Date: Mon, 18 Oct 2010 08:54:33 +0000 (UTC) From: Martin Matuska <mm@FreeBSD.org> To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-stable@freebsd.org, svn-src-stable-8@freebsd.org Subject: svn commit: r214013 - in stable/8: cddl/contrib/opensolaris/cmd/zinject cddl/contrib/opensolaris/cmd/zpool cddl/contrib/opensolaris/lib/libzfs/common sys/cddl/contrib/opensolaris/uts/common/fs/zfs ... Message-ID: <201010180854.o9I8sXgx037212@svn.freebsd.org>
next in thread | raw e-mail | index | archive | help
Author: mm Date: Mon Oct 18 08:54:33 2010 New Revision: 214013 URL: http://svn.freebsd.org/changeset/base/214013 Log: MFC r213197, r213198: MFC r213197: Enable offlining of log devices. OpenSolaris revision and Bug IDs: 9701:cc5b64682e64 6803605 should be able to offline log devices 6726045 vdev_deflate_ratio is not set when offlining a log device 6599442 zpool import has faults in the display MFC r213198: Properly handle IO with B_FAILFAST Retry IO once with ZIO_FLAG_TRYHARD before declaring a pool faulted OpenSolaris revision and Bug IDs: 9725:0bf7402e8022 6843014 ZFS B_FAILFAST handling is broken As FreeBSD does not support B_FAILFAST and we don't use vdev_disk.c, MFC of r213198 is a no-op with the purpose of diff reduction against latest ZFS code and making future updates easier. Approved by: delphij (mentor) Obtained from: OpenSolaris (Bug ID 6803605, 6726045, 6599442, 6843014) Modified: stable/8/cddl/contrib/opensolaris/cmd/zinject/zinject.c stable/8/cddl/contrib/opensolaris/cmd/zpool/zpool_main.c stable/8/cddl/contrib/opensolaris/lib/libzfs/common/libzfs.h stable/8/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_pool.c stable/8/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_util.c stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dsl_scrub.c stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/spa.c stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/vdev.h stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/vdev_impl.h stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zfs_ioctl.h stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zil.h stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zio.h stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev.c stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_disk.c stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_label.c stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_fm.c stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zil.c stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zio.c stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zio_inject.c Directory Properties: stable/8/cddl/contrib/opensolaris/ (props changed) stable/8/sys/ (props changed) stable/8/sys/amd64/include/xen/ (props changed) stable/8/sys/cddl/contrib/opensolaris/ (props changed) stable/8/sys/contrib/dev/acpica/ (props changed) stable/8/sys/contrib/pf/ (props changed) stable/8/sys/dev/xen/xenpci/ (props changed) Modified: stable/8/cddl/contrib/opensolaris/cmd/zinject/zinject.c ============================================================================== --- stable/8/cddl/contrib/opensolaris/cmd/zinject/zinject.c Mon Oct 18 08:36:03 2010 (r214012) +++ stable/8/cddl/contrib/opensolaris/cmd/zinject/zinject.c Mon Oct 18 08:54:33 2010 (r214013) @@ -19,12 +19,10 @@ * CDDL HEADER END */ /* - * Copyright 2008 Sun Microsystems, Inc. All rights reserved. + * Copyright 2009 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ -#pragma ident "%Z%%M% %I% %E% SMI" - /* * ZFS Fault Injector * @@ -227,7 +225,7 @@ usage(void) "\t\tClear the particular record (if given a numeric ID), or\n" "\t\tall records if 'all' is specificed.\n" "\n" - "\tzinject -d device [-e errno] [-L <nvlist|uber>] pool\n" + "\tzinject -d device [-e errno] [-L <nvlist|uber>] [-F] pool\n" "\t\tInject a fault into a particular device or the device's\n" "\t\tlabel. Label injection can either be 'nvlist' or 'uber'.\n" "\t\t'errno' can either be 'nxio' (the default) or 'io'.\n" @@ -519,7 +517,7 @@ main(int argc, char **argv) return (0); } - while ((c = getopt(argc, argv, ":ab:d:f:qhc:t:l:mr:e:uL:")) != -1) { + while ((c = getopt(argc, argv, ":ab:d:f:Fqhc:t:l:mr:e:uL:")) != -1) { switch (c) { case 'a': flags |= ZINJECT_FLUSH_ARC; @@ -556,6 +554,9 @@ main(int argc, char **argv) return (1); } break; + case 'F': + record.zi_failfast = B_TRUE; + break; case 'h': usage(); return (0); Modified: stable/8/cddl/contrib/opensolaris/cmd/zpool/zpool_main.c ============================================================================== --- stable/8/cddl/contrib/opensolaris/cmd/zpool/zpool_main.c Mon Oct 18 08:36:03 2010 (r214012) +++ stable/8/cddl/contrib/opensolaris/cmd/zpool/zpool_main.c Mon Oct 18 08:54:33 2010 (r214013) @@ -980,14 +980,189 @@ max_width(zpool_handle_t *zhp, nvlist_t return (max); } +typedef struct spare_cbdata { + uint64_t cb_guid; + zpool_handle_t *cb_zhp; +} spare_cbdata_t; + +static boolean_t +find_vdev(nvlist_t *nv, uint64_t search) +{ + uint64_t guid; + nvlist_t **child; + uint_t c, children; + + if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) == 0 && + search == guid) + return (B_TRUE); + + if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, + &child, &children) == 0) { + for (c = 0; c < children; c++) + if (find_vdev(child[c], search)) + return (B_TRUE); + } + + return (B_FALSE); +} + +static int +find_spare(zpool_handle_t *zhp, void *data) +{ + spare_cbdata_t *cbp = data; + nvlist_t *config, *nvroot; + + config = zpool_get_config(zhp, NULL); + verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, + &nvroot) == 0); + + if (find_vdev(nvroot, cbp->cb_guid)) { + cbp->cb_zhp = zhp; + return (1); + } + + zpool_close(zhp); + return (0); +} + +/* + * Print out configuration state as requested by status_callback. + */ +void +print_status_config(zpool_handle_t *zhp, const char *name, nvlist_t *nv, + int namewidth, int depth, boolean_t isspare) +{ + nvlist_t **child; + uint_t c, children; + vdev_stat_t *vs; + char rbuf[6], wbuf[6], cbuf[6], repaired[7]; + char *vname; + uint64_t notpresent; + spare_cbdata_t cb; + char *state; + + verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_STATS, + (uint64_t **)&vs, &c) == 0); + + if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, + &child, &children) != 0) + children = 0; + + state = zpool_state_to_name(vs->vs_state, vs->vs_aux); + if (isspare) { + /* + * For hot spares, we use the terms 'INUSE' and 'AVAILABLE' for + * online drives. + */ + if (vs->vs_aux == VDEV_AUX_SPARED) + state = "INUSE"; + else if (vs->vs_state == VDEV_STATE_HEALTHY) + state = "AVAIL"; + } + + (void) printf("\t%*s%-*s %-8s", depth, "", namewidth - depth, + name, state); + + if (!isspare) { + zfs_nicenum(vs->vs_read_errors, rbuf, sizeof (rbuf)); + zfs_nicenum(vs->vs_write_errors, wbuf, sizeof (wbuf)); + zfs_nicenum(vs->vs_checksum_errors, cbuf, sizeof (cbuf)); + (void) printf(" %5s %5s %5s", rbuf, wbuf, cbuf); + } + + if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT, + ¬present) == 0) { + char *path; + verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0); + (void) printf(" was %s", path); + } else if (vs->vs_aux != 0) { + (void) printf(" "); + + switch (vs->vs_aux) { + case VDEV_AUX_OPEN_FAILED: + (void) printf(gettext("cannot open")); + break; + + case VDEV_AUX_BAD_GUID_SUM: + (void) printf(gettext("missing device")); + break; + + case VDEV_AUX_NO_REPLICAS: + (void) printf(gettext("insufficient replicas")); + break; + + case VDEV_AUX_VERSION_NEWER: + (void) printf(gettext("newer version")); + break; + + case VDEV_AUX_SPARED: + verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, + &cb.cb_guid) == 0); + if (zpool_iter(g_zfs, find_spare, &cb) == 1) { + if (strcmp(zpool_get_name(cb.cb_zhp), + zpool_get_name(zhp)) == 0) + (void) printf(gettext("currently in " + "use")); + else + (void) printf(gettext("in use by " + "pool '%s'"), + zpool_get_name(cb.cb_zhp)); + zpool_close(cb.cb_zhp); + } else { + (void) printf(gettext("currently in use")); + } + break; + + case VDEV_AUX_ERR_EXCEEDED: + (void) printf(gettext("too many errors")); + break; + + case VDEV_AUX_IO_FAILURE: + (void) printf(gettext("experienced I/O failures")); + break; + + case VDEV_AUX_BAD_LOG: + (void) printf(gettext("bad intent log")); + break; + + default: + (void) printf(gettext("corrupted data")); + break; + } + } else if (vs->vs_scrub_repaired != 0 && children == 0) { + /* + * Report bytes resilvered/repaired on leaf devices. + */ + zfs_nicenum(vs->vs_scrub_repaired, repaired, sizeof (repaired)); + (void) printf(gettext(" %s %s"), repaired, + (vs->vs_scrub_type == POOL_SCRUB_RESILVER) ? + "resilvered" : "repaired"); + } + + (void) printf("\n"); + + for (c = 0; c < children; c++) { + uint64_t is_log = B_FALSE; + + /* Don't print logs here */ + (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, + &is_log); + if (is_log) + continue; + vname = zpool_vdev_name(g_zfs, zhp, child[c]); + print_status_config(zhp, vname, child[c], + namewidth, depth + 2, isspare); + free(vname); + } +} + /* * Print the configuration of an exported pool. Iterate over all vdevs in the * pool, printing out the name and status for each one. */ void -print_import_config(const char *name, nvlist_t *nv, int namewidth, int depth, - boolean_t print_logs) +print_import_config(const char *name, nvlist_t *nv, int namewidth, int depth) { nvlist_t **child; uint_t c, children; @@ -1044,12 +1219,11 @@ print_import_config(const char *name, nv (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, &is_log); - if ((is_log && !print_logs) || (!is_log && print_logs)) + if (is_log) continue; vname = zpool_vdev_name(g_zfs, NULL, child[c]); - print_import_config(vname, child[c], - namewidth, depth + 2, B_FALSE); + print_import_config(vname, child[c], namewidth, depth + 2); free(vname); } @@ -1075,6 +1249,43 @@ print_import_config(const char *name, nv } /* + * Print log vdevs. + * Logs are recorded as top level vdevs in the main pool child array + * but with "is_log" set to 1. We use either print_status_config() or + * print_import_config() to print the top level logs then any log + * children (eg mirrored slogs) are printed recursively - which + * works because only the top level vdev is marked "is_log" + */ +static void +print_logs(zpool_handle_t *zhp, nvlist_t *nv, int namewidth, boolean_t verbose) +{ + uint_t c, children; + nvlist_t **child; + + if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, + &children) != 0) + return; + + (void) printf(gettext("\tlogs\n")); + + for (c = 0; c < children; c++) { + uint64_t is_log = B_FALSE; + char *name; + + (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, + &is_log); + if (!is_log) + continue; + name = zpool_vdev_name(g_zfs, zhp, child[c]); + if (verbose) + print_status_config(zhp, name, child[c], namewidth, + 2, B_FALSE); + else + print_import_config(name, child[c], namewidth, 2); + free(name); + } +} +/* * Display the status for the given pool. */ static void @@ -1242,11 +1453,9 @@ show_import(nvlist_t *config) if (namewidth < 10) namewidth = 10; - print_import_config(name, nvroot, namewidth, 0, B_FALSE); - if (num_logs(nvroot) > 0) { - (void) printf(gettext("\tlogs\n")); - print_import_config(name, nvroot, namewidth, 0, B_TRUE); - } + print_import_config(name, nvroot, namewidth, 0); + if (num_logs(nvroot) > 0) + print_logs(NULL, nvroot, namewidth, B_FALSE); if (reason == ZPOOL_STATUS_BAD_GUID_SUM) { (void) printf(gettext("\n\tAdditional devices are known to " @@ -2717,182 +2926,6 @@ print_scrub_status(nvlist_t *nvroot) (u_longlong_t)(minutes_left / 60), (uint_t)(minutes_left % 60)); } -typedef struct spare_cbdata { - uint64_t cb_guid; - zpool_handle_t *cb_zhp; -} spare_cbdata_t; - -static boolean_t -find_vdev(nvlist_t *nv, uint64_t search) -{ - uint64_t guid; - nvlist_t **child; - uint_t c, children; - - if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) == 0 && - search == guid) - return (B_TRUE); - - if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, - &child, &children) == 0) { - for (c = 0; c < children; c++) - if (find_vdev(child[c], search)) - return (B_TRUE); - } - - return (B_FALSE); -} - -static int -find_spare(zpool_handle_t *zhp, void *data) -{ - spare_cbdata_t *cbp = data; - nvlist_t *config, *nvroot; - - config = zpool_get_config(zhp, NULL); - verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, - &nvroot) == 0); - - if (find_vdev(nvroot, cbp->cb_guid)) { - cbp->cb_zhp = zhp; - return (1); - } - - zpool_close(zhp); - return (0); -} - -/* - * Print out configuration state as requested by status_callback. - */ -void -print_status_config(zpool_handle_t *zhp, const char *name, nvlist_t *nv, - int namewidth, int depth, boolean_t isspare) -{ - nvlist_t **child; - uint_t c, children; - vdev_stat_t *vs; - char rbuf[6], wbuf[6], cbuf[6], repaired[7]; - char *vname; - uint64_t notpresent; - spare_cbdata_t cb; - char *state; - - verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_STATS, - (uint64_t **)&vs, &c) == 0); - - if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, - &child, &children) != 0) - children = 0; - - state = zpool_state_to_name(vs->vs_state, vs->vs_aux); - if (isspare) { - /* - * For hot spares, we use the terms 'INUSE' and 'AVAILABLE' for - * online drives. - */ - if (vs->vs_aux == VDEV_AUX_SPARED) - state = "INUSE"; - else if (vs->vs_state == VDEV_STATE_HEALTHY) - state = "AVAIL"; - } - - (void) printf("\t%*s%-*s %-8s", depth, "", namewidth - depth, - name, state); - - if (!isspare) { - zfs_nicenum(vs->vs_read_errors, rbuf, sizeof (rbuf)); - zfs_nicenum(vs->vs_write_errors, wbuf, sizeof (wbuf)); - zfs_nicenum(vs->vs_checksum_errors, cbuf, sizeof (cbuf)); - (void) printf(" %5s %5s %5s", rbuf, wbuf, cbuf); - } - - if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT, - ¬present) == 0) { - char *path; - verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0); - (void) printf(" was %s", path); - } else if (vs->vs_aux != 0) { - (void) printf(" "); - - switch (vs->vs_aux) { - case VDEV_AUX_OPEN_FAILED: - (void) printf(gettext("cannot open")); - break; - - case VDEV_AUX_BAD_GUID_SUM: - (void) printf(gettext("missing device")); - break; - - case VDEV_AUX_NO_REPLICAS: - (void) printf(gettext("insufficient replicas")); - break; - - case VDEV_AUX_VERSION_NEWER: - (void) printf(gettext("newer version")); - break; - - case VDEV_AUX_SPARED: - verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, - &cb.cb_guid) == 0); - if (zpool_iter(g_zfs, find_spare, &cb) == 1) { - if (strcmp(zpool_get_name(cb.cb_zhp), - zpool_get_name(zhp)) == 0) - (void) printf(gettext("currently in " - "use")); - else - (void) printf(gettext("in use by " - "pool '%s'"), - zpool_get_name(cb.cb_zhp)); - zpool_close(cb.cb_zhp); - } else { - (void) printf(gettext("currently in use")); - } - break; - - case VDEV_AUX_ERR_EXCEEDED: - (void) printf(gettext("too many errors")); - break; - - case VDEV_AUX_IO_FAILURE: - (void) printf(gettext("experienced I/O failures")); - break; - - case VDEV_AUX_BAD_LOG: - (void) printf(gettext("bad intent log")); - break; - - default: - (void) printf(gettext("corrupted data")); - break; - } - } else if (vs->vs_scrub_repaired != 0 && children == 0) { - /* - * Report bytes resilvered/repaired on leaf devices. - */ - zfs_nicenum(vs->vs_scrub_repaired, repaired, sizeof (repaired)); - (void) printf(gettext(" %s %s"), repaired, - (vs->vs_scrub_type == POOL_SCRUB_RESILVER) ? - "resilvered" : "repaired"); - } - - (void) printf("\n"); - - for (c = 0; c < children; c++) { - uint64_t is_log = B_FALSE; - - /* Don't print logs here */ - (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, - &is_log); - if (is_log) - continue; - vname = zpool_vdev_name(g_zfs, zhp, child[c]); - print_status_config(zhp, vname, child[c], - namewidth, depth + 2, isspare); - free(vname); - } -} - static void print_error_log(zpool_handle_t *zhp) { @@ -2969,39 +3002,6 @@ print_l2cache(zpool_handle_t *zhp, nvlis } /* - * Print log vdevs. - * Logs are recorded as top level vdevs in the main pool child array but with - * "is_log" set to 1. We use print_status_config() to print the top level logs - * then any log children (eg mirrored slogs) are printed recursively - which - * works because only the top level vdev is marked "is_log" - */ -static void -print_logs(zpool_handle_t *zhp, nvlist_t *nv, int namewidth) -{ - uint_t c, children; - nvlist_t **child; - - if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, - &children) != 0) - return; - - (void) printf(gettext("\tlogs\n")); - - for (c = 0; c < children; c++) { - uint64_t is_log = B_FALSE; - char *name; - - (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, - &is_log); - if (!is_log) - continue; - name = zpool_vdev_name(g_zfs, zhp, child[c]); - print_status_config(zhp, name, child[c], namewidth, 2, B_FALSE); - free(name); - } -} - -/* * Display a summary of pool status. Displays a summary such as: * * pool: tank @@ -3229,7 +3229,7 @@ status_callback(zpool_handle_t *zhp, voi namewidth, 0, B_FALSE); if (num_logs(nvroot) > 0) - print_logs(zhp, nvroot, namewidth); + print_logs(zhp, nvroot, namewidth, B_TRUE); if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache) == 0) print_l2cache(zhp, l2cache, nl2cache, namewidth); Modified: stable/8/cddl/contrib/opensolaris/lib/libzfs/common/libzfs.h ============================================================================== --- stable/8/cddl/contrib/opensolaris/lib/libzfs/common/libzfs.h Mon Oct 18 08:36:03 2010 (r214012) +++ stable/8/cddl/contrib/opensolaris/lib/libzfs/common/libzfs.h Mon Oct 18 08:54:33 2010 (r214013) @@ -116,6 +116,7 @@ enum { EZFS_VDEVNOTSUP, /* unsupported vdev type */ EZFS_NOTSUP, /* ops not supported on this dataset */ EZFS_ACTIVE_SPARE, /* pool has active shared spare devices */ + EZFS_UNPLAYED_LOGS, /* log device has unplayed logs */ EZFS_UNKNOWN }; Modified: stable/8/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_pool.c ============================================================================== --- stable/8/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_pool.c Mon Oct 18 08:36:03 2010 (r214012) +++ stable/8/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_pool.c Mon Oct 18 08:54:33 2010 (r214013) @@ -1720,6 +1720,12 @@ zpool_vdev_fault(zpool_handle_t *zhp, ui */ return (zfs_error(hdl, EZFS_NOREPLICAS, msg)); + case EEXIST: + /* + * The log device has unplayed logs + */ + return (zfs_error(hdl, EZFS_UNPLAYED_LOGS, msg)); + default: return (zpool_standard_error(hdl, errno, msg)); } Modified: stable/8/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_util.c ============================================================================== --- stable/8/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_util.c Mon Oct 18 08:36:03 2010 (r214012) +++ stable/8/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_util.c Mon Oct 18 08:54:33 2010 (r214013) @@ -210,6 +210,9 @@ libzfs_error_description(libzfs_handle_t case EZFS_ACTIVE_SPARE: return (dgettext(TEXT_DOMAIN, "pool has active shared spare " "device")); + case EZFS_UNPLAYED_LOGS: + return (dgettext(TEXT_DOMAIN, "log device has unplayed intent " + "logs")); case EZFS_UNKNOWN: return (dgettext(TEXT_DOMAIN, "unknown error")); default: Modified: stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dsl_scrub.c ============================================================================== --- stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dsl_scrub.c Mon Oct 18 08:36:03 2010 (r214012) +++ stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dsl_scrub.c Mon Oct 18 08:54:33 2010 (r214013) @@ -351,7 +351,7 @@ traverse_zil_block(zilog_t *zilog, blkpt return; /* - * One block ("stumpy") can be allocated a long time ago; we + * One block ("stubby") can be allocated a long time ago; we * want to visit that one because it has been allocated * (on-disk) even if it hasn't been claimed (even though for * plain scrub there's nothing to do to it). Modified: stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/spa.c ============================================================================== --- stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/spa.c Mon Oct 18 08:36:03 2010 (r214012) +++ stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/spa.c Mon Oct 18 08:54:33 2010 (r214013) @@ -1110,6 +1110,33 @@ spa_check_removed(vdev_t *vd) } /* + * Load the slog device state from the config object since it's possible + * that the label does not contain the most up-to-date information. + */ +void +spa_load_log_state(spa_t *spa) +{ + nvlist_t *nv, *nvroot, **child; + uint64_t is_log; + uint_t children, c; + vdev_t *rvd = spa->spa_root_vdev; + + VERIFY(load_nvlist(spa, spa->spa_config_object, &nv) == 0); + VERIFY(nvlist_lookup_nvlist(nv, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); + VERIFY(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, + &child, &children) == 0); + + for (c = 0; c < children; c++) { + vdev_t *tvd = rvd->vdev_child[c]; + + if (nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, + &is_log) == 0 && is_log) + vdev_load_log_state(tvd, child[c]); + } + nvlist_free(nv); +} + +/* * Check for missing log devices */ int @@ -1125,13 +1152,7 @@ spa_check_logs(spa_t *spa) return (1); } break; - - case SPA_LOG_CLEAR: - (void) dmu_objset_find(spa->spa_name, zil_clear_log_chain, NULL, - DS_FIND_CHILDREN); - break; } - spa->spa_log_state = SPA_LOG_GOOD; return (0); } @@ -1455,6 +1476,8 @@ spa_load(spa_t *spa, nvlist_t *config, s spa_config_exit(spa, SCL_ALL, FTAG); } + spa_load_log_state(spa); + if (spa_check_logs(spa)) { vdev_set_state(rvd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_BAD_LOG); @@ -1542,6 +1565,7 @@ spa_load(spa_t *spa, nvlist_t *config, s zil_claim, tx, DS_FIND_CHILDREN); dmu_tx_commit(tx); + spa->spa_log_state = SPA_LOG_GOOD; spa->spa_sync_on = B_TRUE; txg_sync_start(spa->spa_dsl_pool); @@ -4222,10 +4246,16 @@ spa_sync(spa_t *spa, uint64_t txg) if (svdcount == SPA_DVAS_PER_BP) break; } - error = vdev_config_sync(svd, svdcount, txg); + error = vdev_config_sync(svd, svdcount, txg, B_FALSE); + if (error != 0) + error = vdev_config_sync(svd, svdcount, txg, + B_TRUE); } else { error = vdev_config_sync(rvd->vdev_child, - rvd->vdev_children, txg); + rvd->vdev_children, txg, B_FALSE); + if (error != 0) + error = vdev_config_sync(rvd->vdev_child, + rvd->vdev_children, txg, B_TRUE); } spa_config_exit(spa, SCL_STATE, FTAG); Modified: stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/vdev.h ============================================================================== --- stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/vdev.h Mon Oct 18 08:36:03 2010 (r214012) +++ stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/vdev.h Mon Oct 18 08:54:33 2010 (r214013) @@ -19,7 +19,7 @@ * CDDL HEADER END */ /* - * Copyright 2008 Sun Microsystems, Inc. All rights reserved. + * Copyright 2009 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ @@ -113,7 +113,8 @@ extern void vdev_queue_io_done(zio_t *zi extern void vdev_config_dirty(vdev_t *vd); extern void vdev_config_clean(vdev_t *vd); -extern int vdev_config_sync(vdev_t **svd, int svdcount, uint64_t txg); +extern int vdev_config_sync(vdev_t **svd, int svdcount, uint64_t txg, + boolean_t); extern void vdev_state_dirty(vdev_t *vd); extern void vdev_state_clean(vdev_t *vd); Modified: stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/vdev_impl.h ============================================================================== --- stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/vdev_impl.h Mon Oct 18 08:36:03 2010 (r214012) +++ stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/vdev_impl.h Mon Oct 18 08:54:33 2010 (r214013) @@ -259,6 +259,7 @@ extern void vdev_remove_parent(vdev_t *c /* * vdev sync load and sync */ +extern void vdev_load_log_state(vdev_t *vd, nvlist_t *nv); extern void vdev_load(vdev_t *vd); extern void vdev_sync(vdev_t *vd, uint64_t txg); extern void vdev_sync_done(vdev_t *vd, uint64_t txg); Modified: stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zfs_ioctl.h ============================================================================== --- stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zfs_ioctl.h Mon Oct 18 08:36:03 2010 (r214012) +++ stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zfs_ioctl.h Mon Oct 18 08:54:33 2010 (r214013) @@ -118,7 +118,7 @@ typedef struct zinject_record { uint32_t zi_error; uint64_t zi_type; uint32_t zi_freq; - uint32_t zi_pad; /* pad out to 64 bit alignment */ + uint32_t zi_failfast; } zinject_record_t; #define ZINJECT_NULL 0x1 Modified: stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zil.h ============================================================================== --- stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zil.h Mon Oct 18 08:36:03 2010 (r214012) +++ stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zil.h Mon Oct 18 08:54:33 2010 (r214013) @@ -366,9 +366,9 @@ extern uint64_t zil_itx_assign(zilog_t * extern void zil_commit(zilog_t *zilog, uint64_t seq, uint64_t oid); +extern int zil_vdev_offline(char *osname, void *txarg); extern int zil_claim(char *osname, void *txarg); extern int zil_check_log_chain(char *osname, void *txarg); -extern int zil_clear_log_chain(char *osname, void *txarg); extern void zil_sync(zilog_t *zilog, dmu_tx_t *tx); extern void zil_clean(zilog_t *zilog); extern int zil_is_committed(zilog_t *zilog); Modified: stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zio.h ============================================================================== --- stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zio.h Mon Oct 18 08:36:03 2010 (r214012) +++ stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zio.h Mon Oct 18 08:54:33 2010 (r214013) @@ -117,31 +117,33 @@ enum zio_compress { #define ZIO_PRIORITY_SCRUB (zio_priority_table[10]) #define ZIO_PRIORITY_TABLE_SIZE 11 -#define ZIO_FLAG_MUSTSUCCEED 0x00000 -#define ZIO_FLAG_CANFAIL 0x00001 -#define ZIO_FLAG_SPECULATIVE 0x00002 -#define ZIO_FLAG_CONFIG_WRITER 0x00004 -#define ZIO_FLAG_DONT_RETRY 0x00008 - -#define ZIO_FLAG_DONT_CACHE 0x00010 -#define ZIO_FLAG_DONT_QUEUE 0x00020 -#define ZIO_FLAG_DONT_AGGREGATE 0x00040 -#define ZIO_FLAG_DONT_PROPAGATE 0x00080 - -#define ZIO_FLAG_IO_BYPASS 0x00100 -#define ZIO_FLAG_IO_REPAIR 0x00200 -#define ZIO_FLAG_IO_RETRY 0x00400 -#define ZIO_FLAG_IO_REWRITE 0x00800 - -#define ZIO_FLAG_SELF_HEAL 0x01000 -#define ZIO_FLAG_RESILVER 0x02000 -#define ZIO_FLAG_SCRUB 0x04000 -#define ZIO_FLAG_SCRUB_THREAD 0x08000 - -#define ZIO_FLAG_PROBE 0x10000 -#define ZIO_FLAG_GANG_CHILD 0x20000 -#define ZIO_FLAG_RAW 0x40000 -#define ZIO_FLAG_GODFATHER 0x80000 +#define ZIO_FLAG_MUSTSUCCEED 0x000000 +#define ZIO_FLAG_CANFAIL 0x000001 +#define ZIO_FLAG_SPECULATIVE 0x000002 +#define ZIO_FLAG_CONFIG_WRITER 0x000004 +#define ZIO_FLAG_DONT_RETRY 0x000008 + +#define ZIO_FLAG_DONT_CACHE 0x000010 +#define ZIO_FLAG_DONT_QUEUE 0x000020 +#define ZIO_FLAG_DONT_AGGREGATE 0x000040 +#define ZIO_FLAG_DONT_PROPAGATE 0x000080 + +#define ZIO_FLAG_IO_BYPASS 0x000100 +#define ZIO_FLAG_IO_REPAIR 0x000200 +#define ZIO_FLAG_IO_RETRY 0x000400 +#define ZIO_FLAG_IO_REWRITE 0x000800 + +#define ZIO_FLAG_SELF_HEAL 0x001000 +#define ZIO_FLAG_RESILVER 0x002000 +#define ZIO_FLAG_SCRUB 0x004000 +#define ZIO_FLAG_SCRUB_THREAD 0x008000 + +#define ZIO_FLAG_PROBE 0x010000 +#define ZIO_FLAG_GANG_CHILD 0x020000 +#define ZIO_FLAG_RAW 0x040000 +#define ZIO_FLAG_GODFATHER 0x080000 + +#define ZIO_FLAG_TRYHARD 0x100000 #define ZIO_FLAG_GANG_INHERIT \ (ZIO_FLAG_CANFAIL | \ @@ -159,7 +161,8 @@ enum zio_compress { (ZIO_FLAG_GANG_INHERIT | \ ZIO_FLAG_IO_REPAIR | \ ZIO_FLAG_IO_RETRY | \ - ZIO_FLAG_PROBE) + ZIO_FLAG_PROBE | \ + ZIO_FLAG_TRYHARD) #define ZIO_FLAG_AGG_INHERIT \ (ZIO_FLAG_DONT_AGGREGATE | \ @@ -440,7 +443,7 @@ extern int zio_inject_list_next(int *id, struct zinject_record *record); extern int zio_clear_fault(int id); extern int zio_handle_fault_injection(zio_t *zio, int error); -extern int zio_handle_device_injection(vdev_t *vd, int error); +extern int zio_handle_device_injection(vdev_t *vd, zio_t *zio, int error); extern int zio_handle_label_injection(zio_t *zio, int error); #ifdef __cplusplus Modified: stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev.c ============================================================================== --- stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev.c Mon Oct 18 08:36:03 2010 (r214012) +++ stable/8/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev.c Mon Oct 18 08:54:33 2010 (r214013) @@ -39,6 +39,7 @@ #include <sys/zap.h> #include <sys/fs/zfs.h> #include <sys/arc.h> +#include <sys/zil.h> SYSCTL_DECL(_vfs_zfs); SYSCTL_NODE(_vfs_zfs, OID_AUTO, vdev, CTLFLAG_RW, 0, "ZFS VDEV"); @@ -765,6 +766,15 @@ vdev_metaslab_init(vdev_t *vd, uint64_t if (vd->vdev_ms_shift == 0) /* not being allocated from yet */ return (0); + /* + * Compute the raidz-deflation ratio. Note, we hard-code + * in 128k (1 << 17) because it is the current "typical" blocksize. + * Even if SPA_MAXBLOCKSIZE changes, this algorithm must never change, + * or we will inconsistently account for existing bp's. + */ + vd->vdev_deflate_ratio = (1 << 17) / + (vdev_psize_to_asize(vd, 1 << 17) >> SPA_MINBLOCKSHIFT); + ASSERT(oldc <= newc); if (vd->vdev_islog) @@ -918,7 +928,7 @@ vdev_probe(vdev_t *vd, zio_t *zio) vps->vps_flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_PROBE | ZIO_FLAG_DONT_CACHE | ZIO_FLAG_DONT_AGGREGATE | - ZIO_FLAG_DONT_RETRY; + ZIO_FLAG_TRYHARD; if (spa_config_held(spa, SCL_ZIO, RW_WRITER)) { /* @@ -998,6 +1008,8 @@ vdev_open(vdev_t *vd) vd->vdev_state == VDEV_STATE_OFFLINE); vd->vdev_stat.vs_aux = VDEV_AUX_NONE; + vd->vdev_cant_read = B_FALSE; + vd->vdev_cant_write = B_FALSE; if (!vd->vdev_removed && vd->vdev_faulted) { ASSERT(vd->vdev_children == 0); @@ -1013,7 +1025,7 @@ vdev_open(vdev_t *vd) error = vd->vdev_ops->vdev_op_open(vd, &osize, &ashift); if (zio_injection_enabled && error == 0) - error = zio_handle_device_injection(vd, ENXIO); + error = zio_handle_device_injection(vd, NULL, ENXIO); if (error) { if (vd->vdev_removed && @@ -1113,18 +1125,6 @@ vdev_open(vdev_t *vd) } /* - * If this is a top-level vdev, compute the raidz-deflation - * ratio. Note, we hard-code in 128k (1<<17) because it is the - * current "typical" blocksize. Even if SPA_MAXBLOCKSIZE - * changes, this algorithm must never change, or we will - * inconsistently account for existing bp's. - */ - if (vd->vdev_top == vd) { - vd->vdev_deflate_ratio = (1<<17) / - (vdev_psize_to_asize(vd, 1<<17) >> SPA_MINBLOCKSHIFT); - } - - /* * If a leaf vdev has a DTL, and seems healthy, then kick off a * resilver. But don't do this if we are doing a reopen for a scrub, * since this would just restart the scrub we are already doing. @@ -1937,7 +1937,8 @@ vdev_online(spa_t *spa, uint64_t guid, u int vdev_offline(spa_t *spa, uint64_t guid, uint64_t flags) { - vdev_t *vd; + vdev_t *vd, *tvd; + int error; spa_vdev_state_enter(spa); @@ -1947,34 +1948,58 @@ vdev_offline(spa_t *spa, uint64_t guid, if (!vd->vdev_ops->vdev_op_leaf) return (spa_vdev_state_exit(spa, NULL, ENOTSUP)); + tvd = vd->vdev_top; + /* * If the device isn't already offline, try to offline it. */ if (!vd->vdev_offline) { /* * If this device has the only valid copy of some data, - * don't allow it to be offlined. + * don't allow it to be offlined. Log devices are always + * expendable. */ - if (vd->vdev_aux == NULL && vdev_dtl_required(vd)) + if (!tvd->vdev_islog && vd->vdev_aux == NULL && + vdev_dtl_required(vd)) return (spa_vdev_state_exit(spa, NULL, EBUSY)); /* * Offline this device and reopen its top-level vdev. - * If this action results in the top-level vdev becoming - * unusable, undo it and fail the request. + * If the top-level vdev is a log device then just offline + * it. Otherwise, if this action results in the top-level + * vdev becoming unusable, undo it and fail the request. */ vd->vdev_offline = B_TRUE; - vdev_reopen(vd->vdev_top); - if (vd->vdev_aux == NULL && vdev_is_dead(vd->vdev_top)) { + vdev_reopen(tvd); + + if (!tvd->vdev_islog && vd->vdev_aux == NULL && + vdev_is_dead(tvd)) { vd->vdev_offline = B_FALSE; - vdev_reopen(vd->vdev_top); + vdev_reopen(tvd); return (spa_vdev_state_exit(spa, NULL, EBUSY)); } } vd->vdev_tmpoffline = !!(flags & ZFS_OFFLINE_TEMPORARY); - return (spa_vdev_state_exit(spa, vd, 0)); + if (!tvd->vdev_islog || !vdev_is_dead(tvd)) + return (spa_vdev_state_exit(spa, vd, 0)); + + (void) spa_vdev_state_exit(spa, vd, 0); + + error = dmu_objset_find(spa_name(spa), zil_vdev_offline, + NULL, DS_FIND_CHILDREN); + if (error) { + (void) vdev_online(spa, guid, 0, NULL); + return (error); + } + /* + * If we successfully offlined the log device then we need to + * sync out the current txg so that the "stubby" block can be + * removed by zil_sync(). + */ + txg_wait_synced(spa->spa_dsl_pool, 0); + return (0); } /* @@ -2182,6 +2207,16 @@ vdev_stat_update(zio_t *zio, uint64_t ps if (flags & ZIO_FLAG_SPECULATIVE) return; + /* + * If this is an I/O error that is going to be retried, then ignore the + * error. Otherwise, the user may interpret B_FAILFAST I/O errors as + * hard errors, when in reality they can happen for any number of + * innocuous reasons (bus resets, MPxIO link failure, etc). + */ + if (zio->io_error == EIO && + !(zio->io_flags & ZIO_FLAG_IO_RETRY)) + return; + mutex_enter(&vd->vdev_stat_lock); if (type == ZIO_TYPE_READ && !vdev_is_dead(vd)) { if (zio->io_error == ECKSUM) @@ -2279,6 +2314,7 @@ vdev_space_update(vdev_t *vd, int64_t sp * childrens', thus not accurate enough for us. */ ASSERT((dspace_delta & (SPA_MINBLOCKSIZE-1)) == 0); + ASSERT(vd->vdev_deflate_ratio != 0 || vd->vdev_isl2cache); dspace_delta = (dspace_delta >> SPA_MINBLOCKSHIFT) * vd->vdev_deflate_ratio; @@ -2631,11 +2667,7 @@ vdev_set_state(vdev_t *vd, boolean_t iso boolean_t vdev_is_bootable(vdev_t *vd) { -#ifdef __FreeBSD_version - return (B_TRUE); -#else - int c; - +#ifdef sun if (!vd->vdev_ops->vdev_op_leaf) { char *vdev_type = vd->vdev_ops->vdev_op_type; *** DIFF OUTPUT TRUNCATED AT 1000 LINES ***
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201010180854.o9I8sXgx037212>