From 247bea1a8a76a1ac37b84d679d9767bbce4782cd Mon Sep 17 00:00:00 2001
From: Krutika Dhananjay <kdhananj@redhat.com>
Date: Thu, 15 Feb 2018 16:12:12 +0530
Subject: [PATCH 665/675] features/shard: Fix shard inode refcount when it's
part of priv->lru_list.
For as long as a shard's inode is in priv->lru_list, it should have a non-zero
ref-count. This patch achieves it by taking a ref on the inode when it
is added to lru list. When it's time for the inode to be evicted
from the lru list, a corresponding unref is done.
> Upstream: https://review.gluster.org/19608
> BUG: 1468483
> Change-Id: I289ffb41e7be5df7489c989bc1bbf53377433c86
Change-Id: I289ffb41e7be5df7489c989bc1bbf53377433c86
BUG: 1583462
Signed-off-by: Krutika Dhananjay <kdhananj@redhat.com>
Reviewed-on: https://code.engineering.redhat.com/gerrit/140381
Tested-by: RHGS Build Bot <nigelb@redhat.com>
Reviewed-by: Sunil Kumar Heggodu Gopala Acharya <sheggodu@redhat.com>
---
tests/bugs/shard/shard-inode-refcount-test.t | 27 +++++++++++++++++++++++++++
tests/volume.rc | 18 ++++++++++++++++++
xlators/features/shard/src/shard.c | 26 +++++++++++++++++---------
3 files changed, 62 insertions(+), 9 deletions(-)
create mode 100644 tests/bugs/shard/shard-inode-refcount-test.t
diff --git a/tests/bugs/shard/shard-inode-refcount-test.t b/tests/bugs/shard/shard-inode-refcount-test.t
new file mode 100644
index 0000000..6358097
--- /dev/null
+++ b/tests/bugs/shard/shard-inode-refcount-test.t
@@ -0,0 +1,27 @@
+#!/bin/bash
+
+. $(dirname $0)/../../include.rc
+. $(dirname $0)/../../volume.rc
+
+cleanup
+
+TEST glusterd
+TEST pidof glusterd
+TEST $CLI volume create $V0 $H0:$B0/${V0}0
+TEST $CLI volume set $V0 features.shard on
+TEST $CLI volume set $V0 features.shard-block-size 4MB
+TEST $CLI volume start $V0
+
+TEST glusterfs --volfile-id=$V0 --volfile-server=$H0 $M0
+
+TEST dd if=/dev/zero of=$M0/one-plus-five-shards bs=1M count=23
+
+ACTIVE_INODES_BEFORE=$(get_mount_active_size_value $V0)
+TEST rm -f $M0/one-plus-five-shards
+EXPECT `expr $ACTIVE_INODES_BEFORE - 5` get_mount_active_size_value $V0
+
+EXPECT_WITHIN $UMOUNT_TIMEOUT "Y" force_umount $M0
+TEST $CLI volume stop $V0
+TEST $CLI volume delete $V0
+
+cleanup
diff --git a/tests/volume.rc b/tests/volume.rc
index 5d6c96c..9ab31af 100644
--- a/tests/volume.rc
+++ b/tests/volume.rc
@@ -760,3 +760,21 @@ function count_sh_entries()
{
ls $1/.glusterfs/indices/xattrop | grep -v "xattrop-" | wc -l
}
+
+function get_mount_active_size_value {
+ local vol=$1
+ local statedump=$(generate_mount_statedump $vol)
+ sleep 1
+ local val=$(grep "active_size" $statedump | cut -f2 -d'=' | tail -1)
+ rm -f $statedump
+ echo $val
+}
+
+function get_mount_lru_size_value {
+ local vol=$1
+ local statedump=$(generate_mount_statedump $vol)
+ sleep 1
+ local val=$(grep "lru_size" $statedump | cut -f2 -d'=' | tail -1)
+ rm -f $statedump
+ echo $val
+}
diff --git a/xlators/features/shard/src/shard.c b/xlators/features/shard/src/shard.c
index 4351220..c57a426 100644
--- a/xlators/features/shard/src/shard.c
+++ b/xlators/features/shard/src/shard.c
@@ -502,6 +502,10 @@ __shard_update_shards_inode_list (inode_t *linked_inode, xlator_t *this,
* by empty list), and if there is still space in the priv list,
* add this ctx to the tail of the list.
*/
+ /* For as long as an inode is in lru list, we try to
+ * keep it alive by holding a ref on it.
+ */
+ inode_ref (linked_inode);
gf_uuid_copy (ctx->base_gfid, base_inode->gfid);
ctx->block_num = block_num;
list_add_tail (&ctx->ilist, &priv->ilist_head);
@@ -527,8 +531,16 @@ __shard_update_shards_inode_list (inode_t *linked_inode, xlator_t *this,
/* The following unref corresponds to the ref held by
* inode_find() above.
*/
- inode_forget (lru_inode, 0);
inode_unref (lru_inode);
+ /* The following unref corresponds to the ref held at
+ * the time the shard was created or looked up
+ */
+ inode_unref (lru_inode);
+ inode_forget (lru_inode, 0);
+ /* For as long as an inode is in lru list, we try to
+ * keep it alive by holding a ref on it.
+ */
+ inode_ref (linked_inode);
gf_uuid_copy (ctx->base_gfid, base_inode->gfid);
ctx->block_num = block_num;
list_add_tail (&ctx->ilist, &priv->ilist_head);
@@ -1658,11 +1670,6 @@ shard_link_block_inode (shard_local_t *local, int block_num, inode_t *inode,
buf);
inode_lookup (linked_inode);
list_index = block_num - local->first_block;
-
- /* Defer unref'ing the inodes until write is complete. These inodes are
- * unref'd in the event of a failure or after successful fop completion
- * in shard_local_wipe().
- */
local->inode_list[list_index] = linked_inode;
LOCK(&priv->lock);
@@ -2520,10 +2527,11 @@ shard_unlink_block_inode (shard_local_t *local, int shard_block_num)
if (!list_empty (&ctx->ilist)) {
list_del_init (&ctx->ilist);
priv->inode_count--;
+ GF_ASSERT (priv->inode_count >= 0);
+ inode_unlink (inode, priv->dot_shard_inode, block_bname);
+ inode_unref (inode);
+ inode_forget (inode, 0);
}
- GF_ASSERT (priv->inode_count >= 0);
- inode_unlink (inode, priv->dot_shard_inode, block_bname);
- inode_forget (inode, 0);
}
UNLOCK(&priv->lock);
--
1.8.3.1