From 6df6a3ab81fc4c9da4c0a23825605a32194d6945 Mon Sep 17 00:00:00 2001
From: David Moreau Simard <dmsimard@redhat.com>
Date: Thu, 11 Jan 2018 14:41:04 -0500
Subject: [PATCH] Periodically retrieve and back up Zuul's status.json

Since the contents of the status.json is only available in-memory from
the Zuul scheduler, the contents are lost forever if the scheduler,
or the server it is hosted on, crash.

This implements two crons:
1) A curl to download a timestamped status.json, if it is available
2) File deletion so we keep no more than 180 timestamped status files
   at any time.

The files are stored at a location within the scope of the zuul-web
vhost in order to make it easily possible to dump queues with something
like:

    /opt/zuul/tools/zuul-changes.py http://zuulv3.openstack.org/backup/status_timestamp.json

Change-Id: Ia1923bfae0222559b8d2dcb5e0618848b3a618d0
---
 manifests/web.pp | 40 ++++++++++++++++++++++++++++++++++++++++
 1 file changed, 40 insertions(+)

diff --git a/manifests/web.pp b/manifests/web.pp
index 981dd14..424a40c 100644
--- a/manifests/web.pp
+++ b/manifests/web.pp
@@ -19,6 +19,7 @@ class zuul::web (
   $manage_log_conf = true,
   $web_listen_address = '127.0.0.1',
   $web_listen_port = 9000,
+  $enable_status_backups = true,
 ) {
 
   service { 'zuul-web':
@@ -76,6 +77,45 @@ class zuul::web (
     }
   }
 
+  file { '/var/lib/zuul/www/backup':
+    ensure  => directory,
+    require => File['/var/lib/zuul/www'],
+  }
+
+  if $enable_status_backups {
+    # Minutes, hours, days, etc are not specified here because we are
+    # interested in running this *every minute*.
+    # This is a mean of backing up status.json periodically in order to provide
+    # a mean of restoring lost scheduler queues if need be.
+    # We are downloading this file at a location served by the vhost so that we
+    # can query it easily should the need arise.
+    # If the status.json is unavailable for download, no new files are created.
+    if $zuul::proxy_ssl_cert_file_contents != '' {
+      $status = "https://${zuul::vhost_name}/status.json"
+    } else {
+      $status = "http://${zuul::vhost_name}/status.json"
+    }
+    cron { 'zuul_scheduler_status_backup':
+      user    => 'root',
+      command => "timeout -k 5 10 curl ${status} -o /var/lib/zuul/www/backup/status_$(date +\\%s).json",
+      require => [Package['curl'],
+                  User['zuul'],
+                  File['/var/lib/zuul/www/backup']],
+    }
+    # Rotate backups and keep no more than 120 files -- or 2 hours worth of
+    # backup if Zuul has 100% uptime.
+    # We're not basing the rotation on time because the scheduler/web service
+    # could be down for an extended period of time.
+    # This is ran hourly so technically up to ~3 hours worth of backups will
+    # be kept.
+    cron { 'zuul_scheduler_status_prune':
+      user    => 'root',
+      minute  => '0',
+      command => 'flock -n /var/run/status_prune.lock ls -dt -1 /var/lib/zuul/www/backup/* |sed -e "1,120d" |xargs rm -f',
+      require => Cron['zuul_scheduler_status_backup'],
+    }
+  }
+
   file { '/var/lib/zuul/www/static':
     ensure  => directory,
     require => File['/var/lib/zuul/www'],