1
0
mirror of git://sourceware.org/git/lvm2.git synced 2024-12-29 15:22:30 +03:00
Commit Graph

280 Commits

Author SHA1 Message Date
Zdenek Kabelac
fc9d4dd11f config: use int for type
Since the type is used for 'or' operation of enumerated bit fields,
it doesn't not have type cfg_def_type_t - use proper int type for
bitmask.
2013-11-28 12:42:44 +01:00
Peter Rajnoha
ab2f858af7 conf: add allocation/use_blkid_wiping
Add allocation/use_blkid_wiping setting to lvm.conf to select between
LVM2 native code to detect signatures to wipe or blkid library code.
2013-11-27 15:49:14 +01:00
Peter Rajnoha
5b7e543cae conf: add allocation/wipe_signatures_on_new_logical_volumes_when_zeroing
This setting controls whether signature wiping on newly created logical
volumes will follow the state of zeroing (-Z/--zero option).
2013-11-27 15:48:06 +01:00
Jonathan Brassow
d5896f0afd 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  above, the repair command already in progress can continue
and repair the failed mirror.

This patch brings solution .  If solution  is developed later on, the
configuration option created in  can be negated - allowing mirrors to
be scanned for labels by default once again.
2013-10-22 19:14:33 -05:00
Alasdair G Kergon
04d9a52684 release 2.02.103
52 files changed, 598 insertions(+), 264 deletions(-)
2013-10-04 14:32:23 +01:00
Peter Rajnoha
8cf0810d57 thin: rename thin_pool_chunk_size_calculation -> ..size_policy and rename "default" policy to "generic"
Just to be consistent with existing naming we use.
2013-10-04 12:30:33 +02:00
Peter Rajnoha
b4637bd298 fix: make it possible to compile with --disable-devmapper again
Some code has been added recently which makes it impossible to compile
when "configure --disable-devmapper" is used. This patch just shuffles
the code around so it's under proper #ifdef DEVMAPPER_SUPPORT.
2013-09-27 13:58:55 +02:00
Peter Rajnoha
cc9e65c391 thin: use appropriate default value based on allocation/thin_pool_chunk_size_calculation setting
If thin_pool_chunk_size_calculation is set to "default", use 64KiB,
otheriwse 512KiB for "performance".
2013-09-25 16:06:38 +02:00
Peter Rajnoha
8bf425005c conf: add allocation/thin_pool_chunk_size_calculation
Add allocation/thin_pool_chunk_size_calculation lvm.conf
option to select a method for calculating thin pool chunk
sizes and define two possible values - "default" and "performance".
2013-09-25 16:06:38 +02:00
Alasdair G Kergon
a3a5f58c21 reporting: Add devtypes command.
Add internal devtypes reporting command to display built-in recognised
block device types.  (The output does not include any additional
types added by a configuration file.)

> lvm devtypes -o help
  Device Types Fields
  -------------------
    devtype_all            - All fields in this section.
    devtype_name           - Name of Device Type exactly as it appears in /proc/devices.
    devtype_max_partitions - Maximum number of partitions. (How many device minor numbers get reserved for each device.)
    devtype_description    - Description of Device Type.

> lvm devtypes
  DevType       MaxParts Description
  aoe                 16 ATA over Ethernet
  ataraid             16 ATA Raid
  bcache               1 bcache block device cache
  blkext               1 Extended device partitions
...
2013-09-18 01:09:15 +01:00
Peter Rajnoha
2f61478436 workaround: gcc v4.8 on 32 bit param. passing bug when -02 opimization used
gcc -O2 v4.8 on 32 bit architecture is causing a bug in parameter
passing. It does not happen with -01 nor -O0.

The problematic part of the code was strlen use in config.c in
the config_def_check fn and the call for _config_def_check_tree in it:

<snip>
  rplen = strlen(rp);
  if (!_config_def_check_tree(handle, vp, vp + strlen(vp), rp, rp + rplen, CFG_PATH_MAX_LEN - rplen, cn, cmd->cft_def_hash)) ...
</snip>

If compiled with -O0 (correct):

Breakpoint 1, config_def_check (cmd=0x819b050, handle=0x81a04f8) at config/config.c:775
(gdb) p	vp
$1 = 0x8189ee0 <_cfg_path> "config"
(gdb) p	strlen(vp)
$2 = 6
(gdb)
_config_def_check_tree (handle=0x81a04f8, vp=0x8189ee0 <_cfg_path>
"config", pvp=0x8189ee6 <_cfg_path+6> "", rp=0xbfffe1e8 "config",
prp=0xbfffe1ee "", buf_size=58, root=0x81a2568, ht=0x81a65
48) at config/config.c:680
(gdb) p	vp
$4 = 0x8189ee0 <_cfg_path> "config"
(gdb) p	pvp
$5 = 0x8189ee6 <_cfg_path+6> ""

If compiled with -O2 (incorrect):

Breakpoint 1, config_def_check (cmd=cmd@entry=0x8183050, handle=0x81884f8) at config/config.c:775
(gdb) p	vp
$1 = 0x8172fc0 <_cfg_path> "config"
(gdb) p strlen(vp)
$2 = 6
(gdb) p	vp + strlen(vp)
$3 = 0x8172fc6 <_cfg_path+6> ""
(gdb)
_config_def_check_tree (handle=handle@entry=0x81884f8, pvp=0x8172fc7
<_cfg_path+7> "host_list", rp=rp@entry=0xbffff190 "config",
prp=prp@entry=0xbffff196 "", buf_size=buf_size@entry=58, ht=0x
818e548, root=0x818a568, vp=0x8172fc0 <_cfg_path> "config") at
config/config.c:674
(gdb) p	pvp
$4 = 0x8172fc7 <_cfg_path+7> "host_list"

The difference is in passing the "pvp" arg for _config_def_check_tree.
While in the correct case, the value of _cfg_path+6 is passed
(the result of vp + strlen(vp) - see the snippet of the code above),
in the incorrect case, this value is increased by 1 to _cfg_path+7,
hence totally malforming the string that is being processed.

This ends up with incorrect validation check and incorrect warning
messages are issued like:

 "Configuration setting "config/checks" has invalid type. Found integer, expected section."

To workaround this issue, remove the "static" qualifier from the
"static char _cfg_path[CFG_PATH_MAX_LEN]". This causes the optimalizer
to be less aggressive (also shuffling the arg list for
_config_def_check_tree call helps).
2013-08-09 13:24:50 +02:00
Jonathan Brassow
c13d1b11b2 RAID: Make "raid10" the default striped + mirror segment type
When both the '-i' and '-m' arguments are specified on the command
line, use the "raid10" segment type.  This way, the native RAID10
personality is used through dm-raid rather than layering a mirror
on striped LVs.  If the old behavior is desired, the '--type'
argument to use would be "mirror" rather than "raid10".
2013-08-06 14:15:08 -05:00
Jonathan Brassow
7e1083c985 RAID: Make "raid1" the default mirror segment type 2013-08-06 14:13:55 -05:00
Zdenek Kabelac
22fc80982a thin: add thin_repair and thin_dump options
Add new configure lvm.conf options for binaries thin_repair
and thin_dump.

Those are part of device-mapper-persistent-data package
and will be used for recovery of thin_pool.
2013-07-31 15:30:47 +02:00
Alasdair G Kergon
d13e87b9ef cleanup: comments and a message 2013-07-24 22:10:37 +01:00
Zdenek Kabelac
3075784955 thin: add spare lvcreate support
Add --poolmetadataspare option and creates and handles
pool metadata spare lv when thin pool is created.
With default setting 'y' it tries to ensure, spare has
at least the size of created LV.
2013-07-18 18:22:44 +02:00
Peter Rajnoha
b73fbf8f4a config_settings: add comment about possible types 2013-07-16 15:18:40 +02:00
Peter Rajnoha
a9dbe2c4fa fix: apply profile and then get the config path + auto_set_activation not profilable
cfg_def_get_path uses a global static var to store the result (for efficiency).
So we need to apply the profile first and then get the path for the config item
when calling find_config_tree_* fns.

Also activation/auto_set_activation is not profilable (at least not now,
maybe later if we need that).
2013-07-15 13:33:14 +02:00
Peter Rajnoha
8d1e511363 conf: add activation/auto_set_activation_skip
The activation/auto_set_activation_skip enables/disables automatic
adding of the ACTIVATION_SKIP LV flag. By default thin snapshots
are flagged to be skipped during activation.

And by default, the auto_set_activation_skip is enabled.
2013-07-12 20:54:17 +02:00
Peter Rajnoha
618d13705c config: fix unterminated strings in arrays
...when creating config trees while calling config_def_create_tree fn
that constructs a tree out of config_settings.h definition
(CFG_DEF_TREE_NEW/MISSING/DEFAULT/PROFILABLE).
2013-07-09 13:49:58 +02:00
Peter Rajnoha
934616d4c7 dumpconfig: honour --atversion with --type profilable
raw/~ $ lvm dumpconfig --type profilable
  allocation {
	  thin_pool_zero=1
	  thin_pool_discards="passdown"
	  thin_pool_chunk_size=64
  }
  activation {
	  thin_pool_autoextend_threshold=100
	  thin_pool_autoextend_percent=20
  }

  raw/~ $ lvm dumpconfig --type profilable --atversion 2.2.90
  activation {
	  thin_pool_autoextend_threshold=100
	  thin_pool_autoextend_percent=20
  }
2013-07-09 10:23:46 +02:00
Peter Rajnoha
953a438e93 dumpconfig: add --type profilable
The --type profilable shows all config settings that
are customizable by profiles:

  raw/~ $ lvm dumpconfig --type profilable
  allocation {
	  thin_pool_zero=1
	  thin_pool_discards="passdown"
	  thin_pool_chunk_size=64
  }
  activation {
	  thin_pool_autoextend_threshold=100
	  thin_pool_autoextend_percent=20
  }
2013-07-09 10:00:47 +02:00
Peter Rajnoha
f1e2890012 config: make it possible to do a raw config tree merge
Till now, we needed the config tree merge only for merging
tag configs with lvm.conf. However, this type of merging
did a few extra exceptions:

  - leaving out the tags section
  - merging values in activation/volume_list
  - merging values in devices/filter
  - merging values in devices/types

Any other config values were replaced by new values.
However, we'd like to do a 'raw merge' as well, simply
bypassing the exceptions listed above. This will help
us to create a single tree representing the cascaded
configs like CONFIG_STRING -> CONFIG_PROFILE -> ...

The reason for this patch is that when trees are cascaded,
the first value found while traversing the cascade is used,
not making any exceptions like we do for tag configs.
2013-07-08 15:59:25 +02:00
Peter Rajnoha
661406a417 config: fix use of last config check status if creating CFG_DEF_TREE_MISSING tree
When CFG_DEF_TREE_MISSING is created, it needs to know the status
of the check done on the tree used (the CFG_USED flag).

This bug was introduced with f1c292cc38
"make it possible to run several instances of configuration check at
once". This patch separated the CFG_USED and CFG_VALID flags in
a separate 'status' field in struct cft_check_handle.

However, when creating some trees, like CFG_DEF_TREE_MISSING,
we need this status to do a comparison with full config definition
to determine which items are missing and for which default values
were used. Otherwise, all items would be considered missing.

So, pass this status in a new field called 'check_status' in
struct config_def_tree_spec that defines how the (dumpconfig) tree
should be constructed (and this struct is passed to
config_def_create_tree fn then).
2013-07-08 15:57:49 +02:00
Peter Rajnoha
1cb66fe16e config: fix timestamp comparison during merge 2013-07-04 14:51:37 +02:00
Peter Rajnoha
9f6cfc9de4 report: add vg_profile and lv_profile to report the profile attached to VG/LV
vgs -o vg_profile ...
lvs -o lv_profile ...
2013-07-02 15:22:12 +02:00
Peter Rajnoha
9c12a23a99 profile: strictly require profiles to be valid
We must ensure that the profile is valid and contains
only items that are profilable - the check is forced
for profiles.
2013-07-02 15:22:11 +02:00
Peter Rajnoha
24a84549a8 thin: make selected thinp settings profilable
These settins are customizable by profiles:

	allocation/thin_pool_zero
	allocation/thin_pool_discards
	allocation/thin_pool_chunk_size
	activation/thin_pool_autoextend_threshold
	activation/thin_pool_autoextend_percent
2013-07-02 15:22:11 +02:00
Peter Rajnoha
e8832917f6 config: add validation for configuration profiles
Besides the classical configuration checks (type checking and
checking whether the item is recognized by lvm tools) for profiles,
do an extra check whether the configuration setting is customizable
by a profile at all. Give a warning message if not.
2013-07-02 15:22:11 +02:00
Peter Rajnoha
f1c292cc38 config: make it possible to run several instances of configuration check at once
Before, the status of the configuration check (config_def_check fn call)
was saved directly in global configuration definitinion array (as part
of the cfg_def_item_t/flags)

This patch introduces the "struct cft_check_handle" that defines
configuration check parameters as well as separate place to store
the status (status here means CFG_USED and CFG_VALID flags, formerly
saved in cfg_def_item_t/flags). This struct can hold config check
parameters as well as the status for each config tree separately,
thus making it possible to run several instances of config_def_check
without interference.
2013-07-02 15:22:11 +02:00
Peter Rajnoha
d6a91da4be config: add profile arg to find_config_tree_bool 2013-07-02 15:19:09 +02:00
Peter Rajnoha
aeffa4cb5b config: add profile arg to find_config_tree_float 2013-07-02 15:19:09 +02:00
Peter Rajnoha
dbea545ffb config: add profile arg to find_config_tree_int64 2013-07-02 15:19:09 +02:00
Peter Rajnoha
50bf2c0db1 config: add profile arg to find_config_tree_int 2013-07-02 15:19:09 +02:00
Peter Rajnoha
8ac4fcf8ff config: add profile arg to find_config_tree_str_allow_empty 2013-07-02 15:19:09 +02:00
Peter Rajnoha
06dd66af54 config: add profile arg to find_config_tree_str 2013-07-02 15:19:09 +02:00
Peter Rajnoha
eeb7b0f7fa config: add profile arg to find_config_tree_node 2013-07-02 15:19:09 +02:00
Peter Rajnoha
bfde83eb34 config: add override_config_tree_from_profile fn to add the profile to a configuration cascade
When placing the profile in a configuration cascade, this sequence is
used exactly:

CONFIG_STRING -> CONFIG_PROFILE -> CONFIG_FILE/MERGED_FILES

So if the profile is used, it overloads the lvm.conf (and any
existing tag configs). However, if "--config" is used to define
a custom configuration on command line, this overloads even the
profile config!
2013-07-02 15:19:08 +02:00
Peter Rajnoha
ebc236d085 config: add support for loading profiles
This patch adds --profile arg to lvm cmds and adds config/profile_dir
configuration setting to select the directory where profiles are stored
By default it's /etc/lvm/profile.

The profiles are added by using new "add_profile" fn and then loaded
using the "load_profile" fn. All profiles are stored in a cmd context
within the new "struct profile_params":

struct profile_params {
  const char *dir;
  struct profile *global_profile;
  struct dm_list profiles_to_load;
  struct dm_list profiles;
};

...where "dir" is the directory with profiles, "global_profile" is
the profile that is set globally via the --profile arg (IOW, not
set per VG/LV basis based on metadata record) and the "profiles"
is the list with loaded profiles.
2013-07-02 15:19:08 +02:00
Peter Rajnoha
da3ea66a96 config: add config_source_t type to identify configuration source
A helper type that helps with identification of the configuration source
which makes handling the configuration cascade a bit easier, mainly
removing and adding configuration trees to cascade dynamically.

Currently, the possible types are:

  CONFIG_UNDEFINED - configuration is not defined yet (not initialized)
  CONFIG_FILE - one file configuration
  CONFIG_MERGED_FILES - configuration that is a result of merging more files into one
  CONFIG_STRING - configuration string typed on cmd line directly
  CONFIG_PROFILE - profile configuration (the new type of configuration, patches will follow...)

Also, generalize existing "remove_overridden_config_tree" to work with
configuration type identification in a cascade. Before, it was just
the CONFIG_STRING we used. Now, we need some more to add in a
cascade (like the CONFIG_PROFILE). So, we have:

  struct dm_config_tree *remove_config_tree_by_source(struct cmd_context *cmd, config_source_t source);
  config_source_t config_get_source_type(struct dm_config_tree *cft);

... for removing the tree by its source type from the cascade and
simply getting the source type.
2013-07-02 15:19:08 +02:00
Peter Rajnoha
2d784753c0 refactor: factor out common part of the code to open and read config files 2013-07-02 15:19:08 +02:00
Zdenek Kabelac
5e19410d93 cleanup: use simplier structure initialization
And also use static struct for _out_spec.
2013-06-16 00:07:33 +02:00
Zdenek Kabelac
a040dd8cfb config: skip close of stdin
If the used stream has been stdout, do not close it.
2013-06-16 00:07:32 +02:00
Zdenek Kabelac
f319a61e9c config: fix bitfield for section
Since the cfg_def_type_t is used as bitfield in some tests,
use bitshifting to create valid bit sequence.

(in release fix)
2013-05-03 15:42:44 +02:00
Zdenek Kabelac
1445d40794 cleanup: type
oops another type inserted.
2013-04-23 12:35:10 +02:00
Zdenek Kabelac
45eeb70b02 config: merge timestamps
Merging multiple config files together needs to know newest (highest)
timestamp of merged files. Persistent cache file is being used
only in case, the config file is older then .cache file.
2013-04-23 12:31:16 +02:00
Zdenek Kabelac
2925023411 cleanup: improve messages
Add verbose message when we will not obtain devices from udev
(i.e. testing is using different udev dir, and the log was
giving misleading info about using udev)

Add proper error message if zalloc from pull would have failed.

Fix typo obolete -> obsolete
2013-04-23 12:27:33 +02:00
Zdenek Kabelac
35ab841ecf cleanup: drop zeroing of zallocated data 2013-04-23 12:27:33 +02:00
Zdenek Kabelac
4eb8db26ac cleanup: convert set of strncat into a single printf
Use single snprintf for better control of used bufsize,
and also increased buffer to 128char to ensure all
chars will always fit.
2013-04-21 23:14:05 +02:00
Zdenek Kabelac
f787b575b5 lvmetad: fix error paths
Also add missing goto out on error.
Error path missed return NULL leading to double free of enc_value.
2013-04-21 23:04:53 +02:00