Skip to content
Draft
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
127 changes: 125 additions & 2 deletions test/functional/tests/cache_ops/test_cleaning_policy_operation.py
Original file line number Diff line number Diff line change
Expand Up @@ -4,19 +4,24 @@
# SPDX-License-Identifier: BSD-3-Clause
#

import random
import time
import pytest

from datetime import timedelta

from api.cas import casadm
from api.cas.cache_config import (
CacheMode,
CleaningPolicy,
FlushParametersAcp,
FlushParametersAlru,
Time,
SeqCutOffPolicy,
)
from storage_devices.disk import DiskType, DiskTypeSet, DiskTypeLowerThan
from core.test_run import TestRun
from test_tools.iostat import IOstatBasic
from type_def.size import Size, Unit
from test_tools.udev import Udev
from test_tools.fio.fio import Fio
Expand Down Expand Up @@ -106,7 +111,7 @@ def test_cleaning_policies_in_write_back(cleaning_policy: CleaningPolicy):
@pytest.mark.parametrize("cleaning_policy", CleaningPolicy)
@pytest.mark.require_disk("cache", DiskTypeSet([DiskType.optane, DiskType.nand]))
@pytest.mark.require_disk("core", DiskTypeLowerThan("cache"))
def test_cleaning_policies_in_write_through(cleaning_policy):
def test_cleaning_policies_in_write_through(cleaning_policy: CleaningPolicy):
"""
title: Test for cleaning policy operation in Write-Through cache mode.
description: |
Expand Down Expand Up @@ -257,7 +262,7 @@ def check_cleaning_policy_operation(
case CleaningPolicy.nop:
if (
core_writes_after_wait_for_cleaning != Size.zero()
or core_writes_before_wait_for_cleaning != Size.zero()
or core_writes_before_wait_for_cleaning != Size.zero()
):
TestRun.LOGGER.error(
"NOP cleaning policy is not working properly! "
Expand All @@ -275,3 +280,121 @@ def check_cleaning_policy_operation(
"ACP cleaning policy is not working properly! "
"Core writes should increase in time while cleaning dirty data"
)


@pytest.mark.require_disk("cache", DiskTypeSet([DiskType.optane, DiskType.nand]))
@pytest.mark.require_disk("core", DiskTypeLowerThan("cache"))
def test_alru_change_trigger_dirty_ratio_during_io():
"""
title: Test ALRU with trigger dirty ratio param during I/O
description: |
Verify that ALRU is able to start cleaning to expected dirty ratio under constant I/O.
pass_criteria:
- Dirty cache lines are being cleaned after trigger dirty ratio change during I/O.
- Dirty cache lines are cleaned to expected dirty ratio after stopping I/O
"""
iterations = 10

with TestRun.step("Prepare cache and core devices"):
cache_dev = TestRun.disks["cache"]
core_dev = TestRun.disks["core"]

cache_size = Size(1, Unit.GiB)
cache_dev.create_partitions([cache_size])
core_dev.create_partitions([Size(2, Unit.GiB)])

with TestRun.step(f"Disable udev"):
Udev.disable()

with TestRun.step("Start cache"):
cache = casadm.start_cache(cache_dev.partitions[0], force=True, cache_mode=CacheMode.WB)

with TestRun.step("Add core"):
core = cache.add_core(core_dev.partitions[0])

with TestRun.step("Set params for ALRU cleaning policy"):
params = FlushParametersAlru.default_alru_params()
params.wake_up_time = Time(seconds=1)
params.staleness_time = Time(seconds=3600)
cache.set_params_alru(alru_params=params)

with TestRun.step("Disable sequential cut-off and set cleaning to ALRU"):
cache.set_seq_cutoff_policy(SeqCutOffPolicy.never)
cache.set_cleaning_policy(CleaningPolicy.alru)

for _ in TestRun.iteration(
range(0, iterations),
"Start changing trigger dirty ratio param during I/O and check if flush occurred",
):

params = FlushParametersAlru.default_alru_params()

fio = (
Fio()
.create_command()
.io_engine(IoEngine.libaio)
.block_size(Size(4, Unit.KiB))
.target(core)
.direct()
.time_based()
.run_time(timedelta(minutes=10))
.read_write(ReadWrite.randwrite)
)

dirty_ratio_test_threshold = random.randint(0, 100)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Picking a random threshold doesn't guarante that the corner cases with threshold=0 and threshold=100 are tested. Instead of 10 random iterations I suggest 3 predefined cases with both corner cases and and one value from the middle of the range.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

There could also be another testcase to check the behavior when decreasing and increasing the threshold when the flush is in progress


with TestRun.step("Disable cache cleaning"):
cache.set_seq_cutoff_policy(SeqCutOffPolicy.never)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cutoff has been set to never in :322


with TestRun.step("Start running I/O to exported object in background"):
fio_pid = fio.run_in_background()

with TestRun.step("Wait until dirty data on cache exceed dirty ratio threshold"):

while TestRun.executor.check_if_process_exists(fio_pid):
dirty_blocks_percentage_on_cache = cache.get_statistics(
percentage_val=True
).usage_stats.dirty

if dirty_blocks_percentage_on_cache >= dirty_ratio_test_threshold:
break
time.sleep(5)

with TestRun.step("Set cache cleaning policy to alru"):
params.trigger_dirty_ratio = dirty_ratio_test_threshold
cache.set_params_alru(alru_params=params)

with TestRun.step("Check if cleaning started after changing cleaning policy"):
time.sleep(5)

iostat_core = IOstatBasic.get_iostat_list([core_dev.get_device_id()])
core_writes = iostat_core[0].total_writes

if core_writes == Size.zero():
TestRun.fail(f"Cleaning on cache haven`t started")

with TestRun.step("Stop I/O to exported object"):
TestRun.executor.kill_process(fio_pid)

with TestRun.step("Check if cache has been cleaned to set threshold after stopping I/O"):
max_cleaning_time = timedelta(seconds=60)
t_end = time.time() + max_cleaning_time.seconds

while time.time() < t_end:
dirty_blocks_on_cache_before = cache.get_statistics(
percentage_val=True).usage_stats.dirty
time.sleep(20)
dirty_blocks_on_cache_after = cache.get_statistics(
percentage_val=True).usage_stats.dirty
if dirty_blocks_on_cache_before == dirty_blocks_on_cache_after:
break

dirty_blocks_cache_percentage = cache.get_statistics(
percentage_val=True
).usage_stats.dirty

if not dirty_ratio_test_threshold == dirty_blocks_cache_percentage:
TestRun.fail("Dirty block percentage outside of defined range")

with TestRun.step("Purge cache"):
casadm.purge_cache(cache_id=cache.cache_id)