From 7ddb3649317a12f9e5e367b6d4f2bd0e5ca84dbd Mon Sep 17 00:00:00 2001
From: Sage Weil <sage@newdream.net>
Date: Tue, 30 Nov 2021 18:18:31 -0500
Subject: [PATCH] mgr: limit changes to pg_num

We need to avoid making drastic changes to pg_num that outpace pgp_num or
else we will may hit the per-osd pg limits.

Fixes: https://tracker.ceph.com/issues/53442
Signed-off-by: Sage Weil <sage@newdream.net>
(cherry picked from commit 3b2a11249aff6ee608efc95212d6723df180cd07)

Conflicts:
	src/common/options/mgr.yaml.in

Cherry-pick notes:
- Options defined in common/options.cc in Octopus vs. common/options/mgr.yaml.in
---
 src/common/options.cc   |  5 +++++
 src/mgr/DaemonServer.cc | 11 ++++++++++-
 2 files changed, 15 insertions(+), 1 deletion(-)

diff --git a/src/common/options.cc b/src/common/options.cc
index 4fa13e1076b3e..3ddc5fa90c0f2 100644
--- a/src/common/options.cc
+++ b/src/common/options.cc
@@ -5264,6 +5264,11 @@ std::vector<Option> get_global_options() {
     .add_service("mgr")
     .set_description("Path to cephadm utility"),
 
+    Option("mgr_max_pg_num_change", Option::TYPE_UINT, Option::LEVEL_ADVANCED)
+    .set_default(128)
+    .add_service("mgr")
+    .set_description("maximum change in pg_num"),
+
     Option("mgr_module_path", Option::TYPE_STR, Option::LEVEL_ADVANCED)
     .set_default(CEPH_DATADIR "/mgr")
     .add_service("mgr")
diff --git a/src/mgr/DaemonServer.cc b/src/mgr/DaemonServer.cc
index 9b5af411d394c..6d18354effd45 100644
--- a/src/mgr/DaemonServer.cc
+++ b/src/mgr/DaemonServer.cc
@@ -2635,15 +2635,24 @@ void DaemonServer::adjust_pgs()
 	    } else {
 	      active = false;
 	    }
+	    unsigned pg_gap = p.get_pg_num() - p.get_pgp_num();
+        unsigned max_jump = cct->_conf.get_val<uint64_t>("mgr_max_pg_num_change");
 	    if (!active) {
 	      dout(10) << "pool " << i.first
 		       << " pg_num_target " << p.get_pg_num_target()
 		       << " pg_num " << p.get_pg_num()
 		       << " - not all pgs active"
 		       << dendl;
+	    } else if (pg_gap >= max_jump) {
+	      dout(10) << "pool " << i.first
+		       << " pg_num " << p.get_pg_num()
+		       << " - pgp_num " << p.get_pgp_num()
+		       << " gap > max_pg_num_change " << max_jump
+		       << " - must scale pgp_num first"
+		       << dendl;
 	    } else {
 	      unsigned add = std::min(
-		left,
+		std::min(left, max_jump - pg_gap),
 		p.get_pg_num_target() - p.get_pg_num());
 	      unsigned target = p.get_pg_num() + add;
 	      left -= add;
-- 
2.39.5