OpenZFS 9102 - zfs should be able to initialize storage devices

PROBLEM ======== The first access to a block incurs a performance penalty on some platforms (e.g. AWS's EBS, VMware VMDKs). Therefore we recommend that volumes are "thick provisioned", where supported by the platform (VMware). This can create a large delay in getting a new virtual machines up and running (or adding storage to an existing Engine). If the thick provision step is omitted, write performance will be suboptimal until all blocks on the LUN have been written. SOLUTION ========= This feature introduces a way to 'initialize' the disks at install or in the background to make sure we don't incur this first read penalty. When an entire LUN is added to ZFS, we make all space available immediately, and allow ZFS to find unallocated space and zero it out. This works with concurrent writes to arbitrary offsets, ensuring that we don't zero out something that has been (or is in the middle of being) written. This scheme can also be applied to existing pools (affecting only free regions on the vdev). Detailed design: - new subcommand:zpool initialize [-cs] <pool> [<vdev> ...] - start, suspend, or cancel initialization - Creates new open-context thread for each vdev - Thread iterates through all metaslabs in this vdev - Each metaslab: - select a metaslab - load the metaslab - mark the metaslab as being zeroed - walk all free ranges within that metaslab and translate them to ranges on the leaf vdev - issue a "zeroing" I/O on the leaf vdev that corresponds to a free range on the metaslab we're working on - continue until all free ranges for this metaslab have been "zeroed" - reset/unmark the metaslab being zeroed - if more metaslabs exist, then repeat above tasks. - if no more metaslabs, then we're done. - progress for the initialization is stored on-disk in the vdev’s leaf zap object. The following information is stored: - the last offset that has been initialized - the state of the initialization process (i.e. active, suspended, or canceled) - the start time for the initialization - progress is reported via the zpool status command and shows information for each of the vdevs that are initializing Porting notes: - Added zfs_initialize_value module parameter to set the pattern written by "zpool initialize". - Added zfs_vdev_{initializing,removal}_{min,max}_active module options. Authored by: George Wilson <george.wilson@delphix.com> Reviewed by: John Wren Kennedy <john.kennedy@delphix.com> Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Reviewed by: Prakash Surya <prakash.surya@delphix.com> Reviewed by: loli10K <ezomori.nozomu@gmail.com> Reviewed by: Brian Behlendorf <behlendorf1@llnl.gov> Approved by: Richard Lowe <richlowe@richlowe.net> Signed-off-by: Tim Chase <tim@chase2k.com> Ported-by: Tim Chase <tim@chase2k.com> OpenZFS-issue: https://www.illumos.org/issues/9102 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/c3963210eb Closes #8230
2026-05-22 02:27:36 +03:00 · 2018-12-19 07:54:59 -07:00
parent c87db59196
commit 619f097693
54 changed files with 2743 additions and 30 deletions
@@ -97,6 +97,7 @@ static int zpool_do_detach(int, char **);
 static int zpool_do_replace(int, char **);
 static int zpool_do_split(int, char **);

+static int zpool_do_initialize(int, char **);
 static int zpool_do_scrub(int, char **);
 static int zpool_do_resilver(int, char **);

@@ -150,6 +151,7 @@ typedef enum {
 	HELP_ONLINE,
 	HELP_REPLACE,
 	HELP_REMOVE,
+	HELP_INITIALIZE,
 	HELP_SCRUB,
 	HELP_RESILVER,
 	HELP_STATUS,
@@ -278,6 +280,7 @@ static zpool_command_t command_table[] = {
 	{ "replace",	zpool_do_replace,	HELP_REPLACE		},
 	{ "split",	zpool_do_split,		HELP_SPLIT		},
 	{ NULL },
+	{ "initialize",	zpool_do_initialize,	HELP_INITIALIZE		},
 	{ "scrub",	zpool_do_scrub,		HELP_SCRUB		},
 	{ "resilver",	zpool_do_resilver,	HELP_RESILVER		},
 	{ NULL },
@@ -360,6 +363,8 @@ get_usage(zpool_help_t idx)
 		return (gettext("\tremove [-nps] <pool> <device> ...\n"));
 	case HELP_REOPEN:
 		return (gettext("\treopen [-n] <pool>\n"));
+	case HELP_INITIALIZE:
+		return (gettext("\tinitialize [-cs] <pool> [<device> ...]\n"));
 	case HELP_SCRUB:
 		return (gettext("\tscrub [-s | -p] <pool> ...\n"));
 	case HELP_RESILVER:
@@ -393,6 +398,27 @@ get_usage(zpool_help_t idx)
 	/* NOTREACHED */
 }

+static void
+zpool_collect_leaves(zpool_handle_t *zhp, nvlist_t *nvroot, nvlist_t *res)
+{
+	uint_t children = 0;
+	nvlist_t **child;
+	uint_t i;
+
+	(void) nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
+	    &child, &children);
+
+	if (children == 0) {
+		char *path = zpool_vdev_name(g_zfs, zhp, nvroot, B_FALSE);
+		fnvlist_add_boolean(res, path);
+		free(path);
+		return;
+	}
+
+	for (i = 0; i < children; i++) {
+		zpool_collect_leaves(zhp, child[i], res);
+	}
+}

 /*
 * Callback routine that will print out a pool property value.
@@ -479,6 +505,97 @@ usage(boolean_t requested)
 	exit(requested ? 0 : 2);
 }

+/*
+ * zpool initialize [-cs] <pool> [<vdev> ...]
+ * Initialize all unused blocks in the specified vdevs, or all vdevs in the pool
+ * if none specified.
+ *
+ *	-c	Cancel. Ends active initializing.
+ *	-s	Suspend. Initializing can then be restarted with no flags.
+ */
+int
+zpool_do_initialize(int argc, char **argv)
+{
+	int c;
+	char *poolname;
+	zpool_handle_t *zhp;
+	nvlist_t *vdevs;
+	int err = 0;
+
+	struct option long_options[] = {
+		{"cancel",	no_argument,		NULL, 'c'},
+		{"suspend",	no_argument,		NULL, 's'},
+		{0, 0, 0, 0}
+	};
+
+	pool_initialize_func_t cmd_type = POOL_INITIALIZE_DO;
+	while ((c = getopt_long(argc, argv, "cs", long_options, NULL)) != -1) {
+		switch (c) {
+		case 'c':
+			if (cmd_type != POOL_INITIALIZE_DO) {
+				(void) fprintf(stderr, gettext("-c cannot be "
+				    "combined with other options\n"));
+				usage(B_FALSE);
+			}
+			cmd_type = POOL_INITIALIZE_CANCEL;
+			break;
+		case 's':
+			if (cmd_type != POOL_INITIALIZE_DO) {
+				(void) fprintf(stderr, gettext("-s cannot be "
+				    "combined with other options\n"));
+				usage(B_FALSE);
+			}
+			cmd_type = POOL_INITIALIZE_SUSPEND;
+			break;
+		case '?':
+			if (optopt != 0) {
+				(void) fprintf(stderr,
+				    gettext("invalid option '%c'\n"), optopt);
+			} else {
+				(void) fprintf(stderr,
+				    gettext("invalid option '%s'\n"),
+				    argv[optind - 1]);
+			}
+			usage(B_FALSE);
+		}
+	}
+
+	argc -= optind;
+	argv += optind;
+
+	if (argc < 1) {
+		(void) fprintf(stderr, gettext("missing pool name argument\n"));
+		usage(B_FALSE);
+		return (-1);
+	}
+
+	poolname = argv[0];
+	zhp = zpool_open(g_zfs, poolname);
+	if (zhp == NULL)
+		return (-1);
+
+	vdevs = fnvlist_alloc();
+	if (argc == 1) {
+		/* no individual leaf vdevs specified, so add them all */
+		nvlist_t *config = zpool_get_config(zhp, NULL);
+		nvlist_t *nvroot = fnvlist_lookup_nvlist(config,
+		    ZPOOL_CONFIG_VDEV_TREE);
+		zpool_collect_leaves(zhp, nvroot, vdevs);
+	} else {
+		int i;
+		for (i = 1; i < argc; i++) {
+			fnvlist_add_boolean(vdevs, argv[i]);
+		}
+	}
+
+	err = zpool_initialize(zhp, cmd_type, vdevs);
+
+	fnvlist_free(vdevs);
+	zpool_close(zhp);
+
+	return (err);
+}
+
 /*
 * print a pool vdev config for dry runs
 */
@@ -1923,6 +2040,43 @@ print_status_config(zpool_handle_t *zhp, status_cbdata_t *cb, const char *name,
 		}
 	}

+	if ((vs->vs_initialize_state == VDEV_INITIALIZE_ACTIVE ||
+	    vs->vs_initialize_state == VDEV_INITIALIZE_SUSPENDED ||
+	    vs->vs_initialize_state == VDEV_INITIALIZE_COMPLETE) &&
+	    !vs->vs_scan_removing) {
+		char zbuf[1024];
+		char tbuf[256];
+		struct tm zaction_ts;
+
+		time_t t = vs->vs_initialize_action_time;
+		int initialize_pct = 100;
+		if (vs->vs_initialize_state != VDEV_INITIALIZE_COMPLETE) {
+			initialize_pct = (vs->vs_initialize_bytes_done * 100 /
+			    (vs->vs_initialize_bytes_est + 1));
+		}
+
+		(void) localtime_r(&t, &zaction_ts);
+		(void) strftime(tbuf, sizeof (tbuf), "%c", &zaction_ts);
+
+		switch (vs->vs_initialize_state) {
+		case VDEV_INITIALIZE_SUSPENDED:
+			(void) snprintf(zbuf, sizeof (zbuf),
+			    ", suspended, started at %s", tbuf);
+			break;
+		case VDEV_INITIALIZE_ACTIVE:
+			(void) snprintf(zbuf, sizeof (zbuf),
+			    ", started at %s", tbuf);
+			break;
+		case VDEV_INITIALIZE_COMPLETE:
+			(void) snprintf(zbuf, sizeof (zbuf),
+			    ", completed at %s", tbuf);
+			break;
+		}
+
+		(void) printf(gettext("  (%d%% initialized%s)"),
+		    initialize_pct, zbuf);
+	}
+
 	(void) printf("\n");

 	for (c = 0; c < children; c++) {
@@ -106,6 +106,7 @@
 #include <sys/zil_impl.h>
 #include <sys/vdev_impl.h>
 #include <sys/vdev_file.h>
+#include <sys/vdev_initialize.h>
 #include <sys/spa_impl.h>
 #include <sys/metaslab_impl.h>
 #include <sys/dsl_prop.h>
@@ -374,6 +375,7 @@ ztest_func_t ztest_spa_upgrade;
 ztest_func_t ztest_device_removal;
 ztest_func_t ztest_remap_blocks;
 ztest_func_t ztest_spa_checkpoint_create_discard;
+ztest_func_t ztest_initialize;
 ztest_func_t ztest_fletcher;
 ztest_func_t ztest_fletcher_incr;
 ztest_func_t ztest_verify_dnode_bt;
@@ -427,6 +429,7 @@ ztest_info_t ztest_info[] = {
 	ZTI_INIT(ztest_device_removal, 1, &zopt_sometimes),
 	ZTI_INIT(ztest_remap_blocks, 1, &zopt_sometimes),
 	ZTI_INIT(ztest_spa_checkpoint_create_discard, 1, &zopt_rarely),
+	ZTI_INIT(ztest_initialize, 1, &zopt_sometimes),
 	ZTI_INIT(ztest_fletcher, 1, &zopt_rarely),
 	ZTI_INIT(ztest_fletcher_incr, 1, &zopt_rarely),
 	ZTI_INIT(ztest_verify_dnode_bt, 1, &zopt_sometimes),
@@ -6343,6 +6346,97 @@ ztest_get_zdb_bin(char *bin, int len)
 	strcpy(bin, "zdb");
 }

+static vdev_t *
+ztest_random_concrete_vdev_leaf(vdev_t *vd)
+{
+	if (vd == NULL)
+		return (NULL);
+
+	if (vd->vdev_children == 0)
+		return (vd);
+
+	vdev_t *eligible[vd->vdev_children];
+	int eligible_idx = 0, i;
+	for (i = 0; i < vd->vdev_children; i++) {
+		vdev_t *cvd = vd->vdev_child[i];
+		if (cvd->vdev_top->vdev_removing)
+			continue;
+		if (cvd->vdev_children > 0 ||
+		    (vdev_is_concrete(cvd) && !cvd->vdev_detached)) {
+			eligible[eligible_idx++] = cvd;
+		}
+	}
+	VERIFY(eligible_idx > 0);
+
+	uint64_t child_no = ztest_random(eligible_idx);
+	return (ztest_random_concrete_vdev_leaf(eligible[child_no]));
+}
+
+/* ARGSUSED */
+void
+ztest_initialize(ztest_ds_t *zd, uint64_t id)
+{
+	spa_t *spa = ztest_spa;
+	int error = 0;
+
+	mutex_enter(&ztest_vdev_lock);
+
+	spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER);
+
+	/* Random leaf vdev */
+	vdev_t *rand_vd = ztest_random_concrete_vdev_leaf(spa->spa_root_vdev);
+	if (rand_vd == NULL) {
+		spa_config_exit(spa, SCL_VDEV, FTAG);
+		mutex_exit(&ztest_vdev_lock);
+		return;
+	}
+
+	/*
+	 * The random vdev we've selected may change as soon as we
+	 * drop the spa_config_lock. We create local copies of things
+	 * we're interested in.
+	 */
+	uint64_t guid = rand_vd->vdev_guid;
+	char *path = strdup(rand_vd->vdev_path);
+	boolean_t active = rand_vd->vdev_initialize_thread != NULL;
+
+	zfs_dbgmsg("vd %p, guid %llu", rand_vd, guid);
+	spa_config_exit(spa, SCL_VDEV, FTAG);
+
+	uint64_t cmd = ztest_random(POOL_INITIALIZE_FUNCS);
+	error = spa_vdev_initialize(spa, guid, cmd);
+	switch (cmd) {
+	case POOL_INITIALIZE_CANCEL:
+		if (ztest_opts.zo_verbose >= 4) {
+			(void) printf("Cancel initialize %s", path);
+			if (!active)
+				(void) printf(" failed (no initialize active)");
+			(void) printf("\n");
+		}
+		break;
+	case POOL_INITIALIZE_DO:
+		if (ztest_opts.zo_verbose >= 4) {
+			(void) printf("Start initialize %s", path);
+			if (active && error == 0)
+				(void) printf(" failed (already active)");
+			else if (error != 0)
+				(void) printf(" failed (error %d)", error);
+			(void) printf("\n");
+		}
+		break;
+	case POOL_INITIALIZE_SUSPEND:
+		if (ztest_opts.zo_verbose >= 4) {
+			(void) printf("Suspend initialize %s", path);
+			if (!active)
+				(void) printf(" failed (no initialize active)");
+			(void) printf("\n");
+		}
+		break;
+	}
+	free(path);
+	mutex_exit(&ztest_vdev_lock);
+}
+
 /*
 * Verify pool integrity by running zdb.
 */