Merge pull request #26942 from dzafman/wip-38616

Feature: Improvements to auto repair Reviewed-by: Neha Ojha <nojha@redhat.com>
ceph · Mar 26, 2019 · 3335774 · 3335774
2 parents 0247d56 + 769cdc8
commit 3335774
Show file tree

Hide file tree

Showing 19 changed files with 580 additions and 40 deletions.
diff --git a/doc/dev/placement-group.rst b/doc/dev/placement-group.rst
@@ -186,6 +186,9 @@ User-visible PG States
 *forced_backfill*
   the PG has been marked for highest priority backfill
 
+*failed_repair*
+  an attempt to repair the PG has failed.  Manual intervention is required.
+
 
 OMAP STATISTICS
 ===============

diff --git a/doc/rados/configuration/osd-config-ref.rst b/doc/rados/configuration/osd-config-ref.rst
@@ -361,8 +361,8 @@ scrubbing operations.
 ``osd scrub auto repair``
 
 :Description: Setting this to ``true`` will enable automatic pg repair when errors
-              are found in deep-scrub.  However, if more than ``osd scrub auto repair num errors``
-              errors are found a repair is NOT performed.
+              are found in scrub or deep-scrub.  However, if more than
+              ``osd scrub auto repair num errors`` errors are found a repair is NOT performed.
 :Type: Boolean
 :Default: ``false``
 

diff --git a/qa/standalone/osd/osd-rep-recov-eio.sh b/qa/standalone/osd/osd-rep-recov-eio.sh
@@ -110,18 +110,30 @@ function rados_get_data() {
 
     local poolname=pool-rep
     local objname=obj-$inject-$$
+    local pgid=$(get_pg $poolname $objname)
+
     rados_put $dir $poolname $objname || return 1
     inject_$inject rep data $poolname $objname $dir 0 || return 1
     rados_get $dir $poolname $objname || return 1
 
+    COUNT=$(ceph pg $pgid query | jq '.info.stats.stat_sum.num_objects_repaired')
+    test "$COUNT" = "1" || return 1
+
     inject_$inject rep data $poolname $objname $dir 0 || return 1
     inject_$inject rep data $poolname $objname $dir 1 || return 1
     rados_get $dir $poolname $objname || return 1
 
+    COUNT=$(ceph pg $pgid query | jq '.info.stats.stat_sum.num_objects_repaired')
+    test "$COUNT" = "2" || return 1
+
     inject_$inject rep data $poolname $objname $dir 0 || return 1
     inject_$inject rep data $poolname $objname $dir 1 || return 1
     inject_$inject rep data $poolname $objname $dir 2 || return 1
     rados_get $dir $poolname $objname hang || return 1
+
+    # After hang another repair couldn't happen, so count stays the same
+    COUNT=$(ceph pg $pgid query | jq '.info.stats.stat_sum.num_objects_repaired')
+    test "$COUNT" = "2" || return 1
 }
 
 function TEST_rados_get_with_eio() {