fs/zfs/vdev.c

61 /* maximum scrub/resilver I/O queue per leaf vdev */
65  * Given a vdev type, return the appropriate ops vector.
99  * the vdev's asize rounded to the nearest metaslab. This allows us to
116 	 * The top-level vdev just returns the allocatable size rounded
123 	 * The allocatable space for a raidz vdev is N * sizeof(smallest child),
142 vdev_lookup_top(spa_t *spa, uint64_t vdev)
148 	if (vdev < rvd->vdev_children) {
149 		ASSERT(rvd->vdev_child[vdev] != NULL);
150 		return (rvd->vdev_child[vdev]);
294 			 * The root vdev's guid will also be the pool guid,
300 			 * Any other vdev's guid must be unique within the pool.
325 	    offsetof(struct vdev, vdev_dtl_node));
334  * Allocate a new vdev.  The 'alloctype' is used to control whether we are
335  * creating a new vdev or loading an existing one - the behavior is slightly
356 	 * If this is a load, get the vdev guid from the nvlist.
380 	 * The first allocated vdev must be of type 'root'.
386 	 * Determine whether we're a log vdev.
468 	 * Retrieve the vdev creation time.
474 	 * If we're a top-level vdev, try to load the allocation parameters.
498 	 * If we're a leaf vdev, try to load the DTL object and other state.
567 	 * vdev_free() implies closing the vdev first.  This is simpler than
597 	 * Remove this vdev from its parent's child list.
604 	 * Clean up vdev structure.
644  * Transfer top-level vdev state from svd to tvd.
720  * Add a mirror/replacing vdev above an existing vdev.
752  * Remove a 1-way mirror/replacing vdev from the tree.
772 	 * If cvd will replace mvd as a top-level vdev, preserve mvd's guid.
775 	 * instead of a different version of the same top-level vdev.
808 	 * This vdev is not being allocated from yet or is a hole.
864 	 * If the vdev is being removed we don't activate
954  * to several known locations: the pad regions of each vdev label
975 	 * this vdev will become parents of the probe io.
1013 		 * We can't change the vdev state in this context, so we
1121 	 * If this vdev is not removed, check its fault status.  If it's
1141 	 * the vdev on error.
1161 	 * the vdev is accessible.  If we're faulted, bail.
1257 	 * vdev open for business.
1267 	 * If a leaf vdev has a DTL, and seems healthy, then kick off a
1285  * /etc/zfs/zpool.cache was readonly at the time.  Otherwise, the vdev state
1316 		 * Determine if this vdev has been split off into another
1341 		 * If this vdev just became a top-level vdev because its
1343 		 * vdev guid -- but the label may or may not be on disk yet.
1345 		 * same top guid, so if we're a top-level vdev, we can
1348 		 * If we split this vdev off instead, then we also check the
1349 		 * original pool's guid.  We don't want to consider the vdev
1384 		 * If we were able to open and validate a vdev that was
1473 	/* set the reopening flag unless we're taking the vdev offline */
1494 	 * Reassess parent vdev's health.
1532 	 * Aim for roughly 200 metaslabs per vdev.
1558  * A vdev's DTL (dirty time log) is the set of transaction groups for which
1559  * the vdev has less than perfect replication.  There are four kinds of DTL:
1561  * DTL_MISSING: txgs for which the vdev has no valid copies of the data
1579  * A vdev's DTL_PARTIAL is the union of its children's DTL_PARTIALs, because
1581  * A vdev's DTL_MISSING is a modified union of its children's DTL_MISSINGs,
1830  * Determine whether the specified vdev can be offlined/detached/removed
1848 	 * whether this results in any DTL outages in the top-level vdev.
1916 	 * If this is a top-level vdev, initialize its metaslabs.
1925 	 * If this is a leaf vdev, load its DTL.
1933  * The special vdev case is used for hot spares and l2cache devices.  Its
1934  * sole purpose it to set the vdev state for the associated vdev.  To do this,
2046 	 * Remove the metadata associated with this vdev once it's empty.
2069  * Mark the given vdev faulted.  A faulted vdev behaves as if the device could
2104 	 * back off and simply mark the vdev as degraded instead.
2124  * Mark the given vdev degraded.  A degraded vdev is purely an indication to the
2125  * user that something is wrong.  The vdev continues to operate as normal as far
2142 	 * If the vdev is already faulted, then don't do anything.
2156  * Online the given vdev.  If 'unspare' is set, it implies two things.  First,
2248 		 * then proceed.  We check that the vdev's metaslab group
2250 		 * added this vdev but not yet initialized its metaslabs.
2278 		 * Offline this device and reopen its top-level vdev.
2279 		 * If the top-level vdev is a log device then just offline
2281 		 * vdev becoming unusable, undo it and fail the request.
2319  * Clear the error counts associated with this vdev.  Unlike vdev_online() and
2343 	 * also mark the vdev config dirty, so that the new faulted state is
2419 	 * the proper locks.  Note that we have to get the vdev state
2445  * Get statistics for the given vdev.
2462 	 * If we're getting stats on the root vdev, aggregate the I/O counts
2529 		 * (Holes never create vdev children, so all the counters
2535 		 * one top-level vdev does not imply a root-level error.
2645  * Update the in-core space usage stats for this vdev, its metaslab class,
2646  * and the root vdev.
2662 	 * factor.  We must calculate this here and not at the root vdev
2663 	 * because the root vdev's psize-to-asize is simply the max of its
2695  * Mark a top-level vdev's config as dirty, placing it on the dirty list
2696  * so that it will be written out next time the vdev configuration is synced.
2697  * If the root vdev is specified (vdev_top == NULL), dirty all top-level vdevs.
2709 	 * If this is an aux vdev (as with l2cache and spare devices), then we
2710 	 * update the vdev config manually and set the sync flag.
2786  * Mark a top-level vdev's state as dirty, so that the next pass of
2827  * Propagate vdev state up from children to parent.
2852 				 * device, treat the root vdev as if it were
2870 		 * Root special: if there is a top-level vdev that cannot be
2872 		 * vdev's aux state as 'corrupt' rather than 'insufficient
2886  * Set a vdev's state.  If this is during an open, we don't update the parent
2910 	 * If we are setting the vdev state to anything but an open state, then
2924 	 * If we have brought this vdev back into service, we need
2930 	 * double-check the state of the vdev before repairing it.
2954 		 * If we fail to open a vdev during an import or recovery, we
3021  * Check the vdev configuration to ensure that it's capable of supporting
3023  * In addition, only a single top-level vdev is allowed and none of the leaves
3051  * Load the state from the original vdev tree (ovd) which
3053  * vdev was offline or faulted then we transfer that state to the
3054  * device in the current vdev tree (nvd).
3070 		 * Restore the persistent vdev state
3080  * Determine if a log device has valid content.  If the vdev was
3099  * Expand a vdev if possible.
3114  * Split a vdev.