summaryrefslogtreecommitdiffstats
path: root/xlators/cluster/dht/src/dht-rebalance.c
diff options
context:
space:
mode:
authorDan Lambright <dlambrig@redhat.com>2015-10-05 19:52:02 +0000
committerDan Lambright <dlambrig@redhat.com>2015-10-21 15:44:35 -0700
commit36974c36fa4231df3f0e9428a9da6d1aa33348ab (patch)
treecbd0c604e4f4fdcd39dbeef2f2713b3d494957f6 /xlators/cluster/dht/src/dht-rebalance.c
parent802b24d4d9da09dfb1479eb54bd4a74f678bc97e (diff)
cluster/tier: add pause tier for snapshots
Snaps of tiered volumes cannot handle files undergoing migration. We implement a helper mechanism to "pause" migration. Any files undergoing migration are aborted. Clean up is done to remove sticky bits and data at the destination. Migration is restarted after snap completes. For testing an internal switch is added. It is not exposed externally. gluster volume set vol1 tier-pause [true|false] Change-Id: Ia85bbf89ac142e9b7e73fcbef98bb9da86097799 BUG: 1267950 Signed-off-by: Dan Lambright <dlambrig@redhat.com> Reviewed-on: http://review.gluster.org/12304 Reviewed-by: N Balachandran <nbalacha@redhat.com> Tested-by: NetBSD Build System <jenkins@build.gluster.org> Tested-by: Gluster Build System <jenkins@build.gluster.com>
Diffstat (limited to 'xlators/cluster/dht/src/dht-rebalance.c')
-rw-r--r--xlators/cluster/dht/src/dht-rebalance.c125
1 files changed, 122 insertions, 3 deletions
diff --git a/xlators/cluster/dht/src/dht-rebalance.c b/xlators/cluster/dht/src/dht-rebalance.c
index b3c25ba9ee2..fe648f07e8e 100644
--- a/xlators/cluster/dht/src/dht-rebalance.c
+++ b/xlators/cluster/dht/src/dht-rebalance.c
@@ -727,7 +727,7 @@ out:
static int
__dht_rebalance_migrate_data (xlator_t *from, xlator_t *to, fd_t *src, fd_t *dst,
- uint64_t ia_size, int hole_exists)
+ uint64_t ia_size, int hole_exists)
{
int ret = 0;
int count = 0;
@@ -779,6 +779,68 @@ __dht_rebalance_migrate_data (xlator_t *from, xlator_t *to, fd_t *src, fd_t *dst
return ret;
}
+static int
+__tier_migrate_data (gf_defrag_info_t *defrag, xlator_t *from, xlator_t *to, fd_t *src, fd_t *dst,
+ uint64_t ia_size, int hole_exists)
+{
+ int ret = 0;
+ int count = 0;
+ off_t offset = 0;
+ struct iovec *vector = NULL;
+ struct iobref *iobref = NULL;
+ uint64_t total = 0;
+ size_t read_size = 0;
+
+ /* if file size is '0', no need to enter this loop */
+ while (total < ia_size) {
+
+ read_size = (((ia_size - total) > DHT_REBALANCE_BLKSIZE) ?
+ DHT_REBALANCE_BLKSIZE : (ia_size - total));
+
+ ret = syncop_readv (from, src, read_size,
+ offset, 0, &vector, &count, &iobref, NULL,
+ NULL);
+ if (!ret || (ret < 0)) {
+ break;
+ }
+
+ if (hole_exists)
+ ret = dht_write_with_holes (to, dst, vector, count,
+ ret, offset, iobref);
+ else
+ ret = syncop_writev (to, dst, vector, count,
+ offset, iobref, 0, NULL, NULL);
+ if (defrag->tier_conf.request_pause) {
+ gf_msg ("tier", GF_LOG_INFO, 0,
+ DHT_MSG_TIER_PAUSED,
+ "Migrate file paused");
+ ret = -1;
+ }
+
+ if (ret < 0) {
+ break;
+ }
+ offset += ret;
+ total += ret;
+
+ GF_FREE (vector);
+ if (iobref)
+ iobref_unref (iobref);
+ iobref = NULL;
+ vector = NULL;
+ }
+ if (iobref)
+ iobref_unref (iobref);
+ GF_FREE (vector);
+
+ if (ret >= 0)
+ ret = 0;
+ else
+ ret = -1;
+
+ return ret;
+}
+
static int
__dht_rebalance_open_src_file (xlator_t *from, xlator_t *to, loc_t *loc,
@@ -1251,8 +1313,14 @@ dht_migrate_file (xlator_t *this, loc_t *loc, xlator_t *from, xlator_t *to,
/* All I/O happens in this function */
- ret = __dht_rebalance_migrate_data (from, to, src_fd, dst_fd,
- stbuf.ia_size, file_has_holes);
+ if (defrag->cmd == GF_DEFRAG_CMD_START_TIER) {
+ ret = __tier_migrate_data (defrag, from, to, src_fd, dst_fd,
+ stbuf.ia_size, file_has_holes);
+ } else {
+ ret = __dht_rebalance_migrate_data (from, to, src_fd, dst_fd,
+ stbuf.ia_size, file_has_holes);
+ }
+
if (ret) {
gf_msg (this->name, GF_LOG_ERROR, 0,
DHT_MSG_MIGRATE_FILE_FAILED,
@@ -3415,6 +3483,57 @@ out:
}
int
+gf_defrag_pause_tier (xlator_t *this, gf_defrag_info_t *defrag)
+{
+ int poll = 0;
+ int ret = 0;
+ int usec_sleep = 100000; /* 1/10th of a sec */
+ int poll_max = 15; /* 15 times = wait at most 3/2 sec */
+
+ if (defrag->defrag_status != GF_DEFRAG_STATUS_STARTED)
+ goto out;
+
+ /*
+ * Set flag requesting to pause tiering. Wait a finite time for
+ * tiering to actually stop as indicated by the "paused" boolean,
+ * before returning success or failure.
+ */
+ defrag->tier_conf.request_pause = 1;
+
+ for (poll = 0; poll < poll_max; poll++) {
+ if ((defrag->tier_conf.paused == _gf_true) ||
+ (defrag->defrag_status != GF_DEFRAG_STATUS_STARTED)) {
+ goto out;
+ }
+
+ usleep (usec_sleep);
+ }
+
+ ret = -1;
+
+out:
+
+ gf_msg (this->name, GF_LOG_DEBUG, 0,
+ DHT_MSG_TIER_PAUSED,
+ "Pause tiering ret=%d", ret);
+
+ return ret;
+}
+
+int
+gf_defrag_resume_tier (xlator_t *this, gf_defrag_info_t *defrag)
+{
+ gf_msg (this->name, GF_LOG_DEBUG, 0,
+ DHT_MSG_TIER_RESUME,
+ "Resume tiering");
+
+ defrag->tier_conf.request_pause = 0;
+ defrag->tier_conf.paused = _gf_false;
+
+ return 0;
+}
+
+int
gf_defrag_start_detach_tier (gf_defrag_info_t *defrag)
{
defrag->cmd = GF_DEFRAG_CMD_START_DETACH_TIER;