2005-11-19 20:25:18 +00:00
|
|
|
/*-
|
2017-11-27 15:17:37 +00:00
|
|
|
* SPDX-License-Identifier: BSD-2-Clause-FreeBSD
|
|
|
|
*
|
2005-11-19 20:25:18 +00:00
|
|
|
* Copyright (c) 2005 Chris Jones
|
|
|
|
* All rights reserved.
|
2005-11-20 10:40:06 +00:00
|
|
|
*
|
|
|
|
* This software was developed for the FreeBSD Project by Chris Jones
|
|
|
|
* thanks to the support of Google's Summer of Code program and
|
|
|
|
* mentoring by Lukas Ertl.
|
|
|
|
*
|
2005-11-19 20:25:18 +00:00
|
|
|
* Redistribution and use in source and binary forms, with or without
|
|
|
|
* modification, are permitted provided that the following conditions
|
|
|
|
* are met:
|
|
|
|
* 1. Redistributions of source code must retain the above copyright
|
|
|
|
* notice, this list of conditions and the following disclaimer.
|
|
|
|
* 2. Redistributions in binary form must reproduce the above copyright
|
|
|
|
* notice, this list of conditions and the following disclaimer in the
|
|
|
|
* documentation and/or other materials provided with the distribution.
|
2005-11-20 10:40:06 +00:00
|
|
|
*
|
2005-11-19 20:25:18 +00:00
|
|
|
* THIS SOFTWARE IS PROVIDED BY AUTHOR AND CONTRIBUTORS ``AS IS'' AND
|
|
|
|
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
|
|
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
|
|
* ARE DISCLAIMED. IN NO EVENT SHALL AUTHOR OR CONTRIBUTORS BE LIABLE
|
|
|
|
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|
|
|
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
|
|
|
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
|
|
|
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
|
|
|
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
|
|
|
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
|
|
|
* SUCH DAMAGE.
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <sys/cdefs.h>
|
|
|
|
__FBSDID("$FreeBSD$");
|
|
|
|
|
|
|
|
#include <sys/libkern.h>
|
|
|
|
#include <sys/malloc.h>
|
|
|
|
|
|
|
|
#include <geom/geom.h>
|
2019-08-07 19:28:35 +00:00
|
|
|
#include <geom/geom_dbg.h>
|
2005-11-19 20:25:18 +00:00
|
|
|
#include <geom/vinum/geom_vinum_var.h>
|
|
|
|
#include <geom/vinum/geom_vinum.h>
|
|
|
|
|
|
|
|
void
|
|
|
|
gv_move(struct g_geom *gp, struct gctl_req *req)
|
|
|
|
{
|
|
|
|
struct gv_softc *sc;
|
|
|
|
struct gv_sd *s;
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
struct gv_drive *d;
|
2005-11-19 20:25:18 +00:00
|
|
|
char buf[20], *destination, *object;
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
int *argc, *flags, i, type;
|
2005-11-19 20:25:18 +00:00
|
|
|
|
|
|
|
sc = gp->softc;
|
|
|
|
|
|
|
|
argc = gctl_get_paraml(req, "argc", sizeof(*argc));
|
2008-11-25 19:13:58 +00:00
|
|
|
if (argc == NULL) {
|
|
|
|
gctl_error(req, "no arguments given");
|
|
|
|
return;
|
|
|
|
}
|
2005-11-19 20:25:18 +00:00
|
|
|
flags = gctl_get_paraml(req, "flags", sizeof(*flags));
|
2008-11-25 19:13:58 +00:00
|
|
|
if (flags == NULL) {
|
|
|
|
gctl_error(req, "no flags given");
|
|
|
|
return;
|
|
|
|
}
|
2005-11-19 20:25:18 +00:00
|
|
|
destination = gctl_get_param(req, "destination", NULL);
|
|
|
|
if (destination == NULL) {
|
|
|
|
gctl_error(req, "no destination given");
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
if (gv_object_type(sc, destination) != GV_TYPE_DRIVE) {
|
|
|
|
gctl_error(req, "destination '%s' is not a drive", destination);
|
|
|
|
return;
|
|
|
|
}
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
d = gv_find_drive(sc, destination);
|
2005-11-19 20:25:18 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* We start with 1 here, because argv[0] on the command line is the
|
|
|
|
* destination drive.
|
|
|
|
*/
|
|
|
|
for (i = 1; i < *argc; i++) {
|
|
|
|
snprintf(buf, sizeof(buf), "argv%d", i);
|
|
|
|
object = gctl_get_param(req, buf, NULL);
|
|
|
|
if (object == NULL)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
type = gv_object_type(sc, object);
|
|
|
|
if (type != GV_TYPE_SD) {
|
|
|
|
gctl_error(req, "you can only move subdisks; "
|
2009-10-05 08:44:31 +00:00
|
|
|
"'%s' is not a subdisk", object);
|
2005-11-19 20:25:18 +00:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
s = gv_find_sd(sc, object);
|
|
|
|
if (s == NULL) {
|
|
|
|
gctl_error(req, "unknown subdisk '%s'", object);
|
|
|
|
return;
|
|
|
|
}
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
gv_post_event(sc, GV_EVENT_MOVE_SD, s, d, *flags, 0);
|
2005-11-19 20:25:18 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Move a subdisk. */
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
int
|
|
|
|
gv_move_sd(struct gv_softc *sc, struct gv_sd *cursd,
|
|
|
|
struct gv_drive *destination, int flags)
|
2005-11-19 20:25:18 +00:00
|
|
|
{
|
|
|
|
struct gv_drive *d;
|
|
|
|
struct gv_sd *newsd, *s, *s2;
|
|
|
|
struct gv_plex *p;
|
2005-11-20 10:40:06 +00:00
|
|
|
int err;
|
2005-11-19 20:25:18 +00:00
|
|
|
|
2005-11-20 10:40:06 +00:00
|
|
|
g_topology_assert();
|
|
|
|
KASSERT(cursd != NULL, ("gv_move_sd: NULL cursd"));
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
KASSERT(destination != NULL, ("gv_move_sd: NULL destination"));
|
2005-11-19 20:25:18 +00:00
|
|
|
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
d = cursd->drive_sc;
|
2005-11-19 20:25:18 +00:00
|
|
|
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
if ((gv_consumer_is_open(d->consumer) ||
|
|
|
|
gv_consumer_is_open(destination->consumer)) &&
|
2010-10-01 06:12:13 +00:00
|
|
|
!(flags & GV_FLAG_F)) {
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
G_VINUM_DEBUG(0, "consumers on current and destination drive "
|
|
|
|
" still open");
|
|
|
|
return (GV_ERR_ISBUSY);
|
2005-11-19 20:25:18 +00:00
|
|
|
}
|
|
|
|
|
2010-10-01 06:12:13 +00:00
|
|
|
if (!(flags & GV_FLAG_F)) {
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
G_VINUM_DEBUG(1, "-f flag not passed; move would be "
|
2005-11-19 20:25:18 +00:00
|
|
|
"destructive");
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
return (GV_ERR_INVFLAG);
|
2005-11-19 20:25:18 +00:00
|
|
|
}
|
|
|
|
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
if (destination == cursd->drive_sc) {
|
|
|
|
G_VINUM_DEBUG(1, "subdisk '%s' already on drive '%s'",
|
|
|
|
cursd->name, destination->name);
|
|
|
|
return (GV_ERR_ISATTACHED);
|
2005-11-19 20:25:18 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/* XXX: Does it have to be part of a plex? */
|
|
|
|
p = gv_find_plex(sc, cursd->plex);
|
|
|
|
if (p == NULL) {
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
G_VINUM_DEBUG(0, "subdisk '%s' is not part of a plex",
|
2005-11-19 20:25:18 +00:00
|
|
|
cursd->name);
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
return (GV_ERR_NOTFOUND);
|
2005-11-19 20:25:18 +00:00
|
|
|
}
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
|
2005-11-19 20:25:18 +00:00
|
|
|
/* Stale the old subdisk. */
|
|
|
|
err = gv_set_sd_state(cursd, GV_SD_STALE,
|
|
|
|
GV_SETSTATE_FORCE | GV_SETSTATE_CONFIG);
|
|
|
|
if (err) {
|
2009-10-05 08:44:31 +00:00
|
|
|
G_VINUM_DEBUG(0, "unable to set the subdisk '%s' to state "
|
2005-11-19 20:25:18 +00:00
|
|
|
"'stale'", cursd->name);
|
|
|
|
return (err);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Create new subdisk. Ideally, we'd use gv_new_sd, but that requires
|
|
|
|
* us to create a string for it to parse, which is silly.
|
|
|
|
* TODO: maybe refactor gv_new_sd such that this is no longer the case.
|
|
|
|
*/
|
|
|
|
newsd = g_malloc(sizeof(struct gv_sd), M_WAITOK | M_ZERO);
|
|
|
|
newsd->plex_offset = cursd->plex_offset;
|
2005-11-20 10:40:06 +00:00
|
|
|
newsd->size = cursd->size;
|
2005-11-19 20:25:18 +00:00
|
|
|
newsd->drive_offset = -1;
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
strlcpy(newsd->name, cursd->name, sizeof(newsd->name));
|
|
|
|
strlcpy(newsd->drive, destination->name, sizeof(newsd->drive));
|
|
|
|
strlcpy(newsd->plex, cursd->plex, sizeof(newsd->plex));
|
2005-11-19 20:25:18 +00:00
|
|
|
newsd->state = GV_SD_STALE;
|
|
|
|
newsd->vinumconf = cursd->vinumconf;
|
|
|
|
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
err = gv_sd_to_drive(newsd, destination);
|
2005-11-19 20:25:18 +00:00
|
|
|
if (err) {
|
|
|
|
/* XXX not enough free space? */
|
|
|
|
g_free(newsd);
|
|
|
|
return (err);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Replace the old sd by the new one. */
|
|
|
|
LIST_FOREACH_SAFE(s, &p->subdisks, in_plex, s2) {
|
|
|
|
if (s == cursd) {
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
gv_rm_sd(sc, s);
|
2005-11-19 20:25:18 +00:00
|
|
|
}
|
|
|
|
}
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
gv_sd_to_plex(newsd, p);
|
2005-11-19 20:25:18 +00:00
|
|
|
LIST_INSERT_HEAD(&sc->subdisks, newsd, sd);
|
Import the gvinum work that have been done during and after Summer of Code 2007.
The work have been under testing and fixing since then, and it is mature enough
to be put into HEAD for further testing.
A lot have changed in this time, and here are the most important:
- Gvinum now uses one single workerthread instead of one thread for each
volume and each plex. The reason for this is that the previous scheme was
very complex, and was the cause of many of the bugs discovered in gvinum.
Instead, gvinum now uses one worker thread with an event queue, quite
similar to what used in gmirror.
- The rebuild/grow/initialize/parity check routines no longer runs in
separate threads, but are run as regular I/O requests with special flags.
This made it easier to support mounted growing and parity rebuild.
- Support for growing striped and raid5-plexes, meaning that one can extend the
volumes for these plex types in addition to the concat type. Also works while
the volume is mounted.
- Implementation of many of the missing commands from the old vinum:
attach/detach, start (was partially implemented), stop (was partially
implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes
with one plex of these organizations).
- The parity check and rebuild no longer goes between userland/kernel, meaning
that the gvinum command will not stay and wait forever for the rebuild to
finish. You can instead watch the status with the list command.
- Many problems with gvinum have been reported since 5.x, and some has been hard
to fix due to the complicated architecture. Hopefully, it should be more
stable and better handle edge cases that previously made gvinum crash.
- Failed drives no longer disappears entirely, but now leave behind a dummy
drive that makes sure the original state is not forgotten in case the system
is rebooted between drive failures/swaps.
- Update manpage to reflect new commands and extend it with some examples.
Sponsored by: Google Summer of Code 2007
Mentored by: le
Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
|
|
|
/* Update volume size of plex. */
|
|
|
|
if (p->vol_sc != NULL)
|
|
|
|
gv_update_vol_size(p->vol_sc, gv_vol_size(p->vol_sc));
|
|
|
|
gv_save_config(p->vinumconf);
|
2005-11-19 20:25:18 +00:00
|
|
|
return (0);
|
|
|
|
}
|