OpenZFS 8962 - zdb should work on non-idle pools

Currently `zdb` consistently fails to examine non-idle pools as it
fails during the `spa_load()` process. The main problem seems to be
that `spa_load_verify()` fails as can be seen below:

    $ sudo zdb -d -G dcenter
    zdb: can't open 'dcenter': I/O error

    ZFS_DBGMSG(zdb):
    spa_open_common: opening dcenter
    spa_load(dcenter): LOADING
    disk vdev '/dev/dsk/c4t11d0s0': best uberblock found for spa dcenter. txg 40824950
    spa_load(dcenter): using uberblock with txg=40824950
    spa_load(dcenter): UNLOADING
    spa_load(dcenter): RELOADING
    spa_load(dcenter): LOADING
    disk vdev '/dev/dsk/c3t10d0s0': best uberblock found for spa dcenter. txg 40824952
    spa_load(dcenter): using uberblock with txg=40824952
    spa_load(dcenter): FAILED: spa_load_verify failed [error=5]
    spa_load(dcenter): UNLOADING

This change makes `spa_load_verify()` a dryrun when ran from
`zdb`. This is done by creating a global flag in zfs and then setting
it in `zdb`.

Authored by: Pavel Zakharov <[email protected]>
Reviewed by: George Wilson <[email protected]>
Reviewed by: Matthew Ahrens <[email protected]>
Reviewed by: Andy Stormont <[email protected]>
Approved by: Dan McDonald <[email protected]>
Ported-by: Tim Chase <[email protected]>
Signed-off-by: Tim Chase <[email protected]>

OpenZFS-issue: https://illumos.org/issues/8962
OpenZFS-commit: https://github.com/openzfs/openzfs/commit/180ad792
Closes #7459

2 files changed, 25 insertions, 2 deletions

diff --git a/cmd/zdb/zdb.c b/cmd/zdb/zdb.c
index 5f56b10e2..a7bd64ecd 100644
--- a/cmd/zdb/zdb.c
+++ b/cmd/zdb/zdb.c
@@ -96,6 +96,7 @@ extern int reference_tracking_enable;
 extern int zfs_recover;
 extern uint64_t zfs_arc_max, zfs_arc_meta_limit;
 extern int zfs_vdev_async_read_max_active;
+extern boolean_t spa_load_verify_dryrun;
 
 static const char cmdname[] = "zdb";
 uint8_t dump_opt[256];
@@ -5009,6 +5010,12 @@ main(int argc, char **argv)
 	 */
 	reference_tracking_enable = B_FALSE;
 
+	/*
+	 * Do not fail spa_load when spa_load_verify fails. This is needed
+	 * to load non-idle pools.
+	 */
+	spa_load_verify_dryrun = B_TRUE;
+
 	kernel_init(FREAD);
 	if ((g_zfs = libzfs_init()) == NULL) {
 		(void) fprintf(stderr, "%s", libzfs_error_init(errno));
diff --git a/module/zfs/spa.c b/module/zfs/spa.c
index a17da4239..80f0c6f36 100644
--- a/module/zfs/spa.c
+++ b/module/zfs/spa.c
@@ -170,6 +170,12 @@ uint_t		zio_taskq_basedc = 80;		/* base duty cycle */
 boolean_t	spa_create_process = B_TRUE;	/* no process ==> no sysdc */
 
 /*
+ * Report any spa_load_verify errors found, but do not fail spa_load.
+ * This is used by zdb to analyze non-idle pools.
+ */
+boolean_t	spa_load_verify_dryrun = B_FALSE;
+
+/*
  * This (illegal) pool name is used when temporarily importing a spa_t in order
  * to get the vdev stats associated with the imported devices.
  */
@@ -2140,8 +2146,15 @@ spa_load_verify(spa_t *spa)
 	spa->spa_load_meta_errors = sle.sle_meta_count;
 	spa->spa_load_data_errors = sle.sle_data_count;
 
-	if (!error && sle.sle_meta_count <= policy.zrp_maxmeta &&
-	    sle.sle_data_count <= policy.zrp_maxdata) {
+	if (sle.sle_meta_count != 0 || sle.sle_data_count != 0) {
+		spa_load_note(spa, "spa_load_verify found %llu metadata errors "
+		    "and %llu data errors", (u_longlong_t)sle.sle_meta_count,
+		    (u_longlong_t)sle.sle_data_count);
+	}
+
+	if (spa_load_verify_dryrun ||
+	    (!error && sle.sle_meta_count <= policy.zrp_maxmeta &&
+	    sle.sle_data_count <= policy.zrp_maxdata)) {
 		int64_t loss = 0;
 
 		verify_ok = B_TRUE;
@@ -2159,6 +2172,9 @@ spa_load_verify(spa_t *spa)
 		spa->spa_load_max_txg = spa->spa_uberblock.ub_txg;
 	}
 
+	if (spa_load_verify_dryrun)
+		return (0);
+
 	if (error) {
 		if (error != ENXIO && error != EIO)
 			error = SET_ERROR(EIO);