andy/glusterfs
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

protocol/client: Handle failures in lock self healing gracefully (part 1).

Browse Source 
During reopening of fd's and reacquiring of locks on the fd (after a
reconnect), a release on a fd on which reacquiring of locks is in progress
will free up fdctx. This patch will keep fdctx valid until the reacquiring
of locks is in progress.

Change-Id: I0464c751a5aa986abac0b72b48b261fceeba24e8
BUG: 795386
Signed-off-by: Mohammed Junaid <junaid@redhat.com>
Reviewed-on: http://review.gluster.com/2937
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Amar Tumballi <amarts@redhat.com>
Reviewed-by: Vijay Bellur <vijay@gluster.com>
This commit is contained in:
Mohammed Junaid 2012-03-19 21:07:33 +05:30 committed by Vijay Bellur
parent 83277598bd
commit 55c22fbfae
2 changed files with 51 additions and 93 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
xlators/protocol/client/src/client.h
View File

@ -38,6 +38,11 @@
#define GF_MAX_SOCKET_WINDOW_SIZE (1 * GF_UNIT_MB)
#define GF_MIN_SOCKET_WINDOW_SIZE (0)
typedef enum {
GF_LK_HEAL_IN_PROGRESS,
GF_LK_HEAL_DONE,
} lk_heal_state_t;
#define CLIENT_GET_REMOTE_FD(conf, fd, remote_fd, op_errno, label) \
do { \
clnt_fd_ctx_t *fdctx = NULL; \
@ -129,6 +134,7 @@ typedef struct _client_fd_ctx {
int32_t wbflags;
fd_lk_ctx_t *lk_ctx;
pthread_mutex_t mutex;
lk_heal_state_t lk_heal_state;
struct list_head lock_list; /* List of all granted locks on this fd */
} clnt_fd_ctx_t;

138
xlators/protocol/client/src/client3_1-fops.c
View File

@ -352,11 +352,12 @@ client3_1_open_cbk (struct rpc_req *req, struct iovec *iov, int count,
goto out;
}
fdctx->remote_fd = rsp.fd;
fdctx->inode = inode_ref (fd->inode);
fdctx->flags = local->flags;
fdctx->wbflags = local->wbflags;
fdctx->lk_ctx = fd_lk_ctx_ref (fd->lk_ctx);
fdctx->remote_fd = rsp.fd;
fdctx->inode = inode_ref (fd->inode);
fdctx->flags = local->flags;
fdctx->wbflags = local->wbflags;
fdctx->lk_ctx = fd_lk_ctx_ref (fd->lk_ctx);
fdctx->lk_heal_state = GF_LK_HEAL_DONE;
INIT_LIST_HEAD (&fdctx->sfd_pos);
INIT_LIST_HEAD (&fdctx->lock_list);
@ -1648,9 +1649,11 @@ client3_1_create_cbk (struct rpc_req *req, struct iovec *iov, int count,
goto out;
}
fdctx->remote_fd = rsp.fd;
fdctx->inode = inode_ref (inode);
fdctx->flags = local->flags;
fdctx->remote_fd = rsp.fd;
fdctx->inode = inode_ref (inode);
fdctx->flags = local->flags;
fdctx->lk_ctx = fd_lk_ctx_ref (fd->lk_ctx);
fdctx->lk_heal_state = GF_LK_HEAL_DONE;
INIT_LIST_HEAD (&fdctx->sfd_pos);
INIT_LIST_HEAD (&fdctx->lock_list);
@ -2252,13 +2255,14 @@ client3_1_releasedir_cbk (struct rpc_req *req, struct iovec *iov, int count,
int
client_fdctx_destroy (xlator_t *this, clnt_fd_ctx_t *fdctx)
{
clnt_conf_t *conf = NULL;
call_frame_t *fr = NULL;
int32_t ret = -1;
fd_lk_ctx_t *lk_ctx = NULL;
clnt_conf_t *conf = NULL;
call_frame_t *fr = NULL;
int32_t ret = -1;
char parent_down = 0;
fd_lk_ctx_t *lk_ctx = NULL;
if (!fdctx)
goto out;
GF_VALIDATE_OR_GOTO ("client", this, out);
GF_VALIDATE_OR_GOTO (this->name, fdctx, out);
conf = (clnt_conf_t *) this->private;
@ -2269,12 +2273,19 @@ client_fdctx_destroy (xlator_t *this, clnt_fd_ctx_t *fdctx)
pthread_mutex_lock (&conf->lock);
{
lk_ctx = fdctx->lk_ctx;
parent_down = conf->parent_down;
lk_ctx = fdctx->lk_ctx;
fdctx->lk_ctx = NULL;
}
pthread_mutex_unlock (&conf->lock);
fd_lk_ctx_unref (lk_ctx);
if (lk_ctx)
fd_lk_ctx_unref (lk_ctx);
if (!parent_down)
rpc_clnt_ref (conf->rpc);
else
goto out;
fr = create_frame (this, this->ctx->pool);
if (fr == NULL) {
@ -2303,6 +2314,7 @@ client_fdctx_destroy (xlator_t *this, clnt_fd_ctx_t *fdctx)
(xdrproc_t)xdr_gfs3_release_req);
}
rpc_clnt_unref (conf->rpc);
out:
if (fdctx) {
fdctx->remote_fd = -1;
@ -2323,13 +2335,10 @@ client3_1_releasedir (call_frame_t *frame, xlator_t *this,
clnt_conf_t *conf = NULL;
clnt_fd_ctx_t *fdctx = NULL;
clnt_args_t *args = NULL;
gfs3_releasedir_req req = {{0,},};
int64_t remote_fd = -1;
int ret = 0;
char parent_down = 0;
if (!frame || !this || !data)
goto unwind;
goto out;
args = data;
conf = this->private;
@ -2353,37 +2362,9 @@ client3_1_releasedir (call_frame_t *frame, xlator_t *this,
}
pthread_mutex_unlock (&conf->lock);
if (remote_fd != -1) {
pthread_mutex_lock (&conf->lock);
{
parent_down = conf->parent_down;
if (!parent_down) {
rpc_clnt_ref (conf->rpc);
}
}
pthread_mutex_unlock (&conf->lock);
if (!parent_down) {
req.fd = remote_fd;
client_submit_request (this, &req, frame,
conf->fops,
GFS3_OP_RELEASEDIR,
client3_1_releasedir_cbk,
NULL, NULL, 0, NULL, 0,
NULL,
(xdrproc_t)xdr_gfs3_releasedir_req);
rpc_clnt_unref (conf->rpc);
}
inode_unref (fdctx->inode);
GF_FREE (fdctx);
}
unwind:
if (ret)
STACK_DESTROY (frame->root);
if (remote_fd != -1)
client_fdctx_destroy (this, fdctx);
out:
return 0;
}
@ -2392,16 +2373,14 @@ int32_t
client3_1_release (call_frame_t *frame, xlator_t *this,
void *data)
{
int64_t remote_fd = -1;
clnt_conf_t *conf = NULL;
clnt_fd_ctx_t *fdctx = NULL;
clnt_args_t *args = NULL;
gfs3_release_req req = {{0,},};
int ret = 0;
char parent_down = 0;
int64_t remote_fd = -1;
clnt_conf_t *conf = NULL;
clnt_fd_ctx_t *fdctx = NULL;
clnt_args_t *args = NULL;
lk_heal_state_t lk_heal_state = GF_LK_HEAL_DONE;
if (!frame || !this || !data)
goto unwind;
goto out;
args = data;
conf = this->private;
@ -2410,14 +2389,16 @@ client3_1_release (call_frame_t *frame, xlator_t *this,
{
fdctx = this_fd_del_ctx (args->fd, this);
if (fdctx != NULL) {
remote_fd = fdctx->remote_fd;
remote_fd = fdctx->remote_fd;
lk_heal_state = fdctx->lk_heal_state;
/* fdctx->remote_fd == -1 indicates a reopen attempt
in progress. Just mark ->released = 1 and let
reopen_cbk handle releasing
*/
if (remote_fd != -1)
if (remote_fd != -1 &&
lk_heal_state == GF_LK_HEAL_DONE)
list_del_init (&fdctx->sfd_pos);
fdctx->released = 1;
@ -2425,38 +2406,9 @@ client3_1_release (call_frame_t *frame, xlator_t *this,
}
pthread_mutex_unlock (&conf->lock);
if (remote_fd != -1) {
req.fd = remote_fd;
delete_granted_locks_fd (fdctx);
pthread_mutex_lock (&conf->lock);
{
parent_down = conf->parent_down;
if (!parent_down) {
rpc_clnt_ref (conf->rpc);
}
}
pthread_mutex_unlock (&conf->lock);
if (!parent_down) {
client_submit_request (this, &req, frame,
conf->fops,
GFS3_OP_RELEASE,
client3_1_release_cbk,
NULL, NULL,
0, NULL, 0, NULL,
(xdrproc_t)xdr_gfs3_release_req);
rpc_clnt_unref (conf->rpc);
}
inode_unref (fdctx->inode);
GF_FREE (fdctx);
}
unwind:
if (ret)
STACK_DESTROY (frame->root);
if (remote_fd != -1 && lk_heal_state == GF_LK_HEAL_DONE)
client_fdctx_destroy (this, fdctx);
out:
return 0;
}