From 26915bc54f0963b643afa8270d66e1dbfaaefc15 Mon Sep 17 00:00:00 2001
From: Harshit Bansal <harshitbansal2015@gmail.com>
Date: Tue, 22 Aug 2017 07:44:33 +0000
Subject: [PATCH] scripts: Make `purge-old-deployments` configurable.

Expands `purge-old-deployments` such that now it accepts the threshold
days as argument. Also `clean-unused-caches` script is automatically
run after purging the old deployments so that the orphaned caches
gets automatically cleaned.

Fixes: #5726.
---
 scripts/purge-old-deployments | 85 +++++++++++++++++++++--------------
 1 file changed, 52 insertions(+), 33 deletions(-)

diff --git a/scripts/purge-old-deployments b/scripts/purge-old-deployments
index 6933d1b5cb..34b5864db4 100755
--- a/scripts/purge-old-deployments
+++ b/scripts/purge-old-deployments
@@ -1,40 +1,59 @@
 #!/usr/bin/env python3
-import sys
+import argparse
 import os
-import logging
-import datetime
-import shutil
+import subprocess
+import sys
 
-sys.path.append(os.path.join(os.path.dirname(__file__), '..'))
-from scripts.lib.zulip_tools import DEPLOYMENTS_DIR, TIMESTAMP_FORMAT
+if False:
+    from typing import Set, Text
 
-logging.basicConfig(format="%(asctime)s purge-deployments: %(message)s",
-                    level=logging.INFO)
+ZULIP_PATH = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+sys.path.append(ZULIP_PATH)
+from scripts.lib.zulip_tools import DEPLOYMENTS_DIR, get_recent_deployments, \
+    may_be_perform_purging
 
-deployments_in_use = set()
-for basename in ['current', 'last', 'next']:
-    # Note which symlinks are current
-    path = os.path.abspath(os.readlink(os.path.join(DEPLOYMENTS_DIR, basename)))
-    deployments_in_use.add(path)
+def parse_args():
+    # type: () -> argparse.Namespace
+    parser = argparse.ArgumentParser(
+        description="This script can be used for cleaning old unused deployments.",
+        epilog="Orphaned/unused caches older than threshold days will be automatically "
+        "examined and removed.")
+    parser.add_argument(
+        "--threshold", dest="threshold_days", type=int, default=14,
+        nargs="?", metavar="<days>", help="Deployments older than "
+        "threshold days will be deleted. (defaults to 14)")
+    parser.add_argument(
+        "--dry-run", dest="dry_run", action="store_true",
+        help="If specified then script will only print the deployments and "
+        "caches that it will delete/keep back. It will not delete anything.")
+    args = parser.parse_args()
+    return args
 
-one_week_ago = datetime.datetime.now() - datetime.timedelta(days=7)
-to_purge = []
-for filename in os.listdir(DEPLOYMENTS_DIR):
-    try:
-        date = datetime.datetime.strptime(filename, TIMESTAMP_FORMAT)
-        if date < one_week_ago:
-            if os.path.abspath(os.path.join(DEPLOYMENTS_DIR, filename)) in deployments_in_use:
-                # Never purge an in-use deployment
-                continue
-            to_purge.append(filename)
-    except ValueError:
-        pass
+def get_deployments_to_be_purged(recent_deployments):
+    # type: (Set[Text]) -> Set[Text]
+    all_deployments = set([os.path.join(DEPLOYMENTS_DIR, deployment)
+                           for deployment in os.listdir(DEPLOYMENTS_DIR)])
+    deployments_to_purge = set()
+    for deployment in all_deployments:
+        if deployment not in recent_deployments:
+            deployments_to_purge.add(deployment)
+    return deployments_to_purge
 
-if to_purge:
-    to_purge.sort()
-    logging.info("Purging the following old deployments directories: %s" % (", ".join(to_purge),))
-    for filename in to_purge:
-        shutil.rmtree(os.path.join(DEPLOYMENTS_DIR, filename))
-        logging.info("Finished %s" % (filename))
-else:
-    logging.info("No old deployment directories to purge")
+def main():
+    # type: () -> None
+    args = parse_args()
+    deployments_to_keep = get_recent_deployments(args.threshold_days)
+    deployments_to_purge = get_deployments_to_be_purged(deployments_to_keep)
+
+    may_be_perform_purging(deployments_to_purge, deployments_to_keep, "deployment", args.dry_run)
+
+    if not args.dry_run:
+        print("Deployments cleaned successfully...")
+        print("Cleaning orphaned/unused caches...")
+
+    # Call 'clean-unused-caches' script to clean any orphaned/unused caches.
+    subprocess.check_call(["./scripts/clean-unused-caches"] + sys.argv[1:])
+    print("Done!\n")
+
+if __name__ == "__main__":
+    main()