432f Signed-off-by: Vitaliy Mysak <vitaliy.mysak@intel.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455424 Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-05-21 21:21:03 +00:00
# Sorting bdevs because we dont guarantee that they are going to be
# in the same order after shutdown
rpc: Rename ocf get_bdevs and get_stats related rpcs. Rename get_ocf_stats to bdev_ocf_get_stats. Rename get_ocf_bdevs to bdev_ocf_get_bdevs. Change-Id: I627d9ff7bcc3f30ec19d245e5caa06238cda3a9e Signed-off-by: Pawel Kaminski <pawelx.kaminski@intel.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/466519 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2019-08-28 08:43:22 +00:00
($rpc_py bdev_ocf_get_bdevs | jq '(.. | arrays) |= sort') > ./ocf_bdevs
test/ocf: add persistent metadata tests We check if get_ocf_bdevs after shutdown is the same as before, which shows that state for all vbdevs wes correctly restored from metadata. Change-Id: I9089642aeb840fc940b07908aa39c8193887432f Signed-off-by: Vitaliy Mysak <vitaliy.mysak@intel.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455424 Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-05-21 21:21:03 +00:00
trap - SIGINT SIGTERM EXIT
killprocess $spdk_pid
# Check for ocf persistency after restart
$rootdir/app/iscsi_tgt/iscsi_tgt -c $curdir/config &
spdk_pid=$!
test/ocf: change backend device in persistent-metadata test This patch is related to issue #815. To avoid hung and re-enable testing this part of ocf, this patch changes backed device from aio to nvme. Hung in this test was caused by implementation of AIO bdev and bdev layer. All AIO bdevs are using the same shared context with limited queue depth, so in some cases AIO can return NOMEM status. It's ok when we have more than one IO from bdev layer perspective, if not we will stuck because nothing will triger retry io procedure. This patch only enables testing and are not fixing root cause of hung. To prevent before unexpected behavior of this and next one test we need to clear ocf metadata on nvme device. This patch also reenables this test in CI. Signed-off-by: Marcin Dziegielewski <marcin.dziegielewski@intel.com> Change-Id: Ibab4aefb9aaf33d725db20345bd5c09c1e5eebdd Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463605 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Vitaliy Mysak <vitaliy.mysak@intel.com>
2019-08-22 15:55:46 +00:00
trap 'killprocess $spdk_pid; rm -f $curdir/config ocf_bdevs ocf_bdevs_verify; exit 1' SIGINT SIGTERM EXIT
test/ocf: add persistent metadata tests We check if get_ocf_bdevs after shutdown is the same as before, which shows that state for all vbdevs wes correctly restored from metadata. Change-Id: I9089642aeb840fc940b07908aa39c8193887432f Signed-off-by: Vitaliy Mysak <vitaliy.mysak@intel.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455424 Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-05-21 21:21:03 +00:00
waitforlisten $spdk_pid
# OCF should be loaded now as well
rpc: Rename ocf get_bdevs and get_stats related rpcs. Rename get_ocf_stats to bdev_ocf_get_stats. Rename get_ocf_bdevs to bdev_ocf_get_bdevs. Change-Id: I627d9ff7bcc3f30ec19d245e5caa06238cda3a9e Signed-off-by: Pawel Kaminski <pawelx.kaminski@intel.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/466519 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2019-08-28 08:43:22 +00:00
($rpc_py bdev_ocf_get_bdevs | jq '(.. | arrays) |= sort') > ./ocf_bdevs_verify
test/ocf: add persistent metadata tests We check if get_ocf_bdevs after shutdown is the same as before, which shows that state for all vbdevs wes correctly restored from metadata. Change-Id: I9089642aeb840fc940b07908aa39c8193887432f Signed-off-by: Vitaliy Mysak <vitaliy.mysak@intel.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455424 Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-05-21 21:21:03 +00:00
diff ocf_bdevs ocf_bdevs_verify
trap - SIGINT SIGTERM EXIT
killprocess $spdk_pid
test/ocf: change backend device in persistent-metadata test This patch is related to issue #815. To avoid hung and re-enable testing this part of ocf, this patch changes backed device from aio to nvme. Hung in this test was caused by implementation of AIO bdev and bdev layer. All AIO bdevs are using the same shared context with limited queue depth, so in some cases AIO can return NOMEM status. It's ok when we have more than one IO from bdev layer perspective, if not we will stuck because nothing will triger retry io procedure. This patch only enables testing and are not fixing root cause of hung. To prevent before unexpected behavior of this and next one test we need to clear ocf metadata on nvme device. This patch also reenables this test in CI. Signed-off-by: Marcin Dziegielewski <marcin.dziegielewski@intel.com> Change-Id: Ibab4aefb9aaf33d725db20345bd5c09c1e5eebdd Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463605 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Vitaliy Mysak <vitaliy.mysak@intel.com>
2019-08-22 15:55:46 +00:00
rm -f $curdir/config ocf_bdevs ocf_bdevs_verify
clear_nvme $bdf
Reference in New Issue Copy Permalink