summaryrefslogtreecommitdiffstats
path: root/xlators/mgmt/glusterd/src/glusterd-utils.c
diff options
context:
space:
mode:
authorAvra Sengupta <asengupt@redhat.com>2014-03-06 08:23:01 +0000
committerRajesh Joseph <rjoseph@redhat.com>2014-04-02 06:03:50 -0700
commitbff28bca0e79d67ecdec0ff7a240570b56fdcd7c (patch)
treeb5f63c1e39e34933cb5c775375f9e3c018e13163 /xlators/mgmt/glusterd/src/glusterd-utils.c
parent0ce369a0aa511e98fd71c0337181a5577b2d8a1f (diff)
glusterd/snapshot: Recording the snapshots missed in each brick.
Persisting missing snapshot info on disk as well as in memory in the following format: -------------NODE-UUID--------------:---------SNAP-UUID--------------=BRICKNUM:-------BRICKPATH--------:OPERATION:STATUS 927cb5fe-63da-48f5-82f6-e6a09ddc81c4:a17b4fe42c5a45f7a916438643edaa13= 3 :/brick/brick-dirs/brick3: 1 : 1 927cb5fe-63da-48f5-82f6-e6a09ddc81c4:a17b4fe42c5a45f7a916438643edaa13= 3 :/brick/brick-dirs/brick3: 3 : 1 927cb5fe-63da-48f5-82f6-e6a09ddc81c4:83a3cc05453b46b2a7eda4c9a9208638= 3 :/brick/brick-dirs/brick3: 1 : 1 This data will be stored on disk at /var/lib/glusterd/snaps/missed_snaps_list In memory we maintain the data as a list of glusterd_missed_snap_info in conf, the key for this list are the first two fields, i.e NODE-UUID:SNAP-UUID. For every NODE-UUID:SNAP-UUID, there can be multiple operations missed on multiple bricks. So we maintain a list of glusterd_snap_op_t for evert node of glusterd_missed_snap_info This list is maintained or updated during snapshot create, delete, and restore operations which are the only operations that if missed, are recorded in this list. During snapshot create, if a node is down, or a brick is down, we don't receive their mount point infos. snap_status of such bricks is marked as -1, and their brick details are added to this list. During snapshot delete, we check from originator node, if any other nodes, holding bricks of the said snap are down. Those are also added to the list. Also if the node is up, but the snapshot was pending for a snap brick, and its snap_status is -1, we add that to the list too. When a subsequent delete entry is processed for an already existing create entry, we just mark the create entries status as done (2), and don't add the delete entry to the list. During snapshot restore, we check from originator node, if any other nodes, holding bricks of the said snap are down. Those are also added to the list. Also if the node is up, but the snapshot was pending for a snap brick, and its snap_status is -1, we add that to the list too. Change-Id: I22578d14f81a54e13f6832966b70cd4cfdfd5b44 Signed-off-by: Avra Sengupta <asengupt@redhat.com> Reviewed-on: http://review.gluster.org/7208 Reviewed-by: Vijaikumar Mallikarjuna <vmallika@redhat.com> Reviewed-by: Rajesh Joseph <rjoseph@redhat.com> Tested-by: Rajesh Joseph <rjoseph@redhat.com>
Diffstat (limited to 'xlators/mgmt/glusterd/src/glusterd-utils.c')
-rw-r--r--xlators/mgmt/glusterd/src/glusterd-utils.c201
1 files changed, 200 insertions, 1 deletions
diff --git a/xlators/mgmt/glusterd/src/glusterd-utils.c b/xlators/mgmt/glusterd/src/glusterd-utils.c
index a969c4c84..905ac6e0c 100644
--- a/xlators/mgmt/glusterd/src/glusterd-utils.c
+++ b/xlators/mgmt/glusterd/src/glusterd-utils.c
@@ -588,9 +588,11 @@ out:
* TODO: Duplicate all members of volinfo, e.g. geo-rep sync slaves
*/
int32_t
-glusterd_snap_volinfo_restore (glusterd_volinfo_t *new_volinfo,
+glusterd_snap_volinfo_restore (dict_t *rsp_dict,
+ glusterd_volinfo_t *new_volinfo,
glusterd_volinfo_t *snap_volinfo)
{
+ int32_t brick_count = -1;
int32_t ret = -1;
xlator_t *this = NULL;
glusterd_brickinfo_t *brickinfo = NULL;
@@ -598,10 +600,12 @@ glusterd_snap_volinfo_restore (glusterd_volinfo_t *new_volinfo,
this = THIS;
GF_ASSERT (this);
+ GF_ASSERT (rsp_dict);
GF_VALIDATE_OR_GOTO (this->name, new_volinfo, out);
GF_VALIDATE_OR_GOTO (this->name, snap_volinfo, out);
+ brick_count = 0;
list_for_each_entry (brickinfo, &snap_volinfo->bricks, brick_list) {
ret = glusterd_brickinfo_new (&new_brickinfo);
if (ret) {
@@ -641,10 +645,32 @@ glusterd_snap_volinfo_restore (glusterd_volinfo_t *new_volinfo,
}
}
+ /* If a snapshot is pending for this brick then
+ * restore should also be pending
+ */
+ if (brickinfo->snap_status == -1) {
+ /* Adding missed delete to the dict */
+ ret = glusterd_add_missed_snaps_to_dict
+ (rsp_dict,
+ snap_volinfo->volname,
+ brickinfo,
+ brick_count + 1,
+ GF_SNAP_OPTION_TYPE_RESTORE);
+ if (ret) {
+ gf_log (this->name, GF_LOG_ERROR,
+ "Failed to add missed snapshot info "
+ "for %s:%s in the rsp_dict",
+ brickinfo->hostname,
+ brickinfo->path);
+ goto out;
+ }
+ }
+
list_add_tail (&new_brickinfo->brick_list,
&new_volinfo->bricks);
/* ownership of new_brickinfo is passed to new_volinfo */
new_brickinfo = NULL;
+ brick_count++;
}
/* Regenerate all volfiles */
@@ -2342,6 +2368,17 @@ glusterd_add_volume_to_dict (glusterd_volinfo_t *volinfo,
goto out;
}
+ snprintf (key, sizeof (key), "volume%d.brick%d.device_path",
+ count, i);
+ ret = dict_set_str (dict, key, brickinfo->device_path);
+ if (ret) {
+ gf_log (this->name, GF_LOG_ERROR,
+ "Failed to set snap_device for %s:%s",
+ brickinfo->hostname,
+ brickinfo->path);
+ goto out;
+ }
+
i++;
}
@@ -2861,6 +2898,7 @@ glusterd_import_new_brick (dict_t *vols, int32_t vol_count,
char key[512] = {0,};
int ret = -1;
int32_t snap_status = 0;
+ char *snap_device = NULL;
char *hostname = NULL;
char *path = NULL;
glusterd_brickinfo_t *new_brickinfo = NULL;
@@ -2896,12 +2934,21 @@ glusterd_import_new_brick (dict_t *vols, int32_t vol_count,
goto out;
}
+ snprintf (key, sizeof (key), "volume%d.brick%d.device_path",
+ vol_count, brick_count);
+ ret = dict_get_str (vols, key, &snap_device);
+ if (ret) {
+ snprintf (msg, sizeof (msg), "%s missing in payload", key);
+ goto out;
+ }
+
ret = glusterd_brickinfo_new (&new_brickinfo);
if (ret)
goto out;
strcpy (new_brickinfo->path, path);
strcpy (new_brickinfo->hostname, hostname);
+ strcpy (new_brickinfo->device_path, snap_device);
new_brickinfo->snap_status = snap_status;
//peerinfo might not be added yet
@@ -7848,6 +7895,89 @@ out:
return ret;
}
+/* Aggregate missed_snap_counts from different nodes and save it *
+ * in the req_dict of the originator node */
+int
+glusterd_snap_create_use_rsp_dict (dict_t *dst, dict_t *src)
+{
+ char *buf = NULL;
+ char *tmp_str = NULL;
+ char name_buf[PATH_MAX] = "";
+ int32_t i = -1;
+ int32_t ret = -1;
+ int32_t src_missed_snap_count = -1;
+ int32_t dst_missed_snap_count = -1;
+ xlator_t *this = NULL;
+
+ this = THIS;
+ GF_ASSERT (this);
+
+ if (!dst || !src) {
+ gf_log (this->name, GF_LOG_ERROR, "Source or Destination "
+ "dict is empty.");
+ goto out;
+ }
+
+ ret = dict_get_int32 (src, "missed_snap_count",
+ &src_missed_snap_count);
+ if (ret) {
+ gf_log (this->name, GF_LOG_DEBUG, "No missed snaps");
+ ret = 0;
+ goto out;
+ }
+
+ ret = dict_get_int32 (dst, "missed_snap_count",
+ &dst_missed_snap_count);
+ if (ret) {
+ /* Initialize dst_missed_count for the first time */
+ dst_missed_snap_count = 0;
+ }
+
+ for (i = 0; i < src_missed_snap_count; i++) {
+ snprintf (name_buf, sizeof(name_buf), "missed_snaps_%d",
+ i);
+ ret = dict_get_str (src, name_buf, &buf);
+ if (ret) {
+ gf_log (this->name, GF_LOG_ERROR,
+ "Unable to fetch %s", name_buf);
+ goto out;
+ }
+
+ snprintf (name_buf, sizeof(name_buf), "missed_snaps_%d",
+ dst_missed_snap_count);
+
+ tmp_str = gf_strdup (buf);
+ if (!tmp_str) {
+ ret = -1;
+ goto out;
+ }
+
+ ret = dict_set_dynstr (dst, name_buf, tmp_str);
+ if (ret) {
+ gf_log (this->name, GF_LOG_ERROR,
+ "Unable to set %s", name_buf);
+ goto out;
+ }
+
+ tmp_str = NULL;
+ dst_missed_snap_count++;
+ }
+
+ ret = dict_set_int32 (dst, "missed_snap_count", dst_missed_snap_count);
+ if (ret) {
+ gf_log (this->name, GF_LOG_ERROR,
+ "Unable to set dst_missed_snap_count");
+ goto out;
+ }
+
+out:
+ if (ret && tmp_str)
+ GF_FREE(tmp_str);
+
+ gf_log (this->name, GF_LOG_TRACE, "Returning %d", ret);
+ return ret;
+}
+
int
glusterd_snap_use_rsp_dict (dict_t *dst, dict_t *src)
{
@@ -7868,6 +7998,14 @@ glusterd_snap_use_rsp_dict (dict_t *dst, dict_t *src)
}
switch (snap_command) {
+ case GF_SNAP_OPTION_TYPE_CREATE:
+ case GF_SNAP_OPTION_TYPE_DELETE:
+ ret = glusterd_snap_create_use_rsp_dict (dst, src);
+ if (ret) {
+ gf_log ("", GF_LOG_ERROR, "Unable to use rsp dict");
+ goto out;
+ }
+ break;
case GF_SNAP_OPTION_TYPE_CONFIG:
ret = glusterd_snap_config_use_rsp_dict (dst, src);
if (ret) {
@@ -8695,3 +8833,64 @@ glusterd_compare_snap_vol_time(struct list_head *list1, struct list_head *list2)
return ((int)diff_time);
}
+
+int32_t
+glusterd_missed_snapinfo_new (glusterd_missed_snap_info **missed_snapinfo)
+{
+ glusterd_missed_snap_info *new_missed_snapinfo = NULL;
+ int32_t ret = -1;
+ xlator_t *this = NULL;
+
+ this = THIS;
+ GF_ASSERT (this);
+ GF_ASSERT (missed_snapinfo);
+
+ new_missed_snapinfo = GF_CALLOC (1, sizeof(*new_missed_snapinfo),
+ gf_gld_mt_missed_snapinfo_t);
+
+ if (!new_missed_snapinfo)
+ goto out;
+
+ new_missed_snapinfo->node_snap_info = NULL;
+ INIT_LIST_HEAD (&new_missed_snapinfo->missed_snaps);
+ INIT_LIST_HEAD (&new_missed_snapinfo->snap_ops);
+
+ *missed_snapinfo = new_missed_snapinfo;
+
+ ret = 0;
+
+out:
+ gf_log (this->name, GF_LOG_TRACE, "Returning %d", ret);
+ return ret;
+}
+
+int32_t
+glusterd_missed_snap_op_new (glusterd_snap_op_t **snap_op)
+{
+ glusterd_snap_op_t *new_snap_op = NULL;
+ int32_t ret = -1;
+ xlator_t *this = NULL;
+
+ this = THIS;
+ GF_ASSERT (this);
+ GF_ASSERT (snap_op);
+
+ new_snap_op = GF_CALLOC (1, sizeof(*new_snap_op),
+ gf_gld_mt_missed_snapinfo_t);
+
+ if (!new_snap_op)
+ goto out;
+
+ new_snap_op->brick_path = NULL;
+ new_snap_op->brick_num = -1;
+ new_snap_op->op = -1;
+ new_snap_op->status = -1;
+ INIT_LIST_HEAD (&new_snap_op->snap_ops_list);
+
+ *snap_op = new_snap_op;
+
+ ret = 0;
+out:
+ gf_log (this->name, GF_LOG_TRACE, "Returning %d", ret);
+ return ret;
+}