diff options
author | Pranith Kumar K <pkarampu@redhat.com> | 2015-07-26 11:19:28 +0530 |
---|---|---|
committer | Pranith Kumar Karampuri <pkarampu@redhat.com> | 2015-08-30 00:45:28 -0700 |
commit | b0e125e937cbf4bb02baaa799ff4968a4d1cb1d0 (patch) | |
tree | 0e66f3a507e847279b22cb9d34827e9b970011f0 | |
parent | d304916ddf3d6848787c3a668cc36e3395b32069 (diff) |
cluster/afr: Make [f]xattrop metadata transaction
Problem:
When xlators above afr do [f]xattrop when one of the bricks is down, after the
brick comes backup, the metadata is not healed because [f]xattrop is not
considered a transaction.
Fix:
Treat [f]xattrop as transaction so that changes done by xlators above afr are
marked for heal when some of the bricks were down at the time of [f]xattrop.
Change-Id: Iea180f9a456509847c3cd8d5d59a0cdc2712d334
BUG: 1248887
Signed-off-by: Pranith Kumar K <pkarampu@redhat.com>
Reviewed-on: http://review.gluster.org/11809
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Ravishankar N <ravishankar@redhat.com>
Tested-by: NetBSD Build System <jenkins@build.gluster.org>
-rw-r--r-- | tests/bugs/shard/bug-1248887.t | 38 | ||||
-rw-r--r-- | xlators/cluster/afr/src/afr-common.c | 172 | ||||
-rw-r--r-- | xlators/cluster/afr/src/afr-inode-write.c | 229 | ||||
-rw-r--r-- | xlators/cluster/afr/src/afr-inode-write.h | 8 | ||||
-rw-r--r-- | xlators/cluster/afr/src/afr.h | 8 |
5 files changed, 272 insertions, 183 deletions
diff --git a/tests/bugs/shard/bug-1248887.t b/tests/bugs/shard/bug-1248887.t new file mode 100644 index 00000000000..4528e9b9ac2 --- /dev/null +++ b/tests/bugs/shard/bug-1248887.t @@ -0,0 +1,38 @@ +#!/bin/bash + +. $(dirname $0)/../../include.rc +. $(dirname $0)/../../volume.rc + +cleanup; + +TEST glusterd +TEST pidof glusterd +TEST $CLI volume create $V0 replica 2 $H0:$B0/${V0}{0,1} +TEST $CLI volume set $V0 features.shard on +TEST $CLI volume start $V0 + +TEST $GFS --volfile-id=$V0 --volfile-server=$H0 $M0 +#Create a file. +TEST touch $M0/foo +#Write some data into it. +TEST `echo "abc" > $M0/foo` +EXPECT "0000000000400000" get_hex_xattr trusted.glusterfs.shard.block-size $B0/${V0}0/foo +EXPECT "0000000000000004000000000000000000000000000000010000000000000000" get_hex_xattr trusted.glusterfs.shard.file-size $B0/${V0}0/foo +EXPECT "0000000000400000" get_hex_xattr trusted.glusterfs.shard.block-size $B0/${V0}1/foo +EXPECT "0000000000000004000000000000000000000000000000010000000000000000" get_hex_xattr trusted.glusterfs.shard.file-size $B0/${V0}1/foo +TEST kill_brick $V0 $H0 $B0/${V0}0 +TEST `echo "abc" >> $M0/foo` +EXPECT "0000000000400000" get_hex_xattr trusted.glusterfs.shard.block-size $B0/${V0}1/foo +EXPECT "0000000000000008000000000000000000000000000000010000000000000000" get_hex_xattr trusted.glusterfs.shard.file-size $B0/${V0}1/foo +TEST $CLI volume start $V0 force +EXPECT_WITHIN $PROCESS_UP_TIMEOUT "Y" glustershd_up_status +EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 0 +EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 1 +TEST $CLI volume heal $V0 +EXPECT_WITHIN $HEAL_TIMEOUT "0" get_pending_heal_count $V0 +EXPECT "0000000000400000" get_hex_xattr trusted.glusterfs.shard.block-size $B0/${V0}0/foo +EXPECT "0000000000000008000000000000000000000000000000010000000000000000" get_hex_xattr trusted.glusterfs.shard.file-size $B0/${V0}0/foo +EXPECT "0000000000400000" get_hex_xattr trusted.glusterfs.shard.block-size $B0/${V0}1/foo +EXPECT "0000000000000008000000000000000000000000000000010000000000000000" get_hex_xattr trusted.glusterfs.shard.file-size $B0/${V0}1/foo + +cleanup; diff --git a/xlators/cluster/afr/src/afr-common.c b/xlators/cluster/afr/src/afr-common.c index e9a7c8c3649..f74d530a7aa 100644 --- a/xlators/cluster/afr/src/afr-common.c +++ b/xlators/cluster/afr/src/afr-common.c @@ -1221,6 +1221,11 @@ afr_replies_wipe (struct afr_reply *replies, int count) dict_unref (replies[i].xdata); replies[i].xdata = NULL; } + + if (replies[i].xattr) { + dict_unref (replies[i].xattr); + replies[i].xattr = NULL; + } } } @@ -1273,6 +1278,9 @@ afr_local_cleanup (afr_local_t *local, xlator_t *this) if (local->xattr_req) dict_unref (local->xattr_req); + if (local->xattr_rsp) + dict_unref (local->xattr_rsp); + if (local->dict) dict_unref (local->dict); @@ -1350,10 +1358,6 @@ afr_local_cleanup (afr_local_t *local, xlator_t *this) if (local->cont.xattrop.xattr) dict_unref (local->cont.xattrop.xattr); } - { /* fxattrop */ - if (local->cont.fxattrop.xattr) - dict_unref (local->cont.fxattrop.xattr); - } { /* symlink */ GF_FREE (local->cont.symlink.linkpath); } @@ -2905,166 +2909,6 @@ out: /* }}} */ -/* {{{ xattrop */ - -int32_t -afr_xattrop_cbk (call_frame_t *frame, void *cookie, - xlator_t *this, int32_t op_ret, int32_t op_errno, - dict_t *xattr, dict_t *xdata) -{ - afr_local_t *local = NULL; - int call_count = -1; - - local = frame->local; - - LOCK (&frame->lock); - { - if (op_ret == 0) { - if (!local->cont.xattrop.xattr) - local->cont.xattrop.xattr = dict_ref (xattr); - - if (!local->xdata_rsp && xdata) - local->xdata_rsp = dict_ref (xdata); - - local->op_ret = 0; - } - - local->op_errno = op_errno; - } - UNLOCK (&frame->lock); - - call_count = afr_frame_return (frame); - - if (call_count == 0) - AFR_STACK_UNWIND (xattrop, frame, local->op_ret, local->op_errno, - local->cont.xattrop.xattr, local->xdata_rsp); - - return 0; -} - - -int32_t -afr_xattrop (call_frame_t *frame, xlator_t *this, loc_t *loc, - gf_xattrop_flags_t optype, dict_t *xattr, dict_t *xdata) -{ - afr_private_t *priv = NULL; - afr_local_t *local = NULL; - int i = 0; - int32_t call_count = 0; - int32_t op_errno = ENOMEM; - - priv = this->private; - - local = AFR_FRAME_INIT (frame, op_errno); - if (!local) - goto out; - - call_count = local->call_count; - if (!call_count) { - op_errno = ENOTCONN; - goto out; - } - - for (i = 0; i < priv->child_count; i++) { - if (local->child_up[i]) { - STACK_WIND (frame, afr_xattrop_cbk, - priv->children[i], - priv->children[i]->fops->xattrop, - loc, optype, xattr, xdata); - if (!--call_count) - break; - } - } - - return 0; -out: - AFR_STACK_UNWIND (xattrop, frame, -1, op_errno, NULL, NULL); - - return 0; -} - -/* }}} */ - -/* {{{ fxattrop */ - -int32_t -afr_fxattrop_cbk (call_frame_t *frame, void *cookie, - xlator_t *this, int32_t op_ret, int32_t op_errno, - dict_t *xattr, dict_t *xdata) -{ - afr_local_t *local = NULL; - - int call_count = -1; - - local = frame->local; - - LOCK (&frame->lock); - { - if (op_ret == 0) { - if (!local->cont.fxattrop.xattr) - local->cont.fxattrop.xattr = dict_ref (xattr); - - if (!local->xdata_rsp && xdata) - local->xdata_rsp = dict_ref (xdata); - local->op_ret = 0; - } - - local->op_errno = op_errno; - } - UNLOCK (&frame->lock); - - call_count = afr_frame_return (frame); - - if (call_count == 0) - AFR_STACK_UNWIND (fxattrop, frame, local->op_ret, local->op_errno, - local->cont.fxattrop.xattr, local->xdata_rsp); - - return 0; -} - - -int32_t -afr_fxattrop (call_frame_t *frame, xlator_t *this, fd_t *fd, - gf_xattrop_flags_t optype, dict_t *xattr, dict_t *xdata) -{ - afr_private_t *priv = NULL; - afr_local_t *local = NULL; - int i = 0; - int32_t call_count = 0; - int32_t op_errno = 0; - - priv = this->private; - - local = AFR_FRAME_INIT (frame, op_errno); - if (!local) - goto out; - - call_count = local->call_count; - if (!call_count) { - op_errno = ENOTCONN; - goto out; - } - - for (i = 0; i < priv->child_count; i++) { - if (local->child_up[i]) { - STACK_WIND (frame, afr_fxattrop_cbk, - priv->children[i], - priv->children[i]->fops->fxattrop, - fd, optype, xattr, xdata); - if (!--call_count) - break; - } - } - - return 0; -out: - AFR_STACK_UNWIND (fxattrop, frame, -1, op_errno, NULL, NULL); - - return 0; -} - -/* }}} */ - int32_t afr_unlock_partial_inodelk_cbk (call_frame_t *frame, void *cookie, xlator_t *this, int32_t op_ret, diff --git a/xlators/cluster/afr/src/afr-inode-write.c b/xlators/cluster/afr/src/afr-inode-write.c index a1bf5b65491..5dd52cf9c0d 100644 --- a/xlators/cluster/afr/src/afr-inode-write.c +++ b/xlators/cluster/afr/src/afr-inode-write.c @@ -92,6 +92,12 @@ __afr_inode_write_finalize (call_frame_t *frame, xlator_t *this) local->xdata_rsp = dict_ref (local->replies[i].xdata); } + if (local->replies[i].xattr) { + if (local->xattr_rsp) + dict_unref (local->xattr_rsp); + local->xattr_rsp = + dict_ref (local->replies[i].xattr); + } } } @@ -102,7 +108,8 @@ __afr_inode_write_finalize (call_frame_t *frame, xlator_t *this) static void __afr_inode_write_fill (call_frame_t *frame, xlator_t *this, int child_index, int op_ret, int op_errno, - struct iatt *prebuf, struct iatt *postbuf, dict_t *xdata) + struct iatt *prebuf, struct iatt *postbuf, + dict_t *xattr, dict_t *xdata) { afr_local_t *local = NULL; @@ -117,6 +124,8 @@ __afr_inode_write_fill (call_frame_t *frame, xlator_t *this, int child_index, local->replies[child_index].prestat = *prebuf; if (postbuf) local->replies[child_index].poststat = *postbuf; + if (xattr) + local->replies[child_index].xattr = dict_ref (xattr); if (xdata) local->replies[child_index].xdata = dict_ref (xdata); } else { @@ -130,7 +139,7 @@ __afr_inode_write_fill (call_frame_t *frame, xlator_t *this, int child_index, static int __afr_inode_write_cbk (call_frame_t *frame, void *cookie, xlator_t *this, int32_t op_ret, int32_t op_errno, struct iatt *prebuf, - struct iatt *postbuf, dict_t *xdata) + struct iatt *postbuf, dict_t *xattr, dict_t *xdata) { afr_local_t *local = NULL; int child_index = (long) cookie; @@ -141,7 +150,8 @@ __afr_inode_write_cbk (call_frame_t *frame, void *cookie, xlator_t *this, LOCK (&frame->lock); { __afr_inode_write_fill (frame, this, child_index, op_ret, - op_errno, prebuf, postbuf, xdata); + op_errno, prebuf, postbuf, xattr, + xdata); } UNLOCK (&frame->lock); @@ -250,7 +260,7 @@ afr_writev_wind_cbk (call_frame_t *frame, void *cookie, xlator_t *this, LOCK (&frame->lock); { __afr_inode_write_fill (frame, this, child_index, op_ret, - op_errno, prebuf, postbuf, xdata); + op_errno, prebuf, postbuf, NULL, xdata); if (op_ret == -1 || !xdata) goto unlock; @@ -504,7 +514,7 @@ afr_truncate_wind_cbk (call_frame_t *frame, void *cookie, xlator_t *this, local->stable_write = _gf_false; return __afr_inode_write_cbk (frame, cookie, this, op_ret, op_errno, - prebuf, postbuf, xdata); + prebuf, postbuf, NULL, xdata); } @@ -623,7 +633,7 @@ afr_ftruncate_wind_cbk (call_frame_t *frame, void *cookie, xlator_t *this, local->stable_write = _gf_false; return __afr_inode_write_cbk (frame, cookie, this, op_ret, op_errno, - prebuf, postbuf, xdata); + prebuf, postbuf, NULL, xdata); } @@ -735,7 +745,7 @@ afr_setattr_wind_cbk (call_frame_t *frame, void *cookie, xlator_t *this, struct iatt *preop, struct iatt *postop, dict_t *xdata) { return __afr_inode_write_cbk (frame, cookie, this, op_ret, op_errno, - preop, postop, xdata); + preop, postop, NULL, xdata); } @@ -840,7 +850,7 @@ afr_fsetattr_wind_cbk (call_frame_t *frame, void *cookie, xlator_t *this, struct iatt *preop, struct iatt *postop, dict_t *xdata) { return __afr_inode_write_cbk (frame, cookie, this, op_ret, op_errno, - preop, postop, xdata); + preop, postop, NULL, xdata); } @@ -947,7 +957,7 @@ afr_setxattr_wind_cbk (call_frame_t *frame, void *cookie, xlator_t *this, int32_t op_ret, int32_t op_errno, dict_t *xdata) { return __afr_inode_write_cbk (frame, cookie, this, op_ret, op_errno, - NULL, NULL, xdata); + NULL, NULL, NULL, xdata); } @@ -1494,7 +1504,7 @@ afr_fsetxattr_wind_cbk (call_frame_t *frame, void *cookie, xlator_t *this, int32_t op_ret, int32_t op_errno, dict_t *xdata) { return __afr_inode_write_cbk (frame, cookie, this, op_ret, op_errno, - NULL, NULL, xdata); + NULL, NULL, NULL, xdata); } @@ -1608,7 +1618,7 @@ afr_removexattr_wind_cbk (call_frame_t *frame, void *cookie, xlator_t *this, int32_t op_ret, int32_t op_errno, dict_t *xdata) { return __afr_inode_write_cbk (frame, cookie, this, op_ret, op_errno, - NULL, NULL, xdata); + NULL, NULL, NULL, xdata); } @@ -1716,7 +1726,7 @@ afr_fremovexattr_wind_cbk (call_frame_t *frame, void *cookie, xlator_t *this, int32_t op_ret, int32_t op_errno, dict_t *xdata) { return __afr_inode_write_cbk (frame, cookie, this, op_ret, op_errno, - NULL, NULL, xdata); + NULL, NULL, NULL, xdata); } @@ -1826,7 +1836,7 @@ afr_fallocate_wind_cbk (call_frame_t *frame, void *cookie, xlator_t *this, struct iatt *postbuf, dict_t *xdata) { return __afr_inode_write_cbk (frame, cookie, this, op_ret, op_errno, - prebuf, postbuf, xdata); + prebuf, postbuf, NULL, xdata); } @@ -1940,7 +1950,7 @@ afr_discard_wind_cbk (call_frame_t *frame, void *cookie, xlator_t *this, struct iatt *postbuf, dict_t *xdata) { return __afr_inode_write_cbk (frame, cookie, this, op_ret, op_errno, - prebuf, postbuf, xdata); + prebuf, postbuf, NULL, xdata); } @@ -2050,7 +2060,7 @@ afr_zerofill_wind_cbk (call_frame_t *frame, void *cookie, xlator_t *this, struct iatt *postbuf, dict_t *xdata) { return __afr_inode_write_cbk (frame, cookie, this, op_ret, op_errno, - prebuf, postbuf, xdata); + prebuf, postbuf, NULL, xdata); } @@ -2132,3 +2142,192 @@ out: } /* }}} */ + +int32_t +afr_xattrop_wind_cbk (call_frame_t *frame, void *cookie, + xlator_t *this, int32_t op_ret, int32_t op_errno, + dict_t *xattr, dict_t *xdata) +{ + return __afr_inode_write_cbk (frame, cookie, this, op_ret, op_errno, + NULL, NULL, xattr, xdata); +} + +int +afr_xattrop_wind (call_frame_t *frame, xlator_t *this, int subvol) +{ + afr_local_t *local = NULL; + afr_private_t *priv = NULL; + + local = frame->local; + priv = this->private; + + STACK_WIND_COOKIE (frame, afr_xattrop_wind_cbk, (void *) (long) subvol, + priv->children[subvol], + priv->children[subvol]->fops->xattrop, + &local->loc, local->cont.xattrop.optype, + local->cont.xattrop.xattr, local->xdata_req); + return 0; +} + +int +afr_xattrop_unwind (call_frame_t *frame, xlator_t *this) +{ + afr_local_t *local = NULL; + call_frame_t *main_frame = NULL; + + local = frame->local; + + main_frame = afr_transaction_detach_fop_frame (frame); + if (!main_frame) + return 0; + + AFR_STACK_UNWIND (xattrop, main_frame, local->op_ret, local->op_errno, + local->xattr_rsp, local->xdata_rsp); + return 0; +} + +int32_t +afr_xattrop (call_frame_t *frame, xlator_t *this, loc_t *loc, + gf_xattrop_flags_t optype, dict_t *xattr, dict_t *xdata) +{ + afr_local_t *local = NULL; + call_frame_t *transaction_frame = NULL; + int ret = -1; + int op_errno = ENOMEM; + + transaction_frame = copy_frame (frame); + if (!transaction_frame) + goto out; + + local = AFR_FRAME_INIT (transaction_frame, op_errno); + if (!local) + goto out; + + local->cont.xattrop.xattr = dict_ref (xattr); + local->cont.xattrop.optype = optype; + if (xdata) + local->xdata_req = dict_ref (xdata); + + local->transaction.wind = afr_xattrop_wind; + local->transaction.fop = __afr_txn_write_fop; + local->transaction.done = __afr_txn_write_done; + local->transaction.unwind = afr_xattrop_unwind; + + loc_copy (&local->loc, loc); + local->inode = inode_ref (loc->inode); + + local->op = GF_FOP_XATTROP; + + local->transaction.main_frame = frame; + local->transaction.start = LLONG_MAX - 1; + local->transaction.len = 0; + + ret = afr_transaction (transaction_frame, this, AFR_METADATA_TRANSACTION); + if (ret < 0) { + op_errno = -ret; + goto out; + } + + return 0; +out: + if (transaction_frame) + AFR_STACK_DESTROY (transaction_frame); + + AFR_STACK_UNWIND (xattrop, frame, -1, op_errno, NULL, NULL); + return 0; +} + +int32_t +afr_fxattrop_wind_cbk (call_frame_t *frame, void *cookie, + xlator_t *this, int32_t op_ret, int32_t op_errno, + dict_t *xattr, dict_t *xdata) +{ + return __afr_inode_write_cbk (frame, cookie, this, op_ret, op_errno, + NULL, NULL, xattr, xdata); +} + +int +afr_fxattrop_wind (call_frame_t *frame, xlator_t *this, int subvol) +{ + afr_local_t *local = NULL; + afr_private_t *priv = NULL; + + local = frame->local; + priv = this->private; + + STACK_WIND_COOKIE (frame, afr_fxattrop_wind_cbk, (void *) (long) subvol, + priv->children[subvol], + priv->children[subvol]->fops->fxattrop, + local->fd, local->cont.xattrop.optype, + local->cont.xattrop.xattr, local->xdata_req); + return 0; +} + +int +afr_fxattrop_unwind (call_frame_t *frame, xlator_t *this) +{ + afr_local_t *local = NULL; + call_frame_t *main_frame = NULL; + + local = frame->local; + + main_frame = afr_transaction_detach_fop_frame (frame); + if (!main_frame) + return 0; + + AFR_STACK_UNWIND (fxattrop, main_frame, local->op_ret, local->op_errno, + local->xattr_rsp, local->xdata_rsp); + return 0; +} + +int32_t +afr_fxattrop (call_frame_t *frame, xlator_t *this, fd_t *fd, + gf_xattrop_flags_t optype, dict_t *xattr, dict_t *xdata) +{ + afr_local_t *local = NULL; + call_frame_t *transaction_frame = NULL; + int ret = -1; + int op_errno = ENOMEM; + + transaction_frame = copy_frame (frame); + if (!transaction_frame) + goto out; + + local = AFR_FRAME_INIT (transaction_frame, op_errno); + if (!local) + goto out; + + local->cont.xattrop.xattr = dict_ref (xattr); + local->cont.xattrop.optype = optype; + if (xdata) + local->xdata_req = dict_ref (xdata); + + local->transaction.wind = afr_fxattrop_wind; + local->transaction.fop = __afr_txn_write_fop; + local->transaction.done = __afr_txn_write_done; + local->transaction.unwind = afr_fxattrop_unwind; + + local->fd = fd_ref (fd); + local->inode = inode_ref (fd->inode); + + local->op = GF_FOP_FXATTROP; + + local->transaction.main_frame = frame; + local->transaction.start = LLONG_MAX - 1; + local->transaction.len = 0; + + ret = afr_transaction (transaction_frame, this, + AFR_METADATA_TRANSACTION); + if (ret < 0) { + op_errno = -ret; + goto out; + } + + return 0; +out: + if (transaction_frame) + AFR_STACK_DESTROY (transaction_frame); + + AFR_STACK_UNWIND (fxattrop, frame, -1, op_errno, NULL, NULL); + return 0; +} diff --git a/xlators/cluster/afr/src/afr-inode-write.h b/xlators/cluster/afr/src/afr-inode-write.h index 7b1fc552880..e174cc2d610 100644 --- a/xlators/cluster/afr/src/afr-inode-write.h +++ b/xlators/cluster/afr/src/afr-inode-write.h @@ -79,4 +79,12 @@ afr_fallocate (call_frame_t *frame, xlator_t *this, fd_t *fd, int32_t mode, int afr_zerofill(call_frame_t *frame, xlator_t *this, fd_t *fd, off_t offset, off_t len, dict_t *xdata); + +int32_t +afr_xattrop (call_frame_t *frame, xlator_t *this, loc_t *loc, + gf_xattrop_flags_t optype, dict_t *xattr, dict_t *xdata); + +int32_t +afr_fxattrop (call_frame_t *frame, xlator_t *this, fd_t *fd, + gf_xattrop_flags_t optype, dict_t *xattr, dict_t *xdata); #endif /* __INODE_WRITE_H__ */ diff --git a/xlators/cluster/afr/src/afr.h b/xlators/cluster/afr/src/afr.h index 7f995c731d7..3a4a8794698 100644 --- a/xlators/cluster/afr/src/afr.h +++ b/xlators/cluster/afr/src/afr.h @@ -257,6 +257,7 @@ struct afr_reply { int valid; int32_t op_ret; int32_t op_errno; + dict_t *xattr;/*For xattrop*/ dict_t *xdata; struct iatt poststat; struct iatt postparent; @@ -563,12 +564,9 @@ typedef struct _afr_local { struct { dict_t *xattr; + gf_xattrop_flags_t optype; } xattrop; - struct { - dict_t *xattr; - } fxattrop; - /* dir write */ struct { @@ -729,6 +727,8 @@ typedef struct _afr_local { dict_t *xdata_req; dict_t *xdata_rsp; + dict_t *xattr_rsp; /*for [f]xattrop*/ + mode_t umask; int xflag; gf_boolean_t do_discovery; |