Illumos 4891 - want zdb option to dump all metadata

4891 want zdb option to dump all metadata
Reviewed by: Sonu Pillai <[email protected]>
Reviewed by: George Wilson <[email protected]>
Reviewed by: Christopher Siden <[email protected]>
Reviewed by: Dan McDonald <[email protected]>
Reviewed by: Richard Lowe <[email protected]>
Approved by: Garrett D'Amore <[email protected]>

We'd like a way for zdb to dump metadata in a machine-readable
format, so that we can bring that back from a customer site for
in-house diagnosis.  Think of it as a crash dump for zpools,
which can be used for post-mortem analysis of a malfunctioning
pool

References:
  https://www.illumos.org/issues/4891
  https://github.com/illumos/illumos-gate/commit/df15e41

Porting notes:
- [cmd/zdb/zdb.c]
  - a5778ea zdb: Introduce -V for verbatim import
  - In main() getopt 'opt' variable removed and the code was
    brought back in line with illumos.
- [lib/libzpool/kernel.c]
  - 1e33ac1 Fix Solaris thread dependency by using pthreads
  - f0e324f Update utsname support
  - 4d58b69 Fix vn_open/vn_rdwr error handling
  - In vn_open() allocate 'dumppath' on heap instead of stack
  - Properly handle 'dump_fd == -1' error path
  - Free 'realpath' after added vn_dumpdir_code block

Ported-by: kernelOfTruth [email protected]
Signed-off-by: Brian Behlendorf <[email protected]>

1 files changed, 33 insertions, 2 deletions

diff --git a/lib/libzpool/kernel.c b/lib/libzpool/kernel.c
index a45102699..31d3336a8 100644
--- a/lib/libzpool/kernel.c
+++ b/lib/libzpool/kernel.c
@@ -29,6 +29,7 @@
 #include <stdlib.h>
 #include <string.h>
 #include <zlib.h>
+#include <libgen.h>
 #include <sys/signal.h>
 #include <sys/spa.h>
 #include <sys/stat.h>
@@ -50,6 +51,9 @@ char hw_serial[HW_HOSTID_LEN];
 struct utsname hw_utsname;
 vmem_t *zio_arena = NULL;
 
+/* If set, all blocks read will be copied to the specified directory. */
+char *vn_dumpdir = NULL;
+
 /* this only exists to have its address taken */
 struct proc p0;
 
@@ -588,6 +592,7 @@ int
 vn_open(char *path, int x1, int flags, int mode, vnode_t **vpp, int x2, int x3)
 {
 	int fd;
+	int dump_fd;
 	vnode_t *vp;
 	int old_umask = 0;
 	char *realpath;
@@ -655,13 +660,31 @@ vn_open(char *path, int x1, int flags, int mode, vnode_t **vpp, int x2, int x3)
 	 * FREAD and FWRITE to the corresponding O_RDONLY, O_WRONLY, and O_RDWR.
 	 */
 	fd = open64(realpath, flags - FREAD, mode);
-	free(realpath);
+	err = errno;
 
 	if (flags & FCREAT)
 		(void) umask(old_umask);
 
+	if (vn_dumpdir != NULL) {
+		char *dumppath = umem_zalloc(MAXPATHLEN, UMEM_NOFAIL);
+		(void) snprintf(dumppath, MAXPATHLEN,
+		    "%s/%s", vn_dumpdir, basename(realpath));
+		dump_fd = open64(dumppath, O_CREAT | O_WRONLY, 0666);
+		umem_free(dumppath, MAXPATHLEN);
+		if (dump_fd == -1) {
+			err = errno;
+			free(realpath);
+			close(fd);
+			return (err);
+		}
+	} else {
+		dump_fd = -1;
+	}
+
+	free(realpath);
+
 	if (fd == -1)
-		return (errno);
+		return (err);
 
 	if (fstat64_blk(fd, &st) == -1) {
 		err = errno;
@@ -676,6 +699,7 @@ vn_open(char *path, int x1, int flags, int mode, vnode_t **vpp, int x2, int x3)
 	vp->v_fd = fd;
 	vp->v_size = st.st_size;
 	vp->v_path = spa_strdup(path);
+	vp->v_dump_fd = dump_fd;
 
 	return (0);
 }
@@ -708,6 +732,11 @@ vn_rdwr(int uio, vnode_t *vp, void *addr, ssize_t len, offset_t offset,
 
 	if (uio == UIO_READ) {
 		rc = pread64(vp->v_fd, addr, len, offset);
+		if (vp->v_dump_fd != -1) {
+			int status =
+			    pwrite64(vp->v_dump_fd, addr, rc, offset);
+			ASSERT(status != -1);
+		}
 	} else {
 		/*
 		 * To simulate partial disk writes, we split writes into two
@@ -750,6 +779,8 @@ void
 vn_close(vnode_t *vp)
 {
 	close(vp->v_fd);
+	if (vp->v_dump_fd != -1)
+		close(vp->v_dump_fd);
 	spa_strfree(vp->v_path);
 	umem_free(vp, sizeof (vnode_t));
 }