1
0
mirror of git://sourceware.org/git/lvm2.git synced 2024-12-21 13:34:40 +03:00
lvm2/lib/misc/lvm-globals.c

395 lines
7.4 KiB
C
Raw Normal View History

/*
* Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
* Copyright (C) 2004-2007 Red Hat, Inc. All rights reserved.
*
* This file is part of LVM2.
*
* This copyrighted material is made available to anyone wishing to use,
* modify, copy, or redistribute it subject to the terms and conditions
* of the GNU Lesser General Public License v.2.1.
*
* You should have received a copy of the GNU Lesser General Public License
* along with this program; if not, write to the Free Software Foundation,
* Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "lib/misc/lib.h"
#include "lib/device/device.h"
#include "lib/misc/lvm-string.h"
#include "lib/config/defaults.h"
#include "lib/metadata/metadata-exported.h"
#include <stdarg.h>
static int _verbose_level = VERBOSE_BASE_LEVEL;
static int _silent = 0;
static int _test = 0;
static int _use_aio = 0;
static int _md_filtering = 0;
static int _internal_filtering = 0;
static int _fwraid_filtering = 0;
static int _pvmove = 0;
static int _obtain_device_list_from_udev = DEFAULT_OBTAIN_DEVICE_LIST_FROM_UDEV;
static enum dev_ext_e _external_device_info_source = DEV_EXT_NONE;
static int _debug_level = 0;
static int _debug_classes_logged = 0;
static int _security_level = SECURITY_LEVEL;
static char _log_command_info[40] = "";
static char _log_command_file[40] = "";
static char _cmd_name[30] = "none";
static int _mirror_in_sync = 0;
static int _dmeventd_monitor = DEFAULT_DMEVENTD_MONITOR;
/* When set, disables update of _dmeventd_monitor & _ignore_suspended_devices */
static int _disable_dmeventd_monitoring = 0;
static int _background_polling = DEFAULT_BACKGROUND_POLLING;
static int _ignore_suspended_devices = 0;
Mirror: Fix hangs and lock-ups caused by attempting label reads of mirrors There is a problem with the way mirrors have been designed to handle failures that is resulting in stuck LVM processes and hung I/O. When mirrors encounter a write failure, they block I/O and notify userspace to reconfigure the mirror to remove failed devices. This process is open to a couple races: 1) Any LVM process other than the one that is meant to deal with the mirror failure can attempt to read the mirror, fail, and block other LVM commands (including the repair command) from proceeding due to holding a lock on the volume group. 2) If there are multiple mirrors that suffer a failure in the same volume group, a repair can block while attempting to read the LVM label from one mirror while trying to repair the other. Mitigation of these races has been attempted by disallowing label reading of mirrors that are either suspended or are indicated as blocking by the kernel. While this has closed the window of opportunity for hitting the above problems considerably, it hasn't closed it completely. This is because it is still possible to start an LVM command, read the status of the mirror as healthy, and then perform the read for the label at the moment after a the failure is discovered by the kernel. I can see two solutions to this problem: 1) Allow users to configure whether mirrors can be candidates for LVM labels (i.e. whether PVs can be created on mirror LVs). If the user chooses to allow label scanning of mirror LVs, it will be at the expense of a possible hang in I/O or LVM processes. 2) Instrument a way to allow asynchronous label reading - allowing blocked label reads to be ignored while continuing to process the LVM command. This would action would allow LVM commands to continue even though they would have otherwise blocked trying to read a mirror. They can then release their lock and allow a repair command to commence. In the event of #2 above, the repair command already in progress can continue and repair the failed mirror. This patch brings solution #1. If solution #2 is developed later on, the configuration option created in #1 can be negated - allowing mirrors to be scanned for labels by default once again.
2013-10-23 04:14:33 +04:00
static int _ignore_lvm_mirrors = DEFAULT_IGNORE_LVM_MIRRORS;
static int _error_message_produced = 0;
static unsigned _is_static = 0;
static int _udev_checking = 1;
static int _retry_deactivation = DEFAULT_RETRY_DEACTIVATION;
static int _activation_checks = 0;
static char _sysfs_dir_path[PATH_MAX] = "";
static uint64_t _pv_min_size = (DEFAULT_PV_MIN_SIZE_KB * 1024L >> SECTOR_SHIFT);
static const char *_unknown_device_name = DEFAULT_UNKNOWN_DEVICE_NAME;
static int _io_memory_size_kb = DEFAULT_IO_MEMORY_SIZE_KB;
void init_verbose(int level)
{
_verbose_level = level;
}
void init_silent(int silent)
{
_silent = silent;
}
void init_test(int level)
{
if (!_test && level)
log_warn("TEST MODE: Metadata will NOT be updated and volumes will not be (de)activated.");
_test = level;
}
void init_use_aio(int useaio)
{
_use_aio = useaio;
}
void init_md_filtering(int level)
{
_md_filtering = level;
}
void init_internal_filtering(int level)
{
_internal_filtering = level;
}
void init_fwraid_filtering(int level)
{
_fwraid_filtering = level;
}
void init_pvmove(int level)
{
_pvmove = level;
}
void init_obtain_device_list_from_udev(int device_list_from_udev)
{
_obtain_device_list_from_udev = device_list_from_udev;
}
void init_external_device_info_source(enum dev_ext_e src)
{
_external_device_info_source = src;
}
void init_security_level(int level)
{
_security_level = level;
}
void init_mirror_in_sync(int in_sync)
{
_mirror_in_sync = in_sync;
}
void init_dmeventd_monitor(int reg)
{
if (!_disable_dmeventd_monitoring)
_dmeventd_monitor = reg;
}
void init_disable_dmeventd_monitoring(int reg)
{
_disable_dmeventd_monitoring = reg;
}
void init_background_polling(int polling)
{
_background_polling = polling;
}
void init_ignore_suspended_devices(int ignore)
{
if (!_disable_dmeventd_monitoring)
_ignore_suspended_devices = ignore;
}
Mirror: Fix hangs and lock-ups caused by attempting label reads of mirrors There is a problem with the way mirrors have been designed to handle failures that is resulting in stuck LVM processes and hung I/O. When mirrors encounter a write failure, they block I/O and notify userspace to reconfigure the mirror to remove failed devices. This process is open to a couple races: 1) Any LVM process other than the one that is meant to deal with the mirror failure can attempt to read the mirror, fail, and block other LVM commands (including the repair command) from proceeding due to holding a lock on the volume group. 2) If there are multiple mirrors that suffer a failure in the same volume group, a repair can block while attempting to read the LVM label from one mirror while trying to repair the other. Mitigation of these races has been attempted by disallowing label reading of mirrors that are either suspended or are indicated as blocking by the kernel. While this has closed the window of opportunity for hitting the above problems considerably, it hasn't closed it completely. This is because it is still possible to start an LVM command, read the status of the mirror as healthy, and then perform the read for the label at the moment after a the failure is discovered by the kernel. I can see two solutions to this problem: 1) Allow users to configure whether mirrors can be candidates for LVM labels (i.e. whether PVs can be created on mirror LVs). If the user chooses to allow label scanning of mirror LVs, it will be at the expense of a possible hang in I/O or LVM processes. 2) Instrument a way to allow asynchronous label reading - allowing blocked label reads to be ignored while continuing to process the LVM command. This would action would allow LVM commands to continue even though they would have otherwise blocked trying to read a mirror. They can then release their lock and allow a repair command to commence. In the event of #2 above, the repair command already in progress can continue and repair the failed mirror. This patch brings solution #1. If solution #2 is developed later on, the configuration option created in #1 can be negated - allowing mirrors to be scanned for labels by default once again.
2013-10-23 04:14:33 +04:00
void init_ignore_lvm_mirrors(int scan)
{
_ignore_lvm_mirrors = scan;
}
void init_log_command(int log_name, int log_pid)
{
memset(_log_command_info, 0, sizeof(_log_command_info));
memset(_log_command_file, 0, sizeof(_log_command_file));
/*
* Always include command name and pid in file and verbose output.
*/
(void) dm_snprintf(_log_command_file, sizeof(_log_command_file), "%s[%d]",
_cmd_name, getpid());
/*
* This is the prefix that can be configured for each line of stdout.
*/
if (!log_name && !log_pid)
return;
else if (log_name && !log_pid)
(void) dm_strncpy(_log_command_info, _cmd_name, sizeof(_log_command_info));
else if (!log_name && log_pid)
(void) dm_snprintf(_log_command_info, sizeof(_log_command_info), "%d", getpid());
else
(void) dm_snprintf(_log_command_info, sizeof(_log_command_info), "%s[%d]",
_cmd_name, getpid());
}
void init_is_static(unsigned value)
{
_is_static = value;
}
void init_udev_checking(int checking)
{
if ((_udev_checking = checking))
log_debug_activation("LVM udev checking enabled");
else
log_debug_activation("LVM udev checking disabled");
}
void init_retry_deactivation(int retry)
{
_retry_deactivation = retry;
}
void init_activation_checks(int checks)
{
if ((_activation_checks = checks))
log_debug_activation("LVM activation checks enabled");
else
log_debug_activation("LVM activation checks disabled");
}
void init_pv_min_size(uint64_t sectors)
{
_pv_min_size = sectors;
}
void set_cmd_name(const char *cmd)
{
(void) dm_strncpy(_cmd_name, cmd, sizeof(_cmd_name));
}
const char *get_cmd_name(void)
{
return _cmd_name;
}
void set_sysfs_dir_path(const char *path)
{
(void) dm_strncpy(_sysfs_dir_path, path, sizeof(_sysfs_dir_path));
}
const char *log_command_info(void)
{
return _log_command_info;
}
const char *log_command_file(void)
{
return _log_command_file;
}
void init_error_message_produced(int value)
{
_error_message_produced = value;
}
int error_message_produced(void)
{
return _error_message_produced;
}
int test_mode(void)
{
return _test;
}
int use_aio(void)
{
return _use_aio;
}
int md_filtering(void)
{
return _md_filtering;
}
int internal_filtering(void)
{
return _internal_filtering;
}
int fwraid_filtering(void)
{
return _fwraid_filtering;
}
int pvmove_mode(void)
{
return _pvmove;
}
int obtain_device_list_from_udev(void)
{
return _obtain_device_list_from_udev;
}
enum dev_ext_e external_device_info_source(void)
{
return _external_device_info_source;
}
int background_polling(void)
{
return _background_polling;
}
int security_level(void)
{
return _security_level;
}
int mirror_in_sync(void)
{
return _mirror_in_sync;
}
int dmeventd_monitor_mode(void)
{
return _dmeventd_monitor;
}
int ignore_suspended_devices(void)
{
return _ignore_suspended_devices;
}
Mirror: Fix hangs and lock-ups caused by attempting label reads of mirrors There is a problem with the way mirrors have been designed to handle failures that is resulting in stuck LVM processes and hung I/O. When mirrors encounter a write failure, they block I/O and notify userspace to reconfigure the mirror to remove failed devices. This process is open to a couple races: 1) Any LVM process other than the one that is meant to deal with the mirror failure can attempt to read the mirror, fail, and block other LVM commands (including the repair command) from proceeding due to holding a lock on the volume group. 2) If there are multiple mirrors that suffer a failure in the same volume group, a repair can block while attempting to read the LVM label from one mirror while trying to repair the other. Mitigation of these races has been attempted by disallowing label reading of mirrors that are either suspended or are indicated as blocking by the kernel. While this has closed the window of opportunity for hitting the above problems considerably, it hasn't closed it completely. This is because it is still possible to start an LVM command, read the status of the mirror as healthy, and then perform the read for the label at the moment after a the failure is discovered by the kernel. I can see two solutions to this problem: 1) Allow users to configure whether mirrors can be candidates for LVM labels (i.e. whether PVs can be created on mirror LVs). If the user chooses to allow label scanning of mirror LVs, it will be at the expense of a possible hang in I/O or LVM processes. 2) Instrument a way to allow asynchronous label reading - allowing blocked label reads to be ignored while continuing to process the LVM command. This would action would allow LVM commands to continue even though they would have otherwise blocked trying to read a mirror. They can then release their lock and allow a repair command to commence. In the event of #2 above, the repair command already in progress can continue and repair the failed mirror. This patch brings solution #1. If solution #2 is developed later on, the configuration option created in #1 can be negated - allowing mirrors to be scanned for labels by default once again.
2013-10-23 04:14:33 +04:00
int ignore_lvm_mirrors(void)
{
return _ignore_lvm_mirrors;
}
void init_debug(int level)
{
_debug_level = level;
}
void init_debug_classes_logged(int classes)
{
_debug_classes_logged = classes;
}
int debug_class_is_logged(int class)
{
/* If no class given, log it */
if (!class)
return 1;
return (_debug_classes_logged & class) ? 1 : 0;
}
int verbose_level(void)
{
return _verbose_level;
}
int debug_level(void)
{
return _debug_level;
}
int silent_mode(void)
{
return _silent;
}
unsigned is_static(void)
{
return _is_static;
}
int udev_checking(void)
{
return _udev_checking;
}
int retry_deactivation(void)
{
return _retry_deactivation;
}
int activation_checks(void)
{
return _activation_checks;
}
const char *sysfs_dir_path(void)
{
return _sysfs_dir_path;
}
uint64_t pv_min_size(void)
{
return _pv_min_size;
}
const char *unknown_device_name(void)
{
return _unknown_device_name;
}
void init_unknown_device_name(const char *name)
{
_unknown_device_name = name;
}
int io_memory_size(void)
{
return _io_memory_size_kb;
}
void init_io_memory_size(int val)
{
_io_memory_size_kb = val;
}