osd: increase default hard pg limit

The hard limit was too easy to hit during failure cases, especially
due to mismatched device sizes causing high variance in pgs/osd.

Fixes: https://tracker.ceph.com/issues/24243
Signed-off-by: Josh Durgin <jdurgin@redhat.com>
(cherry picked from commit dfdedce89521f89046fb13287d9fdd683c91e242)

diff --git a/src/common/options.cc b/src/common/options.cc
index 20acf359bfeab302c6b3670951bb862a4e8f23d2..9b9260b426752df63496c8a082b86213e7dcb362 100644 (file)
--- a/src/common/options.cc
+++ b/src/common/options.cc
@@ -3157,7 +3157,7 @@ std::vector<Option> get_global_options() {
     .set_description("Number of entries difference above which to use asynchronous recovery when appropriate"),
 
     Option("osd_max_pg_per_osd_hard_ratio", Option::TYPE_FLOAT, Option::LEVEL_ADVANCED)
-    .set_default(2)
+    .set_default(3)
     .set_min(1)
     .set_description("Maximum number of PG per OSD, a factor of 'mon_max_pg_per_osd'")
     .set_long_description("OSD will refuse to instantiate PG if the number of PG it serves exceeds this number.")