[PATCH 1/2] mds: don't create bloom filter for incomplete dir

["Yan, Zheng" <zheng.z.yan@xxxxxxxxx>]

From: "Yan, Zheng" <zheng.z.yan@xxxxxxxxx>

Creating bloom filter for incomplete dir that was added by log
replay will confuse subsequent dir lookup and can create null
dentry for existing file. The erroneous null dentry confuses the
fragstat accounting and causes undeletable empty directory.

The fix is check if the dir is complete before creating the bloom
filter. For the MDCache::trim_non_auth{,_subtree} cases, just do
not call CDir::add_to_bloom because bloom filter is useless for
replica.

Signed-off-by: Yan, Zheng <zheng.z.yan@xxxxxxxxx>
---
 src/mds/CDir.cc    |  6 +++++-
 src/mds/MDCache.cc | 12 +++++++-----
 2 files changed, 12 insertions(+), 6 deletions(-)

diff --git a/src/mds/CDir.cc b/src/mds/CDir.cc
index c5220ed..835a0e9 100644
--- a/src/mds/CDir.cc
+++ b/src/mds/CDir.cc
@@ -615,8 +615,12 @@ void CDir::unlink_inode_work( CDentry *dn )
 
 void CDir::add_to_bloom(CDentry *dn)
 {
-  if (!bloom)
+  if (!bloom) {
+    /* not create bloom filter for incomplete dir that was added by log replay */
+    if (!is_complete())
+      return;
     bloom = new bloom_filter(100, 0.05, 0);
+  }
   /* This size and false positive probability is completely random.*/
   bloom->insert(dn->name.c_str(), dn->name.size());
 }
diff --git a/src/mds/MDCache.cc b/src/mds/MDCache.cc
index 7152b5e..fcc1ecd 100644
--- a/src/mds/MDCache.cc
+++ b/src/mds/MDCache.cc
@@ -5498,9 +5498,10 @@ void MDCache::trim_dentry(CDentry *dn, map<int, MCacheExpire*>& expiremap)
   // adjust the dir state
   // NOTE: we can safely remove a clean, null dentry without effecting
   //       directory completeness.
-  // (do this _before_ we unlink the inode, below!)
+  // (check this _before_ we unlink the inode, below!)
+  bool clear_complete = false;
   if (!(dnl->is_null() && dn->is_clean())) 
-    dir->state_clear(CDir::STATE_COMPLETE); 
+    clear_complete = true;
   
   // unlink the dentry
   if (dnl->is_remote()) {
@@ -5520,6 +5521,9 @@ void MDCache::trim_dentry(CDentry *dn, map<int, MCacheExpire*>& expiremap)
   // remove dentry
   dir->add_to_bloom(dn);
   dir->remove_dentry(dn);
+
+  if (clear_complete)
+    dir->state_clear(CDir::STATE_COMPLETE);
   
   // reexport?
   if (dir->get_num_head_items() == 0 && dir->is_subtree_root())
@@ -5708,9 +5712,8 @@ void MDCache::trim_non_auth()
       else {
 	assert(dnl->is_null());
       }
-      dir->add_to_bloom(dn);
+
       dir->remove_dentry(dn);
-      
       // adjust the dir state
       dir->state_clear(CDir::STATE_COMPLETE);  // dir incomplete!
     }
@@ -5811,7 +5814,6 @@ bool MDCache::trim_non_auth_subtree(CDir *dir)
         dout(20) << "trim_non_auth_subtree(" << dir << ") removing inode " << in << " with dentry" << dn << dendl;
         dir->unlink_inode(dn);
         remove_inode(in);
-        dir->add_to_bloom(dn);
         dir->remove_dentry(dn);
       } else {
         dout(20) << "trim_non_auth_subtree(" << dir << ") keeping inode " << in << " with dentry " << dn <<dendl;
-- 
1.7.11.7

--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html