[PATCH blktests] block/005,008: do exit if fio did not finish within timeout

[PATCH blktests] block/005,008: do exit if fio did not finish within timeout

[Xiao Liang]

In some bad situation, fio needs taking over several hours to complete
random read operations with specipied size. The test may skip out in such
cases and does not block other cases run.

With this patch, the case will be ended within $TIMEOUT(if set) or 900s.
block/005 => nvme1n1 (switch schedulers while doing IO)      [failed]
    runtime      ...  1800.477s
    read iops    ...
    --- tests/block/005.out	2019-03-31 14:29:39.905449312 +0000
    +++ /home/ec2-user/blktests/results/nvme1n1/block/005.out.bad	2019-05-07 04:10:16.026681842 +0000
    @@ -1,2 +1,4 @@
     Running block/005
    +fio did not finish after 900 seconds which probably caused by
    +lower disk performance
     Test complete

Signed-off-by: Xiao Liang <xiliang@redhat.com>
---
 tests/block/005 | 10 ++++++++++
 tests/block/008 | 10 ++++++++++
 2 files changed, 20 insertions(+)

diff --git a/tests/block/005 b/tests/block/005
index 8ab6791..96b16a4 100755
--- a/tests/block/005
+++ b/tests/block/005
@@ -31,10 +31,20 @@ test_device() {
 	_run_fio_rand_io --filename="$TEST_DEV" --size="$size" &
 
 	# while job is running, switch between schedulers
+	# fio test may take too long time to complete read/write in special size on some bad 
+	# performance disks. Set a timeout here which does not block overall test.
+	start_time=$(date +%s)
+	timeout=${TIMEOUT:=900}
 	while kill -0 $! 2>/dev/null; do
 		idx=$((RANDOM % ${#scheds[@]}))
 		_test_dev_queue_set scheduler "${scheds[$idx]}"
 		sleep .2
+		end_time=$(date +%s)
+		if (( end_time - start_time > timeout )); then
+			echo "fio did not finish after $timeout seconds which probably caused by 
+lower disk performance"
+			break
+		fi
 	done
 
 	FIO_PERF_FIELDS=("read iops")
diff --git a/tests/block/008 b/tests/block/008
index 4a88056..c25b908 100755
--- a/tests/block/008
+++ b/tests/block/008
@@ -45,6 +45,10 @@ test_device() {
 	done
 
 	# while job is running, hotplug CPUs
+	# fio test may take too long time to complete read/write in special size on some bad 
+	# performance disks. Set a timeout here which does not block overall test.
+	start_time=$(date +%s)
+	timeout=${TIMEOUT:=900}
 	while sleep .2; kill -0 $! 2> /dev/null; do
 		if (( offlining && ${#offline_cpus[@]} == max_offline )); then
 			offlining=0
@@ -65,6 +69,12 @@ test_device() {
 			unset offline_cpus["$idx"]
 			offline_cpus=("${offline_cpus[@]}")
 		fi
+		end_time=$(date +%s)
+		if (( end_time - start_time > timeout )); then
+			echo "fio did not finish after $timeout seconds which probably caused by 
+lower disk performance"
+			break
+		fi
 	done
 
 	FIO_PERF_FIELDS=("read iops")
-- 
2.17.2

Re: [PATCH blktests] block/005,008: do exit if fio did not finish within timeout

[Chaitanya Kulkarni]

Hi Xiao,

I like to idea of not blocking the other tests. Couple of comments below.

On 05/06/2019 10:47 PM, Xiao Liang wrote:
> In some bad situation, fio needs taking over several hours to complete
> random read operations with specipied size. The test may skip out in such
> cases and does not block other cases run.
s/specipied/specified/
>
> With this patch, the case will be ended within $TIMEOUT(if set) or 900s.
> block/005 => nvme1n1 (switch schedulers while doing IO)      [failed]
>      runtime      ...  1800.477s
>      read iops    ...
>      --- tests/block/005.out	2019-03-31 14:29:39.905449312 +0000
>      +++ /home/ec2-user/blktests/results/nvme1n1/block/005.out.bad	2019-05-07 04:10:16.026681842 +0000
>      @@ -1,2 +1,4 @@
>       Running block/005
>      +fio did not finish after 900 seconds which probably caused by
>      +lower disk performance
>       Test complete
>
> Signed-off-by: Xiao Liang <xiliang@redhat.com>
> ---
>   tests/block/005 | 10 ++++++++++
>   tests/block/008 | 10 ++++++++++
>   2 files changed, 20 insertions(+)
>
> diff --git a/tests/block/005 b/tests/block/005
> index 8ab6791..96b16a4 100755
> --- a/tests/block/005
> +++ b/tests/block/005
> @@ -31,10 +31,20 @@ test_device() {
>   	_run_fio_rand_io --filename="$TEST_DEV" --size="$size" &
>
>   	# while job is running, switch between schedulers
> +	# fio test may take too long time to complete read/write in special size on some bad
> +	# performance disks. Set a timeout here which does not block overall test.
> +	start_time=$(date +%s)
> +	timeout=${TIMEOUT:=900}
>   	while kill -0 $! 2>/dev/null; do
>   		idx=$((RANDOM % ${#scheds[@]}))
>   		_test_dev_queue_set scheduler "${scheds[$idx]}"
>   		sleep .2
> +		end_time=$(date +%s)
> +		if (( end_time - start_time > timeout )); then
> +			echo "fio did not finish after $timeout seconds which probably caused by
> +lower disk performance"
Can we make this message more generic instead of calling out disk 
performance ?
> +			break
> +		fi
>   	done
>
>   	FIO_PERF_FIELDS=("read iops")
> diff --git a/tests/block/008 b/tests/block/008
> index 4a88056..c25b908 100755
> --- a/tests/block/008
> +++ b/tests/block/008
> @@ -45,6 +45,10 @@ test_device() {
>   	done
>
>   	# while job is running, hotplug CPUs
> +	# fio test may take too long time to complete read/write in special size on some bad
> +	# performance disks. Set a timeout here which does not block overall test.
> +	start_time=$(date +%s)
> +	timeout=${TIMEOUT:=900}
>   	while sleep .2; kill -0 $! 2> /dev/null; do
>   		if (( offlining && ${#offline_cpus[@]} == max_offline )); then
>   			offlining=0
> @@ -65,6 +69,12 @@ test_device() {
>   			unset offline_cpus["$idx"]
>   			offline_cpus=("${offline_cpus[@]}")
>   		fi
> +		end_time=$(date +%s)
> +		if (( end_time - start_time > timeout )); then
> +			echo "fio did not finish after $timeout seconds which probably caused by
> +lower disk performance"
Same here.
> +			break
> +		fi
>   	done
>
>   	FIO_PERF_FIELDS=("read iops")
>