From 6e46b1c0e50ad377d5db058e0452b0d956b3fad6 Mon Sep 17 00:00:00 2001
From: Sage Weil <sage@redhat.com>
Date: Mon, 23 Sep 2019 14:40:48 -0500
Subject: [PATCH] osd/OSDMap: health alert for non-power-of-two pg_num

Fixes: https://tracker.ceph.com/issues/41647
Signed-off-by: Sage Weil <sage@redhat.com>
---
 PendingReleaseNotes                    | 10 ++++++++++
 doc/rados/operations/health-checks.rst | 17 +++++++++++++++++
 src/common/options.cc                  |  5 +++++
 src/mon/OSDMonitor.cc                  |  2 +-
 src/osd/OSDMap.cc                      | 24 +++++++++++++++++++++++-
 src/osd/OSDMap.h                       |  2 +-
 src/tools/osdmaptool.cc                |  2 +-
 7 files changed, 58 insertions(+), 4 deletions(-)

diff --git a/PendingReleaseNotes b/PendingReleaseNotes
index 41a8886cc63..0f305adadac 100644
--- a/PendingReleaseNotes
+++ b/PendingReleaseNotes
@@ -142,6 +142,16 @@
 * New OSD daemon command dump_scrub_reservations which reveals the
   scrub reservations that are held for local (primary) and remote (replica) PGs.
 
+* Ceph will now issue a health warning if a RADOS pool as a ``pg_num``
+  value that is not a power of two.  This can be fixed by adjusting
+  the pool to a nearby power of two::
+
+    ceph osd pool set <pool-name> pg_num <new-pg-num>
+
+  Alternatively, the warning can be silenced with::
+
+    ceph config set global mon_warn_on_pool_pg_num_not_power_of_two false
+
 * The ``pg_autoscale_mode`` is now set to ``on`` by default for newly
   created pools, which means that Ceph will automatically manage the
   number of PGs.  To change this behavior, or to learn more about PG
diff --git a/doc/rados/operations/health-checks.rst b/doc/rados/operations/health-checks.rst
index 0068dc4a585..2536cd4f24d 100644
--- a/doc/rados/operations/health-checks.rst
+++ b/doc/rados/operations/health-checks.rst
@@ -740,6 +740,23 @@ The PG count for existing pools can be increased or new pools can be created.
 Please refer to :ref:`choosing-number-of-placement-groups` for more
 information.
 
+POOL_PG_NUM_NOT_POWER_OF_TWO
+____________________________
+
+One or more pools has a ``pg_num`` value that is not a power of two.
+Although this is not strictly incorrect, it does lead to a less
+balanced distribution of data because some PGs have roughly twice as
+much data as others.
+
+This is easily corrected by setting the ``pg_num`` value for the
+affected pool(s) to a nearby power of two::
+
+  ceph osd pool set <pool-name> pg_num <value>
+
+This health warning can be disabled with::
+
+  ceph config set global mon_warn_on_pool_pg_num_not_power_of_two false
+
 POOL_TOO_FEW_PGS
 ________________
 
diff --git a/src/common/options.cc b/src/common/options.cc
index 0ebbe4b3037..f60182b5fa5 100644
--- a/src/common/options.cc
+++ b/src/common/options.cc
@@ -1702,6 +1702,11 @@ std::vector<Option> get_global_options() {
     .add_service("mgr")
     .set_description("issue POOL_APP_NOT_ENABLED health warning if pool has not application enabled"),
 
+    Option("mon_warn_on_pool_pg_num_not_power_of_two", Option::TYPE_BOOL, Option::LEVEL_DEV)
+    .set_default(true)
+    .add_service("mon")
+    .set_description("issue POOL_PG_NUM_NOT_POWER_OF_TWO warning if pool has a non-power-of-two pg_num value"),
+
     Option("mon_warn_on_misplaced", Option::TYPE_BOOL, Option::LEVEL_ADVANCED)
     .set_default(false)
     .add_service("mgr")
diff --git a/src/mon/OSDMonitor.cc b/src/mon/OSDMonitor.cc
index 9b08b0bba7b..be6894bd003 100644
--- a/src/mon/OSDMonitor.cc
+++ b/src/mon/OSDMonitor.cc
@@ -2014,7 +2014,7 @@ void OSDMonitor::encode_pending(MonitorDBStore::TransactionRef t)
 
   // health
   health_check_map_t next;
-  tmp.check_health(&next);
+  tmp.check_health(g_ceph_context, &next);
   encode_health(next, t);
 }
 
diff --git a/src/osd/OSDMap.cc b/src/osd/OSDMap.cc
index 4e0b7adc609..36444b172f5 100644
--- a/src/osd/OSDMap.cc
+++ b/src/osd/OSDMap.cc
@@ -5472,7 +5472,8 @@ void print_osd_utilization(const OSDMap& osdmap,
   }
 }
 
-void OSDMap::check_health(health_check_map_t *checks) const
+void OSDMap::check_health(CephContext *cct,
+			  health_check_map_t *checks) const
 {
   int num_osds = get_num_osds();
 
@@ -5904,6 +5905,27 @@ void OSDMap::check_health(health_check_map_t *checks) const
       d.detail.swap(nearfull_detail);
     }
   }
+
+  // POOL_PG_NUM_NOT_POWER_OF_TWO
+  if (cct->_conf.get_val<bool>("mon_warn_on_pool_pg_num_not_power_of_two")) {
+    list<string> detail;
+    for (auto it : get_pools()) {
+      if (!isp2(it.second.get_pg_num_target())) {
+	ostringstream ss;
+	ss << "pool '" << get_pool_name(it.first)
+	   << "' pg_num " << it.second.get_pg_num_target()
+	   << " is not a power of two";
+	detail.push_back(ss.str());
+      }
+    }
+    if (!detail.empty()) {
+      ostringstream ss;
+      ss << detail.size() << " pool(s) have non-power-of-two pg_num";
+      auto& d = checks->add("POOL_PG_NUM_NOT_POWER_OF_TWO", HEALTH_WARN,
+			    ss.str(), detail.size());
+      d.detail.swap(detail);
+    }
+  }
 }
 
 int OSDMap::parse_osd_id_list(const vector<string>& ls, set<int> *out,
diff --git a/src/osd/OSDMap.h b/src/osd/OSDMap.h
index 5888f5dd4b3..f0055611b27 100644
--- a/src/osd/OSDMap.h
+++ b/src/osd/OSDMap.h
@@ -1546,7 +1546,7 @@ public:
   static void generate_test_instances(std::list<OSDMap*>& o);
   bool check_new_blacklist_entries() const { return new_blacklist_entries; }
 
-  void check_health(health_check_map_t *checks) const;
+  void check_health(CephContext *cct, health_check_map_t *checks) const;
 
   int parse_osd_id_list(const std::vector<std::string>& ls,
 			std::set<int> *out,
diff --git a/src/tools/osdmaptool.cc b/src/tools/osdmaptool.cc
index 1299d342939..eed36e4e6a3 100644
--- a/src/tools/osdmaptool.cc
+++ b/src/tools/osdmaptool.cc
@@ -659,7 +659,7 @@ int main(int argc, const char **argv)
 
   if (health) {
     health_check_map_t checks;
-    osdmap.check_health(&checks);
+    osdmap.check_health(g_ceph_context, &checks);
     JSONFormatter jf(true);
     jf.dump_object("checks", checks);
     jf.flush(cout);
-- 
2.39.5