From f16246604f8332be03f11c9f45ed7a4d7f985393 Mon Sep 17 00:00:00 2001
From: Dominik Gleich <dominik.gleich@memgraph.io>
Date: Thu, 16 Nov 2017 10:27:29 +0100
Subject: [PATCH] Add recovery speed measurement tool
Reviewers: mferencevic, buda
Reviewed By: buda
Subscribers: pullbot
Differential Revision: https://phabricator.memgraph.io/D985
---
tests/macro_benchmark/databases.py | 10 +++++
tools/requirements.txt | 2 +-
tools/snapshot_recovery_speed.py | 62 ++++++++++++++++++++++++++++++
3 files changed, 73 insertions(+), 1 deletion(-)
create mode 100755 tools/snapshot_recovery_speed.py
diff --git a/tests/macro_benchmark/databases.py b/tests/macro_benchmark/databases.py
index 96ebdcf01..ade5fbfe9 100644
--- a/tests/macro_benchmark/databases.py
+++ b/tests/macro_benchmark/databases.py
@@ -32,6 +32,9 @@ class Memgraph:
default=get_absolute_path("memgraph", "build"))
argp.add_argument("--port", default="7687",
help="Database and client port")
+ argp.add_argument("--snapshot-directory", default=None)
+ argp.add_argument("--snapshot-on-exit", action="store_true")
+ argp.add_argument("--snapshot-recover-on-startup", action="store_true")
self.log.info("Initializing Runner with arguments %r", args)
self.args, _ = argp.parse_known_args(args)
self.config = config
@@ -45,6 +48,13 @@ class Memgraph:
database_args = ["--port", self.args.port]
if self.num_workers:
database_args += ["--num_workers", str(self.num_workers)]
+ if self.args.snapshot_directory:
+ database_args += ["--snapshot-directory",
+ self.args.snapshot_directory]
+ if self.args.snapshot_recover_on_startup:
+ database_args += ["--snapshot-recover-on-startup"]
+ if self.args.snapshot_on_exit:
+ database_args += ["--snapshot-on-exit"]
# find executable path
runner_bin = self.args.runner_bin
diff --git a/tools/requirements.txt b/tools/requirements.txt
index e86db6d17..d3c37c4f1 100644
--- a/tools/requirements.txt
+++ b/tools/requirements.txt
@@ -1,8 +1,8 @@
cycler==0.10.0
matplotlib==2.0.2
numpy==1.13.1
-pkg-resources==0.0.0
pyparsing==2.2.0
python-dateutil==2.6.1
pytz==2017.2
six==1.11.0
+tabulate==0.8.1
diff --git a/tools/snapshot_recovery_speed.py b/tools/snapshot_recovery_speed.py
new file mode 100755
index 000000000..10549fcaf
--- /dev/null
+++ b/tools/snapshot_recovery_speed.py
@@ -0,0 +1,62 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+
+import time
+import sys
+import os
+import tempfile
+from tabulate import tabulate
+from timeit import default_timer as timer
+
+# hackish way to resuse existing start code
+sys.path.append(os.path.dirname(os.path.realpath(__file__)) +
+ "/../tests/macro_benchmark/")
+from databases import *
+from clients import *
+from common import get_absolute_path
+
+def main():
+ path = get_absolute_path("benchmarking.conf", "config")
+ tmp_dir = tempfile.TemporaryDirectory()
+
+ SNAPSHOT_DIR_ARG = ["--snapshot-directory", tmp_dir.name]
+ MAKE_SNAPSHOT_ARGS = ["--snapshot-on-exit"] + SNAPSHOT_DIR_ARG
+ RECOVER_SNAPSHOT_ARGS = ["--snapshot-recover-on-startup"] + SNAPSHOT_DIR_ARG
+ snapshot_memgraph = Memgraph(MAKE_SNAPSHOT_ARGS, path, 1)
+ recover_memgraph = Memgraph(RECOVER_SNAPSHOT_ARGS, path, 1)
+ client = QueryClient(None, 1)
+
+ results = []
+ for node_cnt in [10**6, 5*10**6, 10**7]:
+ for edge_per_node in [0, 1, 3]:
+ for prop_per_node in [0, 1]:
+ snapshot_memgraph.start()
+ properties = "{}".format(",".join(
+ ["p{}: 0".format(x) for x in range(prop_per_node)]))
+ client(["UNWIND RANGE(1, {}) AS _ CREATE ({{ {} }})".format(node_cnt, properties)], snapshot_memgraph)
+ client(["UNWIND RANGE(1, {}) AS _ MATCH (n) CREATE (n)-[:l]->(n)"
+ .format(edge_per_node)], snapshot_memgraph)
+ snapshot_memgraph.stop()
+
+ # This waits for the snapshot to be recovered and then exits
+ start = timer()
+ recover_memgraph.start()
+ recover_memgraph.stop()
+ stop = timer()
+ diff = stop - start
+
+ snapshots = os.listdir(tmp_dir.name)
+ assert len(snapshots) == 1
+
+ snap_path = tmp_dir.name + "/" + snapshots[0]
+ snap_size = round(os.path.getsize(snap_path) / 1024. / 1024., 2)
+ os.remove(snap_path)
+
+ edge_cnt = edge_per_node * node_cnt
+ results.append((node_cnt, edge_cnt, prop_per_node, snap_size, diff))
+
+ print(tabulate(tabular_data=results, headers=["Nodes", "Edges",
+ "Properties", "Snapshot size (MB)", "Elapsed time (s)"]))
+
+if __name__ == "__main__":
+ main()