]> git.apps.os.sepia.ceph.com Git - ceph.git/commitdiff
osd: PgPriorSet: any_survived -> any_is_alive_now
authorSage Weil <sage@newdream.net>
Mon, 17 Oct 2011 23:48:55 +0000 (16:48 -0700)
committerSage Weil <sage@newdream.net>
Mon, 17 Oct 2011 23:48:55 +0000 (16:48 -0700)
Signed-off-by: Sage Weil <sage@newdream.net>
src/osd/PG.cc

index a147febb6cc67c80902266b96ebac7ce2998efb3..4c07de3c979eea6553d579dc380a4185e89e1360 100644 (file)
@@ -4943,7 +4943,7 @@ PG::PgPriorSet::PgPriorSet(int whoami,
       continue;
 
     int need_down = 0;
-    bool any_survived = false;
+    bool any_is_alive_now = false;
 
     // consider UP osds
     for (unsigned i=0; i<interval.up.size(); i++) {
@@ -4977,7 +4977,7 @@ PG::PgPriorSet::PgPriorSet(int whoami,
        cur.insert(o);
 
        // did any osds survive _this_ interval?
-       any_survived = true;
+       any_is_alive_now = true;
       } else if (!pinfo || pinfo->lost_at > interval.first) {
        down.insert(o);
        if (started_since_joining.size()) {
@@ -5014,7 +5014,7 @@ PG::PgPriorSet::PgPriorSet(int whoami,
     // if nobody survived this interval, and we may have gone rw,
     // then we need to wait for one of those osds to recover to
     // ensure that we haven't lost any information.
-    if (!any_survived && need_down && interval.maybe_went_rw) {
+    if (!any_is_alive_now && need_down && interval.maybe_went_rw) {
       // fixme: how do we identify a "clean" shutdown anyway?
       dout(10) << "build_prior  " << need_down
          << " osds possibly went active+rw, no survivors, including" << dendl;