From c82ca2be29ad3cbebc57248c87255351aad3db98 Mon Sep 17 00:00:00 2001
From: David Ames <david.ames@canonical.com>
Date: Tue, 6 Feb 2018 15:24:25 -0800
Subject: [PATCH] Allow setting gmcast.peer_timeout value

In some resource constrained environments particularly during deploy
time percona-cluster nodes can experience time outs during inter-node
communication.

This changes makes the gmcast.peer_timeout configurable based on the
galara cluster documentation:
http://galeracluster.com/documentation-webpages/galeraparameters.html

Warning: Changing this value from the default may have unintended
consequences. This should only be used when constraint's call for it.

Closes-Bug: #1742683
Change-Id: If93d6ba9d0e99b6af59358a7f3ea40e3aa2a6dbc
---
 config.yaml                      | 10 ++++++++++
 hooks/percona_hooks.py           |  6 +++++-
 hooks/percona_utils.py           | 19 ++++++++++++++++++
 unit_tests/test_percona_utils.py | 34 ++++++++++++++++++++++++++++++++
 4 files changed, 68 insertions(+), 1 deletion(-)

diff --git a/config.yaml b/config.yaml
index 7b67fc1..15aba57 100644
--- a/config.yaml
+++ b/config.yaml
@@ -308,3 +308,13 @@ options:
       collisions. Known wait is the amount of time between one node executing
       an operation and another. On slower hardware this value may need to be
       larger than the default of 30 seconds.
+  peer-timeout:
+    type: string
+    default:
+    description: |
+        This setting sets the gmcast.peer_timeout value. Possible values are documented
+        on the galera cluster site http://galeracluster.com/documentation-webpages/galeraparameters.html
+        For very busy clouds or in resource restricted environments this value can be changed.
+        WARNING Please read all documentation before changing the default value which may have
+        unintended consequences. It may be necessary to set this value higher during deploy time
+        (PTS15) and subsequently change it back to the default (PT3S) after deployment.
diff --git a/hooks/percona_hooks.py b/hooks/percona_hooks.py
index 1f3173c..ff5c994 100755
--- a/hooks/percona_hooks.py
+++ b/hooks/percona_hooks.py
@@ -108,6 +108,7 @@ from percona_utils import (
     LeaderNoBootstrapUUIDError,
     update_root_password,
     cluster_wait,
+    get_wsrep_provider_options,
 )
 
 from charmhelpers.core.unitdata import kv
@@ -197,11 +198,14 @@ def render_config(clustered=False, hosts=None):
         # See lp 1380747 for more info. This is intended as a stop gap until
         # percona package is fixed to support ipv6.
         context['bind_address'] = '::'
-        context['wsrep_provider_options'] = 'gmcast.listen_addr=tcp://:::4567;'
         context['ipv6'] = True
     else:
         context['ipv6'] = False
 
+    wsrep_provider_options = get_wsrep_provider_options()
+    if wsrep_provider_options:
+        context['wsrep_provider_options'] = wsrep_provider_options
+
     context.update(PerconaClusterHelper().parse_config())
     render(os.path.basename(config_file),
            config_file, context, perms=0o444)
diff --git a/hooks/percona_utils.py b/hooks/percona_utils.py
index 87f2331..e24e875 100644
--- a/hooks/percona_utils.py
+++ b/hooks/percona_utils.py
@@ -925,3 +925,22 @@ def cluster_wait():
         for rid in relation_ids('cluster'):
             num_nodes += len(related_units(rid))
     distributed_wait(modulo=num_nodes, wait=wait)
+
+
+def get_wsrep_provider_options():
+
+    wsrep_provider_options = []
+
+    if config('prefer-ipv6'):
+        wsrep_provider_options.append('gmcast.listen_addr=tcp://:::4567')
+
+    peer_timeout = config('peer-timeout')
+    if peer_timeout and(not peer_timeout.startswith('PT') or
+                        not peer_timeout.endswith('S')):
+        raise ValueError("Invalid gcast.peer_timeout value: {}"
+                         .format(peer_timeout))
+    elif peer_timeout:
+        wsrep_provider_options.append('gmcast.peer_timeout={}'
+                                      .format(config('peer-timeout')))
+
+    return ';'.join(wsrep_provider_options)
diff --git a/unit_tests/test_percona_utils.py b/unit_tests/test_percona_utils.py
index 3fc1381..789ce4c 100644
--- a/unit_tests/test_percona_utils.py
+++ b/unit_tests/test_percona_utils.py
@@ -241,6 +241,40 @@ class UtilsTests(unittest.TestCase):
         _wsrep_value.side_effect = [True, 2]
         self.assertTrue(percona_utils.cluster_in_sync())
 
+    @mock.patch("percona_utils.config")
+    def test_get_wsrep_provider_options(self, mock_config):
+        # Empty
+        _config = {"min-cluster-size": 3}
+        mock_config.side_effect = lambda key: _config.get(key)
+        expected = ""
+        self.assertEqual(percona_utils.get_wsrep_provider_options(),
+                         expected)
+
+        # IPv6 only
+        _config = {"prefer-ipv6": True}
+        mock_config.side_effect = lambda key: _config.get(key)
+        expected = "gmcast.listen_addr=tcp://:::4567"
+        self.assertEqual(percona_utils.get_wsrep_provider_options(),
+                         expected)
+        # ipv6 and peer_timeout
+        _config = {"peer-timeout": "PT15S",
+                   "prefer-ipv6": True}
+        mock_config.side_effect = lambda key: _config.get(key)
+        expected = ("gmcast.listen_addr=tcp://:::4567;"
+                    "gmcast.peer_timeout=PT15S")
+        self.assertEqual(percona_utils.get_wsrep_provider_options(),
+                         expected)
+
+        # peer_timeout bad setting
+        _config = {"peer-timeout": "10"}
+        mock_config.side_effect = lambda key: _config.get(key)
+        with self.assertRaises(ValueError):
+            percona_utils.get_wsrep_provider_options()
+        _config = {"peer-timeout": "PT10M"}
+        mock_config.side_effect = lambda key: _config.get(key)
+        with self.assertRaises(ValueError):
+            percona_utils.get_wsrep_provider_options()
+
 
 TO_PATCH = [
     'is_sufficient_peers',