From 097e131481d25e5b1f859f4ea556b8bf56155472 Mon Sep 17 00:00:00 2001 From: Gaurav Kumar Garg Date: Wed, 25 Mar 2015 18:07:24 +0530 Subject: glusterd: cli command implementation for bitrot scrub status CLI command for bitrot scrub status will be : gluster volume bitrot scrub status Above command will show the statistics of bitrot scrubber. Upon execution of this command it will show some common scrubber tunable value of volume followed by statistics of scrubber statistics of individual nodes. sample ouput for single node: Volume name : State of scrub: Active Scrub frequency: biweekly Bitrot error log location: /var/log/glusterfs/bitd.log Scrubber error log location: /var/log/glusterfs/scrub.log ========================================================= Node name: Number of Scrubbed files: Number of Unsigned files: Last completed scrub time: Duration of last scrub: Error count: ========================================================= This is just infrastructure. list of bad file, last scrub time, error count value will be taken care by http://review.gluster.org/#/c/12503/ and http://review.gluster.org/#/c/12654/ patches. Change-Id: I3ed3c7057c9d0c894233f4079a7f185d90c202d1 BUG: 1207627 Signed-off-by: Gaurav Kumar Garg Reviewed-on: http://review.gluster.org/10231 Reviewed-by: Atin Mukherjee Tested-by: NetBSD Build System Tested-by: Gluster Build System --- cli/src/cli-cmd-parser.c | 19 +- cli/src/cli-cmd-volume.c | 2 +- cli/src/cli-rpc-ops.c | 163 ++++++++++- glusterfsd/src/glusterfsd-messages.h | 6 +- glusterfsd/src/glusterfsd-mgmt.c | 83 ++++++ libglusterfs/src/globals.h | 4 +- libglusterfs/src/glusterfs.h | 1 + rpc/rpc-lib/src/protocol-common.h | 1 + rpc/xdr/src/cli1-xdr.x | 1 + xlators/features/bit-rot/src/bitd/bit-rot.c | 18 ++ xlators/mgmt/glusterd/src/glusterd-bitrot.c | 46 +++- xlators/mgmt/glusterd/src/glusterd-op-sm.c | 93 +++++++ xlators/mgmt/glusterd/src/glusterd-rpc-ops.c | 2 + xlators/mgmt/glusterd/src/glusterd-syncop.c | 5 +- xlators/mgmt/glusterd/src/glusterd-utils.c | 394 ++++++++++++++++++++++++++- xlators/mgmt/glusterd/src/glusterd-utils.h | 5 + xlators/mgmt/glusterd/src/glusterd.h | 18 ++ 17 files changed, 846 insertions(+), 15 deletions(-) diff --git a/cli/src/cli-cmd-parser.c b/cli/src/cli-cmd-parser.c index 921648f2822..ef07e88b795 100644 --- a/cli/src/cli-cmd-parser.c +++ b/cli/src/cli-cmd-parser.c @@ -5116,7 +5116,7 @@ cli_cmd_bitrot_parse (const char **words, int wordcount, dict_t **options) "biweekly", "monthly", NULL}; char *scrub_values[] = {"pause", "resume", - NULL}; + "status", NULL}; dict_t *dict = NULL; gf_bitrot_type type = GF_BITROT_OPTION_TYPE_NONE; int32_t expiry_time = 0; @@ -5180,6 +5180,23 @@ cli_cmd_bitrot_parse (const char **words, int wordcount, dict_t **options) } } + if ((strcmp (words[3], "scrub") == 0) && + (strcmp (words[4], "status") == 0)) { + if (wordcount == 5) { + type = GF_BITROT_CMD_SCRUB_STATUS; + ret = dict_set_str (dict, "scrub-value", + (char *) words[4]); + if (ret) { + cli_out ("Failed to set dict for scrub status"); + goto out; + } + goto set_type; + } else { + ret = -1; + goto out; + } + } + if (!strcmp (w, "scrub-throttle")) { if (!words[4]) { cli_err ("Missing scrub-throttle value for bitrot " diff --git a/cli/src/cli-cmd-volume.c b/cli/src/cli-cmd-volume.c index 86274a85c1d..4e16ce1528f 100644 --- a/cli/src/cli-cmd-volume.c +++ b/cli/src/cli-cmd-volume.c @@ -2785,7 +2785,7 @@ struct cli_cmd volume_cmds[] = { "volume bitrot scrub-throttle {lazy|normal|aggressive} |\n" "volume bitrot scrub-frequency {hourly|daily|weekly|biweekly" "|monthly} |\n" - "volume bitrot scrub {pause|resume}", + "volume bitrot scrub {pause|resume|status}", cli_cmd_bitrot_cbk, "Bitrot translator specific operation. For more information about " "bitrot command type 'man gluster'" diff --git a/cli/src/cli-rpc-ops.c b/cli/src/cli-rpc-ops.c index ba044efe5d7..27d0595bd3c 100644 --- a/cli/src/cli-rpc-ops.c +++ b/cli/src/cli-rpc-ops.c @@ -10665,11 +10665,157 @@ out: } +int +gf_cli_print_bitrot_scrub_status (dict_t *dict) +{ + int i = 1; + int ret = -1; + int count = 0; + char key[256] = {0,}; + char *volname = NULL; + char *node_name = NULL; + char *scrub_freq = NULL; + char *state_scrub = NULL; + char *scrub_impact = NULL; + char *scrub_log_file = NULL; + char *bitrot_log_file = NULL; + uint64_t scrub_files = 0; + uint64_t unsigned_files = 0; + uint64_t scrub_time = 0; + uint64_t last_scrub = 0; + uint64_t error_count = 0; + + + ret = dict_get_str (dict, "volname", &volname); + if (ret) + gf_log ("cli", GF_LOG_TRACE, "failed to get volume name"); + + ret = dict_get_str (dict, "features.scrub", &state_scrub); + if (ret) + gf_log ("cli", GF_LOG_TRACE, "failed to get scrub state value"); + + ret = dict_get_str (dict, "features.scrub-throttle", &scrub_impact); + if (ret) + gf_log ("cli", GF_LOG_TRACE, "failed to get scrub impact " + "value"); + + ret = dict_get_str (dict, "features.scrub-freq", &scrub_freq); + if (ret) + gf_log ("cli", GF_LOG_TRACE, "failed to get scrub -freq value"); + + ret = dict_get_str (dict, "bitrot_log_file", &bitrot_log_file); + if (ret) + gf_log ("cli", GF_LOG_TRACE, "failed to get bitrot log file " + "location"); + + ret = dict_get_str (dict, "scrub_log_file", &scrub_log_file); + if (ret) + gf_log ("cli", GF_LOG_TRACE, "failed to get scrubber log file " + "location"); + + ret = dict_get_int32 (dict, "count", &count); + if (ret) { + gf_log ("cli", GF_LOG_ERROR, "count not get count value from" + " dictionary"); + goto out; + } + + cli_out ("\n%s: %s\n", "Volume name ", volname); + + cli_out ("%s: %s\n", "State of scrub", state_scrub); + + cli_out ("%s: %s\n", "Scrub impact", scrub_impact); + + cli_out ("%s: %s\n", "Scrub frequency", scrub_freq); + + cli_out ("%s: %s\n", "Bitrot error log location", bitrot_log_file); + + cli_out ("%s: %s\n", "Scrubber error log location", scrub_log_file); + + + for (i = 1; i <= count; i++) { + /* Reset the variables to prevent carryover of values */ + node_name = NULL; + last_scrub = 0; + scrub_time = 0; + error_count = 0; + scrub_files = 0; + unsigned_files = 0; + + memset (key, 0, 256); + snprintf (key, 256, "node-name-%d", i); + ret = dict_get_str (dict, key, &node_name); + if (ret) + gf_log ("cli", GF_LOG_TRACE, "failed to get node-name"); + + memset (key, 0, 256); + snprintf (key, 256, "scrubbed-files-%d", i); + ret = dict_get_uint64 (dict, key, &scrub_files); + if (ret) + gf_log ("cli", GF_LOG_TRACE, "failed to get scrubbed " + "files"); + + memset (key, 0, 256); + snprintf (key, 256, "unsigned-files-%d", i); + ret = dict_get_uint64 (dict, key, &unsigned_files); + if (ret) + gf_log ("cli", GF_LOG_TRACE, "failed to get unsigned " + "files"); + + memset (key, 0, 256); + snprintf (key, 256, "scrub-duration-%d", i); + ret = dict_get_uint64 (dict, key, &scrub_time); + if (ret) + gf_log ("cli", GF_LOG_TRACE, "failed to get last scrub " + "duration"); + + memset (key, 0, 256); + snprintf (key, 256, "last-scrub-time-%d", i); + ret = dict_get_uint64 (dict, key, &last_scrub); + if (ret) + gf_log ("cli", GF_LOG_TRACE, "failed to get last scrub" + " time"); + + memset (key, 0, 256); + snprintf (key, 256, "error-count-%d", i); + ret = dict_get_uint64 (dict, key, &error_count); + if (ret) + gf_log ("cli", GF_LOG_TRACE, "failed to get error " + "count"); + + cli_out ("\n%s\n", "==========================================" + "==============="); + + cli_out ("%s: %s\n", "Node name", node_name); + + cli_out ("%s: %"PRIu64 "\n", "Number of Scrubbed files", + scrub_files); + + cli_out ("%s: %"PRIu64 "\n", "Number of Unsigned files", + unsigned_files); + + cli_out ("%s: %"PRIu64 "\n", "Last completed scrub time", + scrub_time); + + cli_out ("%s: %"PRIu64 "\n", "Duration of last scrub", + last_scrub); + + cli_out ("%s: %"PRIu64 "\n", "Error count", error_count); + + } + cli_out ("%s\n", "==========================================" + "==============="); + +out: + return 0; +} + int gf_cli_bitrot_cbk (struct rpc_req *req, struct iovec *iov, int count, void *myframe) { int ret = -1; + int type = 0; gf_cli_rsp rsp = {0, }; dict_t *dict = NULL; call_frame_t *frame = NULL; @@ -10723,6 +10869,22 @@ gf_cli_bitrot_cbk (struct rpc_req *req, struct iovec *iov, gf_log ("cli", GF_LOG_DEBUG, "Received resp to bit rot command"); + ret = dict_get_int32 (dict, "type", &type); + if (ret) { + gf_log ("cli", GF_LOG_ERROR, "Failed to get command type"); + goto out; + } + + if ((type == GF_BITROT_CMD_SCRUB_STATUS) && + !(global_state->mode & GLUSTER_MODE_XML)) { + ret = gf_cli_print_bitrot_scrub_status (dict); + if (ret) { + gf_log ("cli", GF_LOG_ERROR, "Failed to print bitrot " + "scrub status"); + } + goto out; + } + xml_output: if (global_state->mode & GLUSTER_MODE_XML) { ret = cli_xml_output_vol_profile (dict, rsp.op_ret, @@ -10749,7 +10911,6 @@ out: cli_cmd_broadcast_response (ret); return ret; - } int32_t diff --git a/glusterfsd/src/glusterfsd-messages.h b/glusterfsd/src/glusterfsd-messages.h index 9c6196c1d44..caa999506e6 100644 --- a/glusterfsd/src/glusterfsd-messages.h +++ b/glusterfsd/src/glusterfsd-messages.h @@ -36,7 +36,7 @@ */ #define GLFS_COMP_BASE GLFS_MSGID_COMP_GLUSTERFSD -#define GLFS_NUM_MESSAGES 34 +#define GLFS_NUM_MESSAGES 36 #define GLFS_MSGID_END (GLFS_COMP_BASE + GLFS_NUM_MESSAGES + 1) /* Messaged with message IDs */ #define glfs_msg_start_x GLFS_COMP_BASE, "Invalid: Start of messages" @@ -104,6 +104,10 @@ "was provided" #define glusterfsd_msg_34 (GLFS_COMP_BASE + 34), "memory accounting init" \ " failed." +#define glusterfsd_msg_35 (GLFS_COMP_BASE + 35), "rpc req buffer " \ + " unserialization failed." +#define glusterfsd_msg_36 (GLFS_COMP_BASE + 36), "problem in xlator " \ + " loading." /*------------*/ #define glfs_msg_end_x GLFS_MSGID_END, "Invalid: End of messages" diff --git a/glusterfsd/src/glusterfsd-mgmt.c b/glusterfsd/src/glusterfsd-mgmt.c index 749873872d2..2e44f94b1bd 100644 --- a/glusterfsd/src/glusterfsd-mgmt.c +++ b/glusterfsd/src/glusterfsd-mgmt.c @@ -21,6 +21,7 @@ #include "rpc-clnt.h" #include "protocol-common.h" +#include "glusterfsd-messages.h" #include "glusterfs3.h" #include "portmap-xdr.h" #include "xdr-generic.h" @@ -641,6 +642,87 @@ out: return 0; } +int +glusterfs_handle_bitrot (rpcsvc_request_t *req) +{ + int32_t ret = -1; + gd1_mgmt_brick_op_req xlator_req = {0,}; + dict_t *input = NULL; + dict_t *output = NULL; + xlator_t *any = NULL; + xlator_t *this = NULL; + xlator_t *xlator = NULL; + char msg[2048] = {0,}; + char xname[1024] = {0,}; + glusterfs_ctx_t *ctx = NULL; + glusterfs_graph_t *active = NULL; + + GF_ASSERT (req); + this = THIS; + GF_ASSERT (this); + + ret = xdr_to_generic (req->msg[0], &xlator_req, + (xdrproc_t)xdr_gd1_mgmt_brick_op_req); + + if (ret < 0) { + /*failed to decode msg;*/ + req->rpc_err = GARBAGE_ARGS; + goto out; + } + + ctx = glusterfsd_ctx; + GF_ASSERT (ctx); + + active = ctx->active; + if (!active) { + req->rpc_err = GARBAGE_ARGS; + goto out; + } + + any = active->first; + + input = dict_new (); + if (!input) + goto out; + + ret = dict_unserialize (xlator_req.input.input_val, + xlator_req.input.input_len, + &input); + + if (ret < 0) { + gf_msg (this->name, GF_LOG_ERROR, 0, glusterfsd_msg_35); + goto out; + } + + /* Send scrubber request to bitrot xlator */ + snprintf (xname, sizeof (xname), "%s-bit-rot-0", xlator_req.name); + xlator = xlator_search_by_name (any, xname); + if (!xlator) { + snprintf (msg, sizeof (msg), "xlator %s is not loaded", xname); + gf_msg (this->name, GF_LOG_ERROR, 0, glusterfsd_msg_36); + goto out; + } + + output = dict_new (); + if (!output) { + ret = -1; + goto out; + } + + ret = xlator->notify (xlator, GF_EVENT_SCRUB_STATUS, input, + output); +out: + glusterfs_translator_info_response_send (req, ret, msg, output); + + if (input) + dict_unref (input); + free (xlator_req.input.input_val); /*malloced by xdr*/ + if (output) + dict_unref (output); + free (xlator_req.name); + + return 0; +} int glusterfs_handle_defrag (rpcsvc_request_t *req) @@ -1394,6 +1476,7 @@ rpcsvc_actor_t glusterfs_actors[GLUSTERD_BRICK_MAXVALUE] = { [GLUSTERD_NODE_STATUS] = {"NFS STATUS", GLUSTERD_NODE_STATUS, glusterfs_handle_node_status, NULL, 0, DRC_NA}, [GLUSTERD_VOLUME_BARRIER_OP] = {"VOLUME BARRIER OP", GLUSTERD_VOLUME_BARRIER_OP, glusterfs_handle_volume_barrier_op, NULL, 0, DRC_NA}, [GLUSTERD_BRICK_BARRIER] = {"BARRIER", GLUSTERD_BRICK_BARRIER, glusterfs_handle_barrier, NULL, 0, DRC_NA}, + [GLUSTERD_NODE_BITROT] = {"BITROT", GLUSTERD_NODE_BITROT, glusterfs_handle_bitrot, NULL, 0, DRC_NA}, }; struct rpcsvc_program glusterfs_mop_prog = { diff --git a/libglusterfs/src/globals.h b/libglusterfs/src/globals.h index 88e5f77721b..3a4ff44a335 100644 --- a/libglusterfs/src/globals.h +++ b/libglusterfs/src/globals.h @@ -38,7 +38,7 @@ */ #define GD_OP_VERSION_MIN 1 /* MIN is the fresh start op-version, mostly should not change */ -#define GD_OP_VERSION_MAX GD_OP_VERSION_3_7_6 /* MAX VERSION is the maximum +#define GD_OP_VERSION_MAX GD_OP_VERSION_3_7_7 /* MAX VERSION is the maximum count in VME table, should keep changing with introduction of newer @@ -60,6 +60,8 @@ #define GD_OP_VERSION_3_7_6 30706 /* Op-version for GlusterFS 3.7.6 */ +#define GD_OP_VERSION_3_7_7 30707 /* Op-version for GlusterFS 3.7.7 */ + #define GD_OP_VER_PERSISTENT_AFR_XATTRS GD_OP_VERSION_3_6_0 #include "xlator.h" diff --git a/libglusterfs/src/glusterfs.h b/libglusterfs/src/glusterfs.h index 4e4e5b62ab0..d7deeacbf7f 100644 --- a/libglusterfs/src/glusterfs.h +++ b/libglusterfs/src/glusterfs.h @@ -625,6 +625,7 @@ typedef enum { GF_EVENT_VOLUME_BARRIER_OP, GF_EVENT_UPCALL, GF_EVENT_SOME_CHILD_DOWN, + GF_EVENT_SCRUB_STATUS, GF_EVENT_MAXVAL, } glusterfs_event_t; diff --git a/rpc/rpc-lib/src/protocol-common.h b/rpc/rpc-lib/src/protocol-common.h index 12031738e0c..96d315c5e79 100644 --- a/rpc/rpc-lib/src/protocol-common.h +++ b/rpc/rpc-lib/src/protocol-common.h @@ -217,6 +217,7 @@ enum glusterd_brick_procnum { GLUSTERD_NODE_STATUS, GLUSTERD_VOLUME_BARRIER_OP, GLUSTERD_BRICK_BARRIER, + GLUSTERD_NODE_BITROT, GLUSTERD_BRICK_MAXVALUE, }; diff --git a/rpc/xdr/src/cli1-xdr.x b/rpc/xdr/src/cli1-xdr.x index 56f34bc2dae..231b5261f0e 100644 --- a/rpc/xdr/src/cli1-xdr.x +++ b/rpc/xdr/src/cli1-xdr.x @@ -43,6 +43,7 @@ enum gf_bitrot_type { GF_BITROT_OPTION_TYPE_SCRUB_FREQ, GF_BITROT_OPTION_TYPE_SCRUB, GF_BITROT_OPTION_TYPE_EXPIRY_TIME, + GF_BITROT_CMD_SCRUB_STATUS, GF_BITROT_OPTION_TYPE_MAX }; diff --git a/xlators/features/bit-rot/src/bitd/bit-rot.c b/xlators/features/bit-rot/src/bitd/bit-rot.c index d6ae5e2fdd2..c4ed7dee718 100644 --- a/xlators/features/bit-rot/src/bitd/bit-rot.c +++ b/xlators/features/bit-rot/src/bitd/bit-rot.c @@ -1550,9 +1550,12 @@ int notify (xlator_t *this, int32_t event, void *data, ...) { int idx = -1; + int ret = -1; xlator_t *subvol = NULL; br_child_t *child = NULL; br_private_t *priv = NULL; + dict_t *output = NULL; + va_list ap; subvol = (xlator_t *)data; priv = this->private; @@ -1619,6 +1622,21 @@ notify (xlator_t *this, int32_t event, void *data, ...) default_notify (this, event, data); break; + case GF_EVENT_SCRUB_STATUS: + gf_log (this->name, GF_LOG_INFO, "BitRot scrub status " + "called"); + va_start (ap, data); + output = va_arg (ap, dict_t *); + + /* As of now hardcoding last-scrub-time value. At the time of + * Final patch submission this option value along with other + * few option value will be calculate based on real time */ + ret = dict_set_uint64 (output, "last-scrub-time", 12); + if (ret) { + gf_log (this->name, GF_LOG_DEBUG, "Failed to set last " + "scrub time value"); + } + break; default: default_notify (this, event, data); } diff --git a/xlators/mgmt/glusterd/src/glusterd-bitrot.c b/xlators/mgmt/glusterd/src/glusterd-bitrot.c index c9cf9297bb8..6e91106c8e5 100644 --- a/xlators/mgmt/glusterd/src/glusterd-bitrot.c +++ b/xlators/mgmt/glusterd/src/glusterd-bitrot.c @@ -39,15 +39,16 @@ const char *gd_bitrot_op_list[GF_BITROT_OPTION_TYPE_MAX] = { int __glusterd_handle_bitrot (rpcsvc_request_t *req) { - int32_t ret = -1; - gf_cli_req cli_req = { {0,} }; - dict_t *dict = NULL; - glusterd_op_t cli_op = GD_OP_BITROT; - char *volname = NULL; - int32_t type = 0; + int32_t ret = -1; + gf_cli_req cli_req = { {0,} }; + dict_t *dict = NULL; + glusterd_op_t cli_op = GD_OP_BITROT; + char *volname = NULL; + char *scrub = NULL; + int32_t type = 0; char msg[2048] = {0,}; - xlator_t *this = NULL; - glusterd_conf_t *conf = NULL; + xlator_t *this = NULL; + glusterd_conf_t *conf = NULL; GF_ASSERT (req); @@ -109,6 +110,34 @@ __glusterd_handle_bitrot (rpcsvc_request_t *req) goto out; } + if (type == GF_BITROT_CMD_SCRUB_STATUS) { + /* Backward compatibility handling for scrub status command*/ + if (conf->op_version < GD_OP_VERSION_3_7_7) { + snprintf (msg, sizeof (msg), "Cannot execute command. " + "The cluster is operating at version %d. " + "Bitrot scrub status command unavailable in " + "this version", conf->op_version); + ret = -1; + goto out; + } + + ret = dict_get_str (dict, "scrub-value", &scrub); + if (ret) { + gf_msg (this->name, GF_LOG_ERROR, 0, + GD_MSG_DICT_GET_FAILED, + "Failed to get scrub value."); + ret = -1; + goto out; + } + + if (!strncmp (scrub, "status", strlen ("status"))) { + ret = glusterd_op_begin_synctask (req, + GD_OP_SCRUB_STATUS, + dict); + goto out; + } + } + ret = glusterd_op_begin_synctask (req, GD_OP_BITROT, dict); out: @@ -542,6 +571,7 @@ glusterd_op_bitrot (dict_t *dict, char **op_errstr, dict_t *rsp_dict) op_errstr); if (ret) goto out; + case GF_BITROT_CMD_SCRUB_STATUS: break; default: diff --git a/xlators/mgmt/glusterd/src/glusterd-op-sm.c b/xlators/mgmt/glusterd/src/glusterd-op-sm.c index 7db62e98b16..8bc47fc8c49 100644 --- a/xlators/mgmt/glusterd/src/glusterd-op-sm.c +++ b/xlators/mgmt/glusterd/src/glusterd-op-sm.c @@ -685,6 +685,8 @@ glusterd_node_op_build_payload (glusterd_op_t op, gd1_mgmt_brick_op_req **req, { int ret = -1; gd1_mgmt_brick_op_req *brick_req = NULL; + char xlname[1024] = {0,}; + char *volname = NULL; GF_ASSERT (op < GD_OP_MAX); GF_ASSERT (op > GD_OP_NONE); @@ -713,6 +715,20 @@ glusterd_node_op_build_payload (glusterd_op_t op, gd1_mgmt_brick_op_req **req, break; + case GD_OP_SCRUB_STATUS: + brick_req = GF_CALLOC (1, sizeof(*brick_req), + gf_gld_mt_mop_brick_req_t); + if (!brick_req) + goto out; + + brick_req->op = GLUSTERD_NODE_BITROT; + + ret = dict_get_str (dict, "volname", &volname); + if (ret) + goto out; + + brick_req->name = gf_strdup (volname); + break; default: goto out; } @@ -4035,6 +4051,7 @@ glusterd_op_build_payload (dict_t **req, char **op_errstr, dict_t *op_ctx) case GD_OP_DEFRAG_BRICK_VOLUME: case GD_OP_BARRIER: case GD_OP_BITROT: + case GD_OP_SCRUB_STATUS: { do_common = _gf_true; } @@ -4616,6 +4633,7 @@ glusterd_op_modify_op_ctx (glusterd_op_t op, void *ctx) * same */ case GD_OP_DEFRAG_BRICK_VOLUME: + case GD_OP_SCRUB_STATUS: ret = dict_get_int32 (op_ctx, "count", &count); if (ret) { gf_msg_debug (this->name, 0, @@ -4663,6 +4681,13 @@ glusterd_op_modify_op_ctx (glusterd_op_t op, void *ctx) GD_MSG_CONVERSION_FAILED, "Failed uuid to hostname conversion"); + /* Since Both rebalance and bitrot scrub status are going to + * use same code path till here, we should break in case + * of scrub status */ + if (op == GD_OP_SCRUB_STATUS) { + break; + } + ret = glusterd_op_check_peer_defrag_status (op_ctx, count); if (ret) gf_msg (this->name, GF_LOG_ERROR, 0, @@ -5258,6 +5283,7 @@ glusterd_need_brick_op (glusterd_op_t op) case GD_OP_STATUS_VOLUME: case GD_OP_DEFRAG_BRICK_VOLUME: case GD_OP_HEAL_VOLUME: + case GD_OP_SCRUB_STATUS: ret = _gf_true; break; default: @@ -5520,6 +5546,7 @@ glusterd_op_stage_validate (glusterd_op_t op, dict_t *dict, char **op_errstr, break; case GD_OP_BITROT: + case GD_OP_SCRUB_STATUS: ret = glusterd_op_stage_bitrot (dict, op_errstr, rsp_dict); break; @@ -5644,6 +5671,7 @@ glusterd_op_commit_perform (glusterd_op_t op, dict_t *dict, char **op_errstr, break; case GD_OP_BITROT: + case GD_OP_SCRUB_STATUS: ret = glusterd_op_bitrot (dict, op_errstr, rsp_dict); break; @@ -6808,6 +6836,68 @@ out: return ret; } +static int +glusterd_bricks_select_scrub (dict_t *dict, char **op_errstr, + struct cds_list_head *selected) +{ + int ret = -1; + char *volname = NULL; + char msg[2048] = {0,}; + xlator_t *this = NULL; + glusterd_conf_t *priv = NULL; + glusterd_volinfo_t *volinfo = NULL; + glusterd_brickinfo_t *brickinfo = NULL; + glusterd_pending_node_t *pending_node = NULL; + + this = THIS; + priv = this->private; + GF_ASSERT (this); + GF_ASSERT (priv); + + GF_ASSERT (dict); + + ret = dict_get_str (dict, "volname", &volname); + if (ret) { + gf_msg (this->name, GF_LOG_ERROR, 0, + GD_MSG_DICT_GET_FAILED, "Unable to get" + " volname"); + goto out; + } + + ret = glusterd_volinfo_find (volname, &volinfo); + if (ret) { + snprintf (msg, sizeof (msg), "Volume %s does not exist", + volname); + + *op_errstr = gf_strdup (msg); + gf_msg (this->name, GF_LOG_ERROR, EINVAL, + GD_MSG_VOL_NOT_FOUND, "%s", msg); + goto out; + } + + if (!priv->scrub_svc.online) { + ret = 0; + snprintf (msg, sizeof (msg), "Scrubber daemon is not running"); + + gf_msg_debug (this->name, 0, "%s", msg); + goto out; + } + + pending_node = GF_CALLOC (1, sizeof (*pending_node), + gf_gld_mt_pending_node_t); + if (!pending_node) { + ret = -1; + goto out; + } + + pending_node->node = &(priv->scrub_svc); + pending_node->type = GD_NODE_SCRUB; + cds_list_add_tail (&pending_node->list, selected); + pending_node = NULL; +out: + gf_msg_debug (this->name, 0, "Returning %d", ret); + return ret; +} /* Select the bricks to send the barrier request to. * This selects the bricks of the given volume which are present on this peer * and are running @@ -7021,6 +7111,9 @@ glusterd_op_bricks_select (glusterd_op_t op, dict_t *dict, char **op_errstr, case GD_OP_SNAP: ret = glusterd_bricks_select_snap (dict, op_errstr, selected); break; + case GD_OP_SCRUB_STATUS: + ret = glusterd_bricks_select_scrub (dict, op_errstr, selected); + break; default: break; } diff --git a/xlators/mgmt/glusterd/src/glusterd-rpc-ops.c b/xlators/mgmt/glusterd/src/glusterd-rpc-ops.c index 8a826521f56..fd51255f65b 100644 --- a/xlators/mgmt/glusterd/src/glusterd-rpc-ops.c +++ b/xlators/mgmt/glusterd/src/glusterd-rpc-ops.c @@ -142,6 +142,7 @@ glusterd_op_send_cli_response (glusterd_op_t op, int32_t op_ret, case GD_OP_SNAP: case GD_OP_BARRIER: case GD_OP_BITROT: + case GD_OP_SCRUB_STATUS: { /*nothing specific to be done*/ break; @@ -2234,6 +2235,7 @@ glusterd_brick_op (call_frame_t *frame, xlator_t *this, if ((pending_node->type == GD_NODE_NFS) || (pending_node->type == GD_NODE_QUOTAD) || (pending_node->type == GD_NODE_SNAPD) || + (pending_node->type == GD_NODE_SCRUB) || ((pending_node->type == GD_NODE_SHD) && (req_ctx->op == GD_OP_STATUS_VOLUME))) ret = glusterd_node_op_build_payload diff --git a/xlators/mgmt/glusterd/src/glusterd-syncop.c b/xlators/mgmt/glusterd/src/glusterd-syncop.c index 064077278bd..a0b856160c9 100644 --- a/xlators/mgmt/glusterd/src/glusterd-syncop.c +++ b/xlators/mgmt/glusterd/src/glusterd-syncop.c @@ -304,6 +304,9 @@ glusterd_syncop_aggr_rsp_dict (glusterd_op_t op, dict_t *aggr, dict_t *rsp) goto out; break; + case GD_OP_SCRUB_STATUS: + ret = glusterd_volume_bitrot_scrub_use_rsp_dict (aggr, rsp); + break; default: break; } @@ -932,7 +935,7 @@ gd_syncop_mgmt_brick_op (struct rpc_clnt *rpc, glusterd_pending_node_t *pnode, args.op_errno = ENOTCONN; if ((pnode->type == GD_NODE_NFS) || - (pnode->type == GD_NODE_QUOTAD) || + (pnode->type == GD_NODE_QUOTAD) || (pnode->type == GD_NODE_SCRUB) || ((pnode->type == GD_NODE_SHD) && (op == GD_OP_STATUS_VOLUME))) { ret = glusterd_node_op_build_payload (op, &req, dict_out); diff --git a/xlators/mgmt/glusterd/src/glusterd-utils.c b/xlators/mgmt/glusterd/src/glusterd-utils.c index 6f24e9274ff..c31c394f661 100644 --- a/xlators/mgmt/glusterd/src/glusterd-utils.c +++ b/xlators/mgmt/glusterd/src/glusterd-utils.c @@ -4222,7 +4222,8 @@ glusterd_pending_node_get_rpc (glusterd_pending_node_t *pending_node) } else if (pending_node->type == GD_NODE_SHD || pending_node->type == GD_NODE_NFS || - pending_node->type == GD_NODE_QUOTAD) { + pending_node->type == GD_NODE_QUOTAD || + pending_node->type == GD_NODE_SCRUB) { svc = pending_node->node; rpc = svc->conn.rpc; } else if (pending_node->type == GD_NODE_REBALANCE) { @@ -8240,6 +8241,393 @@ out: return ret; } +int +glusterd_volume_bitrot_scrub_use_rsp_dict (dict_t *aggr, dict_t *rsp_dict) +{ + int ret = -1; + uint64_t value = 0; + int32_t count = 0; + char key[256] = {0,}; + uint64_t error_count = 0; + uint64_t scrubbed_files = 0; + uint64_t unsigned_files = 0; + uint64_t scrub_duration = 0; + uint64_t last_scrub_time = 0; + char *volname = NULL; + char *node_uuid = NULL; + char *node_uuid_str = NULL; + char *bitd_log = NULL; + char *scrub_log = NULL; + char *scrub_freq = NULL; + char *scrub_state = NULL; + char *scrub_impact = NULL; + xlator_t *this = NULL; + glusterd_conf_t *priv = NULL; + glusterd_volinfo_t *volinfo = NULL; + int src_count = 0; + int dst_count = 0; + + this = THIS; + GF_ASSERT (this); + + priv = this->private; + GF_ASSERT (priv); + + ret = dict_get_str (aggr, "volname", &volname); + if (ret) { + gf_msg (this->name, GF_LOG_ERROR, 0, GD_MSG_DICT_GET_FAILED, + "Unable to get volume name"); + goto out; + } + + ret = glusterd_volinfo_find (volname, &volinfo); + if (ret) { + gf_msg (THIS->name, GF_LOG_ERROR, 0, GD_MSG_VOL_NOT_FOUND, + "Unable to find volinfo for volume: %s", volname); + goto out; + } + + ret = dict_get_int32 (aggr, "count", &dst_count); + + ret = dict_get_int32 (rsp_dict, "count", &src_count); + if (ret) { + gf_msg (this->name, GF_LOG_ERROR, 0, GD_MSG_DICT_GET_FAILED, + "failed to get count value"); + ret = 0; + goto out; + } + + ret = dict_set_int32 (aggr, "count", src_count+dst_count); + if (ret) + gf_msg (this->name, GF_LOG_ERROR, 0, GD_MSG_DICT_SET_FAILED, + "Failed to set count in dictonary"); + + snprintf (key, 256, "node-uuid-%d", src_count); + ret = dict_get_str (rsp_dict, key, &node_uuid); + if (!ret) { + node_uuid_str = gf_strdup (node_uuid); + + memset (key, 0, 256); + snprintf (key, 256, "node-uuid-%d", src_count+dst_count); + ret = dict_set_dynstr (aggr, key, node_uuid_str); + if (ret) { + gf_msg_debug (this->name, 0, "failed to set node-uuid"); + } + } + + memset (key, 0, 256); + snprintf (key, 256, "scrubbed-files-%d", src_count); + ret = dict_get_uint64 (rsp_dict, key, &value); + if (!ret) { + memset (key, 0, 256); + snprintf (key, 256, "scrubbed-files-%d", src_count+dst_count); + ret = dict_set_uint64 (aggr, key, value); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "scrubbed-file value"); + } + } + + memset (key, 0, 256); + snprintf (key, 256, "unsigned-files-%d", src_count); + ret = dict_get_uint64 (rsp_dict, key, &value); + if (!ret) { + memset (key, 0, 256); + snprintf (key, 256, "unsigned-files-%d", src_count+dst_count); + ret = dict_set_uint64 (aggr, key, value); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "unsigned-file value"); + } + } + + memset (key, 0, 256); + snprintf (key, 256, "last-scrub-time-%d", src_count); + ret = dict_get_uint64 (rsp_dict, key, &value); + if (!ret) { + memset (key, 0, 256); + snprintf (key, 256, "last-scrub-time-%d", src_count+dst_count); + ret = dict_set_uint64 (aggr, key, value); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "last scrub time value"); + } + } + + memset (key, 0, 256); + snprintf (key, 256, "scrub-duration-%d", src_count); + ret = dict_get_uint64 (rsp_dict, key, &value); + if (!ret) { + memset (key, 0, 256); + snprintf (key, 256, "scrub-duration-%d", src_count+dst_count); + ret = dict_set_uint64 (aggr, key, value); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "scrubbed-duration value"); + } + } + + memset (key, 0, 256); + snprintf (key, 256, "error-count-%d", src_count); + ret = dict_get_uint64 (rsp_dict, key, &value); + if (!ret) { + memset (key, 0, 256); + snprintf (key, 256, "error-count-%d", src_count+dst_count); + ret = dict_set_uint64 (aggr, key, value); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set error " + "count value"); + } + } + + ret = dict_get_str (rsp_dict, "bitrot_log_file", &bitd_log); + if (!ret) { + ret = dict_set_str (aggr, "bitrot_log_file", bitd_log); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "bitrot log file location"); + goto out; + } + } + + ret = dict_get_str (rsp_dict, "scrub_log_file", &scrub_log); + if (!ret) { + ret = dict_set_str (aggr, "scrub_log_file", scrub_log); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "scrubber log file location"); + goto out; + } + } + + ret = dict_get_str (rsp_dict, "features.scrub-freq", &scrub_freq); + if (!ret) { + ret = dict_set_str (aggr, "features.scrub-freq", scrub_freq); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "scrub-frequency value to dictionary"); + goto out; + } + } + + ret = dict_get_str (rsp_dict, "features.scrub-throttle", &scrub_impact); + if (!ret) { + ret = dict_set_str (aggr, "features.scrub-throttle", + scrub_impact); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "scrub-throttle value to dictionary"); + goto out; + } + } + + ret = dict_get_str (rsp_dict, "features.scrub", &scrub_state); + if (!ret) { + ret = dict_set_str (aggr, "features.scrub", scrub_state); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "scrub state value to dictionary"); + goto out; + } + } + + ret = 0; +out: + return ret; +} + +int +glusterd_bitrot_volume_node_rsp (dict_t *aggr, dict_t *rsp_dict) +{ + int ret = -1; + uint64_t value = 0; + int32_t count = 0; + int32_t index = 0; + char key[256] = {0,}; + char buf[1024] = {0,}; + uint64_t error_count = 0; + int32_t i = 0; + uint64_t scrubbed_files = 0; + uint64_t unsigned_files = 0; + uint64_t scrub_duration = 0; + uint64_t last_scrub_time = 0; + char *volname = NULL; + char *node_str = NULL; + char *scrub_freq = NULL; + char *scrub_state = NULL; + char *scrub_impact = NULL; + xlator_t *this = NULL; + glusterd_conf_t *priv = NULL; + glusterd_volinfo_t *volinfo = NULL; + + this = THIS; + GF_ASSERT (this); + + priv = this->private; + GF_ASSERT (priv); + + ret = dict_set_str (aggr, "bitrot_log_file", + (priv->bitd_svc.proc.logfile)); + if (ret) { + gf_msg (this->name, GF_LOG_ERROR, 0, GD_MSG_DICT_SET_FAILED, + "Failed to set bitrot log file location"); + goto out; + } + + ret = dict_set_str (aggr, "scrub_log_file", + (priv->scrub_svc.proc.logfile)); + if (ret) { + gf_msg (this->name, GF_LOG_ERROR, 0, GD_MSG_DICT_SET_FAILED, + "Failed to set scrubber log file location"); + goto out; + } + + ret = dict_get_str (aggr, "volname", &volname); + if (ret) { + gf_msg (this->name, GF_LOG_ERROR, 0, GD_MSG_DICT_GET_FAILED, + "Unable to get volume name"); + goto out; + } + + ret = glusterd_volinfo_find (volname, &volinfo); + if (ret) { + gf_msg (this->name, GF_LOG_ERROR, 0, GD_MSG_VOL_NOT_FOUND, + "Unable to find volinfo for volume: %s", volname); + goto out; + } + + ret = dict_get_int32 (aggr, "count", &i); + i++; + + ret = dict_set_int32 (aggr, "count", i); + if (ret) + gf_msg (this->name, GF_LOG_ERROR, 0, GD_MSG_DICT_SET_FAILED, + "Failed to set count"); + + snprintf (buf, 1024, "%s", uuid_utoa (MY_UUID)); + + snprintf (key, 256, "node-uuid-%d", i); + ret = dict_set_dynstr_with_alloc (aggr, key, buf); + if (ret) + gf_msg (this->name, GF_LOG_ERROR, 0, GD_MSG_DICT_SET_FAILED, + "failed to set node-uuid"); + + ret = dict_get_str (volinfo->dict, "features.scrub-freq", &scrub_freq); + if (!ret) { + ret = dict_set_str (aggr, "features.scrub-freq", scrub_freq); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "scrub-frequency value to dictionary"); + } + } else { + /* By Default scrub-frequency is bi-weekly. So when user + * enable bitrot then scrub-frequency value will not be + * present in volinfo->dict. Setting by-default value of + * scrub-frequency explicitly for presenting it to scrub + * status. + */ + ret = dict_set_dynstr_with_alloc (aggr, "features.scrub-freq", + "biweekly"); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "scrub-frequency value to dictionary"); + } + } + + ret = dict_get_str (volinfo->dict, "features.scrub-throttle", + &scrub_impact); + if (!ret) { + ret = dict_set_str (aggr, "features.scrub-throttle", + scrub_impact); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "scrub-throttle value to dictionary"); + } + } else { + /* By Default scrub-throttle is lazy. So when user + * enable bitrot then scrub-throttle value will not be + * present in volinfo->dict. Setting by-default value of + * scrub-throttle explicitly for presenting it to + * scrub status. + */ + ret = dict_set_dynstr_with_alloc (aggr, + "features.scrub-throttle", + "lazy"); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "scrub-throttle value to dictionary"); + } + } + + ret = dict_get_str (volinfo->dict, "features.scrub", &scrub_state); + if (!ret) { + ret = dict_set_str (aggr, "features.scrub", scrub_state); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "scrub state value to dictionary"); + } + } + + ret = dict_get_uint64 (rsp_dict, "scrubbed-files", &value); + if (!ret) { + memset (key, 0, 256); + snprintf (key, 256, "scrubbed-files-%d", i); + ret = dict_set_uint64 (aggr, key, value); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "scrubbed-file value"); + } + } + + ret = dict_get_uint64 (rsp_dict, "unsigned-files", &value); + if (!ret) { + memset (key, 0, 256); + snprintf (key, 256, "unsigned-files-%d", i); + ret = dict_set_uint64 (aggr, key, value); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "unsigned-file value"); + } + } + + ret = dict_get_uint64 (rsp_dict, "last-scrub-time", &value); + if (!ret) { + memset (key, 0, 256); + snprintf (key, 256, "last-scrub-time-%d", i); + ret = dict_set_uint64 (aggr, key, value); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "last scrub time value"); + } + } + + ret = dict_get_uint64 (rsp_dict, "scrub-duration", &value); + if (!ret) { + memset (key, 0, 256); + snprintf (key, 256, "scrub-duration-%d", i); + ret = dict_set_uint64 (aggr, key, value); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set " + "scrubbed-duration value"); + } + } + + ret = dict_get_uint64 (rsp_dict, "error-count", &value); + if (!ret) { + memset (key, 0, 256); + snprintf (key, 256, "error-count-%d", i); + ret = dict_set_uint64 (aggr, key, value); + if (ret) { + gf_msg_debug (this->name, 0, "Failed to set error " + "count value"); + } + } + + ret = 0; +out: + return ret; +} + int glusterd_volume_rebalance_use_rsp_dict (dict_t *aggr, dict_t *rsp_dict) { @@ -9139,6 +9527,10 @@ glusterd_handle_node_rsp (dict_t *req_dict, void *pending_entry, ret = glusterd_heal_volume_brick_rsp (req_dict, rsp_dict, op_ctx, op_errstr); break; + case GD_OP_SCRUB_STATUS: + ret = glusterd_bitrot_volume_node_rsp (op_ctx, rsp_dict); + + break; default: break; } diff --git a/xlators/mgmt/glusterd/src/glusterd-utils.h b/xlators/mgmt/glusterd/src/glusterd-utils.h index b7302c8cb91..f3895db408b 100644 --- a/xlators/mgmt/glusterd/src/glusterd-utils.h +++ b/xlators/mgmt/glusterd/src/glusterd-utils.h @@ -436,6 +436,11 @@ int32_t glusterd_handle_node_rsp (dict_t *req_ctx, void *pending_entry, glusterd_op_t op, dict_t *rsp_dict, dict_t *op_ctx, char **op_errstr, gd_node_type type); +int +glusterd_volume_bitrot_scrub_use_rsp_dict (dict_t *aggr, dict_t *rsp_dict); + +int +glusterd_volume_heal_use_rsp_dict (dict_t *aggr, dict_t *rsp_dict); int32_t glusterd_check_if_quota_trans_enabled (glusterd_volinfo_t *volinfo); diff --git a/xlators/mgmt/glusterd/src/glusterd.h b/xlators/mgmt/glusterd/src/glusterd.h index 124c6cf0f9f..817112e9aca 100644 --- a/xlators/mgmt/glusterd/src/glusterd.h +++ b/xlators/mgmt/glusterd/src/glusterd.h @@ -116,6 +116,7 @@ typedef enum glusterd_op_ { GD_OP_BITROT, GD_OP_DETACH_TIER, GD_OP_TIER_MIGRATE, + GD_OP_SCRUB_STATUS, GD_OP_MAX, } glusterd_op_t; @@ -276,6 +277,20 @@ typedef struct _auth auth_t; #define CAPS_OFFLOAD_SNAPSHOT 0x00000008 #define CAPS_OFFLOAD_ZERO 0x00000020 +struct glusterd_bitrot_scrub_ { + char *scrub_state; + char *scrub_impact; + char *scrub_freq; + uint64_t scrubbed_files; + uint64_t unsigned_files; + uint64_t last_scrub_time; + uint64_t scrub_duration; + uint64_t error_count; +}; + +typedef struct glusterd_bitrot_scrub_ glusterd_bitrot_scrub_t; + + struct glusterd_rebalance_ { gf_defrag_status_t defrag_status; uint64_t rebalance_files; @@ -382,6 +397,9 @@ struct glusterd_volinfo_ { /* Replace brick status */ glusterd_replace_brick_t rep_brick; + /* Bitrot scrub status*/ + glusterd_bitrot_scrub_t bitrot_scrub; + int version; uint32_t quota_conf_version; uint32_t cksum; -- cgit