Files
scylladb/test/cql-pytest/test_tools.py
Botond Dénes 54c0a387a2 Revert "Merge 'tool/scylla-sstable: more flexibility in obtaining the schema' from Botond Dénes"
This reverts commit 32fff17e19, reversing
changes made to 164afe14ad.

This series proved to be problematic, the new test introduced by it
failing quite often. Revert it until the problems are tracked down and
fixed.
2023-04-03 13:54:00 +03:00

541 lines
22 KiB
Python

# Copyright 2022-present ScyllaDB
#
# SPDX-License-Identifier: AGPL-3.0-or-later
#############################################################################
# Tests for the tools hosted by scylla
#############################################################################
import contextlib
import glob
import json
import nodetool
import os
import pytest
import subprocess
import tempfile
import random
import util
# To run the Scylla tools, we need to run Scylla executable itself, so we
# need to find the path of the executable that was used to run Scylla for
# this test. We do this by trying to find a local process which is listening
# to the address and port to which our our CQL connection is connected.
# If such a process exists, we verify that it is Scylla, and return the
# executable's path. If we can't find the Scylla executable we use
# pytest.skip() to skip tests relying on this executable.
@pytest.fixture(scope="module")
def scylla_path(cql):
pid = util.local_process_id(cql)
if not pid:
pytest.skip("Can't find local Scylla process")
# Now that we know the process id, use /proc to find the executable.
try:
path = os.readlink(f'/proc/{pid}/exe')
except:
pytest.skip("Can't find local Scylla executable")
# Confirm that this executable is a real tool-providing Scylla by trying
# to run it with the "--list-tools" option
try:
subprocess.check_output([path, '--list-tools'])
except:
pytest.skip("Local server isn't Scylla")
return path
# A fixture for finding Scylla's data directory. We get it using the CQL
# interface to Scylla's configuration. Note that if the server is remote,
# the directory retrieved this way may be irrelevant, whether or not it
# exists on the local machine... However, if the same test that uses this
# fixture also uses the scylla_path fixture, the test will anyway be skipped
# if the running Scylla is not on the local machine local.
@pytest.fixture(scope="module")
def scylla_data_dir(cql):
try:
dir = json.loads(cql.execute("SELECT value FROM system.config WHERE name = 'data_file_directories'").one().value)[0]
return dir
except:
pytest.skip("Can't find Scylla sstable directory")
def simple_no_clustering_table(cql, keyspace):
table = util.unique_name()
schema = f"CREATE TABLE {keyspace}.{table} (pk int PRIMARY KEY, v int) WITH compaction = {{'class': 'NullCompactionStrategy'}}"
cql.execute(schema)
for pk in range(0, 10):
x = random.randrange(0, 4)
if x == 0:
# partition tombstone
cql.execute(f"DELETE FROM {keyspace}.{table} WHERE pk = {pk}")
else:
# live row
cql.execute(f"INSERT INTO {keyspace}.{table} (pk, v) VALUES ({pk}, 0)")
if pk == 5:
nodetool.flush(cql, f"{keyspace}.{table}")
nodetool.flush(cql, f"{keyspace}.{table}")
return table, schema
def simple_clustering_table(cql, keyspace):
table = util.unique_name()
schema = f"CREATE TABLE {keyspace}.{table} (pk1 int, pk2 int, ck1 int, ck2 int, v int, s int STATIC, PRIMARY KEY ((pk1, pk2), ck1, ck2)) WITH compaction = {{'class': 'NullCompactionStrategy'}}"
cql.execute(schema)
for pk in range(0, 10):
for ck in range(0, 10):
x = random.randrange(0, 8)
if x == 0:
# ttl
cql.execute(f"INSERT INTO {keyspace}.{table} (pk1, pk2, ck1, ck2, v) VALUES ({pk}, {pk}, {ck}, {ck}, 0) USING TTL 6000")
elif x == 1:
# row tombstone
cql.execute(f"DELETE FROM {keyspace}.{table} WHERE pk1 = {pk} AND pk2 = {pk} AND ck1 = {ck} AND ck2 = {ck}")
elif x == 2:
# cell tombstone
cql.execute(f"DELETE v FROM {keyspace}.{table} WHERE pk1 = {pk} AND pk2 = {pk} AND ck1 = {ck} AND ck2 = {ck}")
elif x == 3:
# range tombstone
l = ck * 10
u = ck * 11
cql.execute(f"DELETE FROM {keyspace}.{table} WHERE pk1 = {pk} AND pk2 = {pk} AND ck1 > {l} AND ck1 < {u}")
else:
# live row
cql.execute(f"INSERT INTO {keyspace}.{table} (pk1, pk2, ck1, ck2, v) VALUES ({pk}, {pk}, {ck}, {ck}, 0)")
if pk == 5:
cql.execute(f"UPDATE {keyspace}.{table} SET s = 10 WHERE pk1 = {pk} AND pk2 = {pk}")
nodetool.flush(cql, f"{keyspace}.{table}")
nodetool.flush(cql, f"{keyspace}.{table}")
return table, schema
def clustering_table_with_collection(cql, keyspace):
table = util.unique_name()
schema = f"CREATE TABLE {keyspace}.{table} (pk int, ck int, v1 map<int, text>, v2 set<int>, v3 list<int>, PRIMARY KEY (pk, ck)) WITH compaction = {{'class': 'NullCompactionStrategy'}}"
cql.execute(schema)
for pk in range(0, 10):
for ck in range(0, 10):
map_vals = {f"{p}: '{c}'" for p in range(0, pk) for c in range(0, ck)}
map_str = ", ".join(map_vals)
set_list_vals = list(range(0, pk))
set_list_str = ", ".join(map(str, set_list_vals))
cql.execute(f"INSERT INTO {keyspace}.{table} (pk, ck, v1, v2, v3) VALUES ({pk}, {ck}, {{{map_str}}}, {{{set_list_str}}}, [{set_list_str}])")
if pk == 5:
nodetool.flush(cql, f"{keyspace}.{table}")
nodetool.flush(cql, f"{keyspace}.{table}")
return table, schema
def clustering_table_with_udt(cql, keyspace):
table = util.unique_name()
create_type_schema = f"CREATE TYPE IF NOT EXISTS {keyspace}.type1 (f1 int, f2 text)"
create_table_schema = f" CREATE TABLE {keyspace}.{table} (pk int, ck int, v type1, PRIMARY KEY (pk, ck)) WITH compaction = {{'class': 'NullCompactionStrategy'}}"
cql.execute(create_type_schema)
cql.execute(create_table_schema)
for pk in range(0, 10):
for ck in range(0, 10):
cql.execute(f"INSERT INTO {keyspace}.{table} (pk, ck, v) VALUES ({pk}, {ck}, {{f1: 100, f2: 'asd'}})")
if pk == 5:
nodetool.flush(cql, f"{keyspace}.{table}")
nodetool.flush(cql, f"{keyspace}.{table}")
return table, "; ".join((create_type_schema, create_table_schema))
def table_with_counters(cql, keyspace):
table = util.unique_name()
schema = f"CREATE TABLE {keyspace}.{table} (pk int PRIMARY KEY, v counter) WITH compaction = {{'class': 'NullCompactionStrategy'}}"
cql.execute(schema)
for pk in range(0, 10):
for c in range(0, 4):
cql.execute(f"UPDATE {keyspace}.{table} SET v = v + 1 WHERE pk = {pk};")
if pk == 5:
nodetool.flush(cql, f"{keyspace}.{table}")
nodetool.flush(cql, f"{keyspace}.{table}")
return table, schema
@contextlib.contextmanager
def scylla_sstable(table_factory, cql, ks, data_dir):
table, schema = table_factory(cql, ks)
schema_file = os.path.join(data_dir, "..", "test_tools_schema.cql")
with open(schema_file, "w") as f:
f.write(schema)
sstables = glob.glob(os.path.join(data_dir, ks, table + '-*', '*-Data.db'))
try:
yield (schema_file, sstables)
finally:
cql.execute(f"DROP TABLE {ks}.{table}")
os.unlink(schema_file)
def one_sstable(sstables):
assert len(sstables) > 1
return [sstables[0]]
def all_sstables(sstables):
assert len(sstables) > 1
return sstables
@pytest.mark.parametrize("what", ["index", "compression-info", "summary", "statistics", "scylla-metadata"])
@pytest.mark.parametrize("which_sstables", [one_sstable, all_sstables])
def test_scylla_sstable_dump_component(cql, test_keyspace, scylla_path, scylla_data_dir, what, which_sstables):
with scylla_sstable(simple_clustering_table, cql, test_keyspace, scylla_data_dir) as (schema_file, sstables):
out = subprocess.check_output([scylla_path, "sstable", f"dump-{what}", "--schema-file", schema_file] + which_sstables(sstables))
print(out)
assert out
assert json.loads(out)
@pytest.mark.parametrize("table_factory", [
simple_no_clustering_table,
simple_clustering_table,
clustering_table_with_collection,
clustering_table_with_udt,
table_with_counters,
])
@pytest.mark.parametrize("merge", [True, False])
@pytest.mark.parametrize("output_format", ["text", "json"])
def test_scylla_sstable_dump_data(cql, test_keyspace, scylla_path, scylla_data_dir, table_factory, merge, output_format):
with scylla_sstable(simple_clustering_table, cql, test_keyspace, scylla_data_dir) as (schema_file, sstables):
args = [scylla_path, "sstable", "dump-data", "--schema-file", schema_file, "--output-format", output_format]
if merge:
args.append("--merge")
out = subprocess.check_output(args + sstables)
print(out)
assert out
if output_format == "json":
assert json.loads(out)
@pytest.mark.parametrize("table_factory", [
simple_no_clustering_table,
simple_clustering_table,
])
def test_scylla_sstable_write(cql, test_keyspace, scylla_path, scylla_data_dir, table_factory):
with scylla_sstable(table_factory, cql, test_keyspace, scylla_data_dir) as (schema_file, sstables):
with tempfile.TemporaryDirectory() as tmp_dir:
dump_common_args = [scylla_path, "sstable", "dump-data", "--schema-file", schema_file, "--output-format", "json", "--merge"]
generation = util.unique_key_int()
original_out = subprocess.check_output(dump_common_args + sstables)
original_json = json.loads(original_out)["sstables"]["anonymous"]
input_file = os.path.join(tmp_dir, 'input.json')
with open(input_file, 'w') as f:
json.dump(original_json, f)
subprocess.check_call([scylla_path, "sstable", "write", "--schema-file", schema_file, "--input-file", input_file, "--output-dir", tmp_dir, "--generation", str(generation), '--logger-log-level', 'scylla-sstable=trace'])
sstable_file = os.path.join(tmp_dir, f"me-{generation}-big-Data.db")
actual_out = subprocess.check_output(dump_common_args + [sstable_file])
actual_json = json.loads(actual_out)["sstables"]["anonymous"]
assert actual_json == original_json
def script_consume_test_table_factory(cql, keyspace):
table = util.unique_name()
schema = f"CREATE TABLE {keyspace}.{table} (pk int, ck int, v int, s int STATIC, PRIMARY KEY (pk, ck)) WITH compaction = {{'class': 'NullCompactionStrategy'}}"
cql.execute(schema)
partitions = 4
for sst in range(0, 2):
for pk in range(sst * partitions, (sst + 1) * partitions):
# static row
cql.execute(f"UPDATE {keyspace}.{table} SET s = 10 WHERE pk = {pk}")
# range tombstone
cql.execute(f"DELETE FROM {keyspace}.{table} WHERE pk = {pk} AND ck >= 0 AND ck <= 4")
# 2 rows
for ck in range(0, 4):
cql.execute(f"INSERT INTO {keyspace}.{table} (pk, ck, v) VALUES ({pk}, {ck}, 0)")
nodetool.flush(cql, f"{keyspace}.{table}")
return table, schema
def test_scylla_sstable_script_consume_sstable(cql, test_keyspace, scylla_path, scylla_data_dir):
script_file = os.path.join(scylla_data_dir, "..", "test_scylla_sstable_script_consume_sstable.lua")
script = """
wr = Scylla.new_json_writer()
i = 0
function arg(args, arg)
ret = nil
wr:key(arg)
if args[arg] then
ret = tonumber(args[arg])
wr:int(ret)
else
wr:null()
end
return ret
end
function basename(path)
s, e = string.find(string.reverse(path), '/', 1, true)
return string.sub(path, #path - s + 2)
end
function consume_stream_start(args)
wr:start_object()
start_sst = arg(args, "start_sst")
end_sst = arg(args, "end_sst")
wr:key("content")
wr:start_array()
end
function consume_sstable_start(sst)
wrote_ps = false
i = i + 1
if i == start_sst then
return false
end
wr:string(basename(sst.filename))
end
function consume_partition_start(ps)
if not wrote_ps then
wr:string("ps")
wrote_ps = true
end
end
function consume_sstable_end()
if i == end_sst then
return false
end
end
function consume_stream_end()
wr:end_array()
wr:end_object()
end
"""
with open(script_file, 'w') as f:
f.write(script)
with scylla_sstable(script_consume_test_table_factory, cql, test_keyspace, scylla_data_dir) as (schema_file, sstables):
sst1 = os.path.basename(sstables[0])
sst2 = os.path.basename(sstables[1])
def run_scenario(script_args, expected):
print(f"Scenario: '{script_args}'\n")
if script_args:
script_args = ["--script-arg", script_args]
else:
script_args = []
script_args = [scylla_path, "sstable", "script", "--schema-file", schema_file, "--script-file", script_file] + script_args + sstables[0:2]
res = json.loads(subprocess.check_output(script_args))
assert res == expected
run_scenario("", {'start_sst': None, 'end_sst': None, 'content': [sst1, "ps", sst2, "ps"]})
run_scenario("start_sst=1", {'start_sst': 1, 'end_sst': None, 'content': [sst2, "ps"]})
run_scenario("start_sst=2", {'start_sst': 2, 'end_sst': None, 'content': [sst1, "ps"]})
run_scenario("start_sst=1:end_sst=1", {'start_sst': 1, 'end_sst': 1, 'content': []})
run_scenario("start_sst=2:end_sst=2", {'start_sst': 2, 'end_sst': 2, 'content': [sst1, "ps"]})
run_scenario("end_sst=1", {'start_sst': None, 'end_sst': 1, 'content': [sst1, "ps"]})
run_scenario("end_sst=2", {'start_sst': None, 'end_sst': 2, 'content': [sst1, "ps", sst2, "ps"]})
def test_scylla_sstable_script_slice(cql, test_keyspace, scylla_path, scylla_data_dir):
class bound:
@staticmethod
def unpack_value(value):
if isinstance(value, tuple):
return value
else:
return None, value
def __init__(self, value, weight):
self.token, self.value = self.unpack_value(value)
self.weight = weight
def tri_cmp(self, value):
if self.token is None and self.value is None:
assert(self.weight)
return -self.weight
token, value = self.unpack_value(value)
if token is None:
res = 0
else:
res = int(token) - int(self.token)
if res == 0 and not value is None and not self.value is None :
res = int(value) - int(self.value)
return res if res else -self.weight
def get_value(self, lookup_table, is_start):
if self.token is None and self.value is None:
return '-inf' if is_start else '+inf'
if self.value is None:
return "t{}".format(int(self.token))
return lookup_table[self.value]
@staticmethod
def before(value):
return bound(value, -1)
@staticmethod
def at(value):
return bound(value, 0)
@staticmethod
def after(value):
return bound(value, 1)
class interval:
def __init__(self, start_bound, end_bound):
self.start = start_bound
self.end = end_bound
def contains(self, value):
return self.start.tri_cmp(value) >= 0 and self.end.tri_cmp(value) <= 0
def summarize_dump(dump):
summary = []
for partition in list(dump["sstables"].items())[0][1]:
partition_summary = {"pk": partition["key"]["value"], "token": partition["key"]["token"], "frags": []}
if "static_row" in partition:
partition_summary["frags"].append(("sr", None))
for clustering_fragment in partition.get("clustering_elements", []):
type_str = "cr" if clustering_fragment["type"] == "clustering-row" else "rtc"
partition_summary["frags"].append((type_str, clustering_fragment["key"]["value"]))
summary.append(partition_summary)
return summary
def filter_summary(summary, partition_ranges, clustering_ranges):
if not partition_ranges:
return summary
filtered_summary = []
for partition in summary:
if any(map(lambda x: interval.contains(x, (partition["token"], partition["pk"])), partition_ranges)):
filtered_summary.append({"pk": partition["pk"], "token": partition["token"], "frags": []})
for (t, k) in partition["frags"]:
if t == "rtc" or k is None or not clustering_ranges or any(map(lambda x: interval.contains(x, k), clustering_ranges)):
filtered_summary[-1]["frags"].append((t, k))
return filtered_summary
def serialize_ranges(prefix, ranges, lookup_table):
serialized_ranges = []
i = 0
for r in ranges:
s = r.start.get_value(lookup_table, True)
e = r.end.get_value(lookup_table, False)
serialized_ranges.append("{}{}={}{},{}{}".format(
prefix,
i,
"(" if s == '-inf' or r.start.weight > 0 else "[",
s,
e,
")" if e == '+inf' or r.end.weight < 0 else "]"))
i = i + 1
return serialized_ranges
scripts_path = os.path.realpath(os.path.join(__file__, '../../../tools/scylla-sstable-scripts'))
script_file = os.path.join(scripts_path, 'slice.lua')
with scylla_sstable(script_consume_test_table_factory, cql, test_keyspace, scylla_data_dir) as (schema_file, sstables):
reference_summary = summarize_dump(json.loads(subprocess.check_output([scylla_path, "sstable", "dump-data", "--schema-file", schema_file, "--merge"] + sstables)))
# same order as in dump
pks = [(p["token"], p["pk"]) for p in reference_summary]
cks = set()
for p in reference_summary:
for t, ck in p["frags"]:
if not ck is None:
cks.add(ck)
cks = sorted(list(cks))
serialized_pk_lookup = {pk: subprocess.check_output([scylla_path, "types", "serialize", "--full-compound", "-t", "Int32Type", "--", pk]).strip().decode() for t, pk in pks}
serialized_ck_lookup = {ck: subprocess.check_output([scylla_path, "types", "serialize", "--prefix-compound", "-t", "Int32Type", "--", ck]).strip().decode() for ck in cks}
script_common_args = [scylla_path, "sstable", "script", "--schema-file", schema_file, "--merge", "--script-file", script_file]
def run_scenario(scenario, partition_ranges, clustering_ranges):
print(f"running scenario {scenario}")
script_args = serialize_ranges("pr", partition_ranges, serialized_pk_lookup) + serialize_ranges("cr", clustering_ranges, serialized_ck_lookup)
if script_args:
script_args = ["--script-arg"] + [":".join(script_args)]
print(f"script_args={script_args}")
expected = filter_summary(reference_summary, partition_ranges, clustering_ranges)
out = subprocess.check_output(script_common_args + script_args + sstables)
summary = summarize_dump(json.loads(out))
assert summary == expected
run_scenario("no args", [], [])
run_scenario("full range", [interval(bound.before(None), bound.after(None))], [])
run_scenario("(pks[0], +inf)", [interval(bound.after(pks[0]), bound.after(None))], [])
run_scenario("(-inf, pks[-3]]", [interval(bound.before(None), bound.after(pks[-3]))], [])
run_scenario("[pks[2], pks[-2]]", [interval(bound.before(pks[2]), bound.after(pks[-2]))], [])
run_scenario("[pks[0], pks[1]], [pks[2], pks[3]]", [interval(bound.before(pks[1]), bound.after(pks[2])), interval(bound.before(pks[3]), bound.after(pks[4]))], [])
run_scenario("[t:pks[2], t:pks[-2]]", [interval(bound.before((pks[2][0], None)), bound.after((pks[-2][0], None)))], [])
run_scenario("full pk range | [-inf, cks[2]]", [interval(bound.before(None), bound.after(None))], [interval(bound.before(None), bound.after(cks[2]))])
run_scenario("[pks[0], pks[1]] | (cks[0], cks[1]], (cks[2], +inf)", [interval(bound.before(pks[1]), bound.after(pks[2]))],
[interval(bound.after(cks[0]), bound.after(cks[1])), interval(bound.after(cks[2]), bound.after(None))])
@pytest.mark.parametrize("table_factory", [
simple_no_clustering_table,
simple_clustering_table,
clustering_table_with_collection,
clustering_table_with_udt,
table_with_counters,
])
def test_scylla_sstable_script(cql, test_keyspace, scylla_path, scylla_data_dir, table_factory):
scripts_path = os.path.realpath(os.path.join(__file__, '../../../tools/scylla-sstable-scripts'))
slice_script_path = os.path.join(scripts_path, 'slice.lua')
dump_script_path = os.path.join(scripts_path, 'dump.lua')
with scylla_sstable(table_factory, cql, test_keyspace, scylla_data_dir) as (schema_file, sstables):
dump_common_args = [scylla_path, "sstable", "dump-data", "--schema-file", schema_file, "--output-format", "json"]
script_common_args = [scylla_path, "sstable", "script", "--schema-file", schema_file]
# without --merge
cxx_json = json.loads(subprocess.check_output(dump_common_args + sstables))
dump_lua_json = json.loads(subprocess.check_output(script_common_args + ["--script-file", dump_script_path] + sstables))
slice_lua_json = json.loads(subprocess.check_output(script_common_args + ["--script-file", slice_script_path] + sstables))
assert dump_lua_json == cxx_json
assert slice_lua_json == cxx_json
# with --merge
cxx_json = json.loads(subprocess.check_output(dump_common_args + ["--merge"] + sstables))
dump_lua_json = json.loads(subprocess.check_output(script_common_args + ["--merge", "--script-file", dump_script_path] + sstables))
slice_lua_json = json.loads(subprocess.check_output(script_common_args + ["--merge", "--script-file", slice_script_path] + sstables))
assert dump_lua_json == cxx_json
assert slice_lua_json == cxx_json