Merge PR #30525 into master

* refs/pull/30525/head:
	qa/tasks/ceph.conf.template: disable power-of-2 warning
	qa/standalone/mon/health-mute: use power of 2 for pg_num
	osd/OSDMap: remove remaining g_conf() usage
	PendingReleaseNotes: add note for 14.2.5 so we can backport this
	osd/OSDMap: health alert for non-power-of-two pg_num

Reviewed-by: Kai Wagner <kwagner@suse.com>
Reviewed-by: Nathan Cutler <ncutler@suse.com>
Reviewed-by: xie xingguo <xie.xingguo@zte.com.cn>

PendingReleaseNotes

@@ -23,7 +23,21 @@
   ensuring that all clients are hammer or newer allows new features
   only supported for `straw2` buckets to be used, including the
   `crush-compat` mode for the :ref:`balancer`.
-
+
+
+14.2.5
+------
+
+* 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
+
 
 >=15.0.0
 --------
@@ -148,6 +162,16 @@
 * 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

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
 ________________
 

qa/standalone/mon/health-mute.sh

@@ -31,7 +31,7 @@ function TEST_mute() {
     run_osd $dir 1 || return 1
     run_osd $dir 2 || return 1
 
-    ceph osd pool create foo 10
+    ceph osd pool create foo 8
     ceph osd pool application enable foo rbd --yes-i-really-mean-it
     wait_for_clean || return 1
 

qa/tasks/ceph.conf.template

@@ -30,6 +30,7 @@
 	mon warn on no sortbitwise = false
 	mon warn on osd down out interval zero = false
 	mon warn on too few osds = false
+	mon_warn_on_pool_pg_num_not_power_of_two = false
 
         osd pool default erasure code profile = "plugin=jerasure technique=reed_sol_van k=2 m=1 ruleset-failure-domain=osd crush-failure-domain=osd"
 

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")

src/mon/OSDMonitor.cc

@@ -2005,7 +2005,7 @@ void OSDMonitor::encode_pending(MonitorDBStore::TransactionRef t)
 
   // health
   health_check_map_t next;
-  tmp.check_health(&next);
+  tmp.check_health(cct, &next);
   encode_health(next, t);
 }
 

src/osd/OSDMap.cc

@@ -26,7 +26,6 @@
 #include "common/errno.h"
 #include "common/Formatter.h"
 #include "common/TextTable.h"
-#include "global/global_context.h"
 #include "include/ceph_features.h"
 #include "include/str_map.h"
 
@@ -5468,7 +5467,8 @@ 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();
 
@@ -5513,7 +5513,7 @@ void OSDMap::check_health(health_check_map_t *checks) const
 	    break;
 	  type = crush->get_bucket_type(parent_id);
 	  if (!subtree_type_is_down(
-		g_ceph_context, parent_id, type,
+		cct, parent_id, type,
 		&down_in_osds, &up_in_osds, &subtree_up, &subtree_type_down))
 	    break;
 	  current = parent_id;
@@ -5645,7 +5645,7 @@ void OSDMap::check_health(health_check_map_t *checks) const
   {
     // An osd could configure failsafe ratio, to something different
     // but for now assume it is the same here.
-    float fsr = g_conf()->osd_failsafe_full_ratio;
+    float fsr = cct->_conf->osd_failsafe_full_ratio;
     if (fsr > 1.0) fsr /= 100;
     float fr = get_full_ratio();
     float br = get_backfillfull_ratio();
@@ -5798,19 +5798,19 @@ void OSDMap::check_health(health_check_map_t *checks) const
   }
 
   // OLD_CRUSH_TUNABLES
-  if (g_conf()->mon_warn_on_legacy_crush_tunables) {
+  if (cct->_conf->mon_warn_on_legacy_crush_tunables) {
     string min = crush->get_min_required_version();
-    if (min < g_conf()->mon_crush_min_required_version) {
+    if (min < cct->_conf->mon_crush_min_required_version) {
       ostringstream ss;
       ss << "crush map has legacy tunables (require " << min
-	 << ", min is " << g_conf()->mon_crush_min_required_version << ")";
+	 << ", min is " << cct->_conf->mon_crush_min_required_version << ")";
       auto& d = checks->add("OLD_CRUSH_TUNABLES", HEALTH_WARN, ss.str(), 0);
       d.detail.push_back("see http://docs.ceph.com/docs/master/rados/operations/crush-map/#tunables");
     }
   }
 
   // OLD_CRUSH_STRAW_CALC_VERSION
-  if (g_conf()->mon_warn_on_crush_straw_calc_version_zero) {
+  if (cct->_conf->mon_warn_on_crush_straw_calc_version_zero) {
     if (crush->get_straw_calc_version() == 0) {
       ostringstream ss;
       ss << "crush map has straw_calc_version=0";
@@ -5821,7 +5821,7 @@ void OSDMap::check_health(health_check_map_t *checks) const
   }
 
   // CACHE_POOL_NO_HIT_SET
-  if (g_conf()->mon_warn_on_cache_pools_without_hit_sets) {
+  if (cct->_conf->mon_warn_on_cache_pools_without_hit_sets) {
     list<string> detail;
     for (auto p = pools.cbegin(); p != pools.cend(); ++p) {
       const pg_pool_t& info = p->second;
@@ -5900,6 +5900,27 @@ 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,

src/osd/OSDMap.h

@@ -1546,7 +1546,7 @@ class OSDMap {
   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,

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(cct.get(), &checks);
     JSONFormatter jf(true);
     jf.dump_object("checks", checks);
     jf.flush(cout);