Peter Rajnoha [Fri, 17 Jun 2016 09:03:14 +0000 (11:03 +0200)]
args: add priorities for grouping args
Groupable args (the ones marked with ARG_GROUPABLE flag) start a new
group of args if:
- this is the first time we hit such a groupable arg,
- or if non-countable arg is repeated.
However, there may be cases where we want to give priorities when
forming groups and hence force new group creation if we hit an arg
with higher grouping priority.
For example, let's assume (for now) hypothetical sequence of args used:
So here "-o" started a new group but since "--configreport" has higher
priority than "-o", it starts fresh new group now and hence the rest of
the command line's args are grouped by --configreport now.
Peter Rajnoha [Tue, 3 May 2016 09:18:16 +0000 (11:18 +0200)]
commands: report: add lvm fullreport command
lvm fullreport executes 5 subreports (vg, pv, lv, pvseg, seg) per each VG
(and so taking one VG lock each time) within one command which makes it
easier to produce full report about LVM entities.
Since all 5 subreports for a VG are done under a VG lock, the output is
more consistent mainly in cases where LVM entities may be changed in
parallel.
Peter Rajnoha [Mon, 2 May 2016 12:37:16 +0000 (14:37 +0200)]
report: add current report to existing group
Add any report (pvs/vgs/lvs) currently processed to current report group
which is part of processing handle and which already contains log
report. This way both log report and pvs/vgs/lvs report will be
reported as a whole within a group, thus having same output format as
selected by --reportformat option.
Peter Rajnoha [Tue, 31 May 2016 10:24:05 +0000 (12:24 +0200)]
toollib: add 'parent' field to processing_handle; init report format only if there's no parent
If there's parent processing handle, we don't need to create completely
new report group and status report - we'll just reuse the one already
initialized for the parent.
Currently, the situation where this matter is when doing internal report
to do the selection for processing commands where we have parent processing
handle for the command itself and processing handle for the selection
part (that is selection for non-reporting tools).
Peter Rajnoha [Mon, 23 May 2016 08:16:29 +0000 (10:16 +0200)]
toollib: add report_group and log_rh to processing_handle and initialize cmd processing log report
Wire up report group creation with log report in struct
processing_handle and call report_format_init during processing handle
initialization (init_processing_handle fn) and destroy it while
destroing processing handle (destroy_processing_handle fn).
This way, all the LVM command processing using processing handle
has access to log report via which the current command log
can be reported as items are processed.
Peter Rajnoha [Wed, 25 May 2016 11:41:42 +0000 (13:41 +0200)]
report: separate common report and per-report args
Separating common report and per-report arguments prepares the code for
handling several reports per one command (for example, the command log
report and LVM command report itself).
Each report can have sort keys, options (fields), list of fields to
compact and selection criteria set individually. Hooks for setting these
per report within one command will be a part of subsequent patches, this
patch only separates new struct single_report_args out of existing
struct report_args.
Peter Rajnoha [Tue, 24 May 2016 10:05:42 +0000 (12:05 +0200)]
conf: add report/output_format config setting
New report/output_format configuration sets the output format used
for all LVM commands globally. Currently, there are 2 formats
recognized:
- basic (the classical basic output with columns and rows, used by default)
- json (output is in json format)
Peter Rajnoha [Mon, 2 May 2016 12:22:02 +0000 (14:22 +0200)]
toollib: add report_format_init fn to create report group and to create/add log report handle
Add new --reportformat option and new report_format_init function that
checks this option and creates new report group accordingly, also
preparing log report handle and adding it to the report group just
created.
Peter Rajnoha [Tue, 10 May 2016 13:15:48 +0000 (15:15 +0200)]
report: add CMDLOG report type
This is a preparation for new CMDLOG report type which is going to be
used for reporting LVM command log.
The new report type introduces several new fields (log_seq_num, log_type,
log_context, log_object_type, log_object_group, log_object_id, object_name,
log_message, log_errno, log_ret_code) as well as new configuration settings
to set this report type (report/command_log_sort and report/command_log_cols
lvm.conf settings).
This patch also introduces internal report_cmdlog helper function
which is a wrapper over dm_report_object to report command log via
CMDLOG report type and which is going to be used throughout the code
to report the log items.
Peter Rajnoha [Mon, 2 May 2016 12:21:40 +0000 (14:21 +0200)]
libdm: report: implement DM_REPORT_GROUP_JSON for JSON report output
This patch introduces DM_REPORT_GROUP_JSON report group type. When using
this group type and when pushing a report to such a group, these flags
are automatically unset:
The whole group is encapsulated in { } for the outermost JSON object
and then each report is reported on output as array of objects where
each object is the row from report:
Peter Rajnoha [Mon, 23 May 2016 08:47:03 +0000 (10:47 +0200)]
libdm: report: implement DM_REPORT_GROUP_BASIC for extended report output
This patch introduces DM_REPORT_GROUP_BASIC report group type. This
type has exactly the classical output format as we know from before
introduction of report groups. However, in addition to that, it allows
to put several reports into a group - this is the very basic grouping
scheme that doesn't change the output format itself:
Report: report1_name
Header1 Header2 ...
value value ...
value value ...
... ... ...
Report: report2_name
Header1 Header2 ...
value value ...
value value ...
... ... ...
Peter Rajnoha [Thu, 16 Jun 2016 09:23:56 +0000 (11:23 +0200)]
libdm: report: implement DM_REPORT_GROUP_SINGLE for a group containing single report only
There's no change in output for this report group type - with this type,
we only make sure there's always only one report in a group at a time,
not more.
Peter Rajnoha [Mon, 2 May 2016 12:21:05 +0000 (14:21 +0200)]
libdm: report: introduce dm_report_group
This patch introduces DM report group (represented by dm_report_group
structure) that is used to group several reports to make a whole. As a
whole, all the reports in the group follow the same settings and/or
formatting used on output and it controls that the output is properly
ordered (e.g. the output from different reports is not interleaved
which would break readability and/or syntax of target output format
used for the whole group).
To support this feature, there are 4 new functions:
- dm_report_group_create
- dm_report_group_push
- dm_report_group_pop
- dm_report_group_destroy
From the naming used (dm_report_group_push/pop), it's clear the reports
are pushed onto a stack. The rule then is that only the report on top
of the stack can be reported (that means calling dm_report_output).
This way we make sure that the output is not interleaved and provides
determinism and control over the output.
Different formats may allow or disallow some of the existing report
flags controlling output itself (DM_REPORT_OUTPUT_*) to be set or not so
once the report is pushed to a group, the grouping code makes sure that
all the reports have compatible flags set and then these flags are
restored once each report is popped from the report group stack.
We also allow to push/pop non-report item in which case such an item
creates a structure (e.g. to put several reports together with any
opening and/or closing lines needed on output which pose as extra
formatting structure besides formatting the reports).
The dm_report_group_push function accepts an argument to pass any
format-specific data needed (e.g. handle, name, structures passed
along while working with reports...).
We can call dm_report_output directly anytime we need (with the only
restriction that we can call dm_report_output only for the report that
is currently on top of the group's stack). Or we don't need to call
dm_report_output explicitly in which case all the reports in a stack are
reported on output automatically once we call dm_report_group_destroy.
David Teigland [Fri, 17 Jun 2016 16:45:08 +0000 (11:45 -0500)]
lvmlockd: fix dropping PVs in rescanning VG
This fixes a problem in commit ae0a8740c. The problem
in that commit was that all existing PVs are initially
dropped from lvmetad. This works if the VG is updated
at the end, which replaces the dropped PVs, but if the
rescan finds that the VG seqno is unchanged, it leaves
the cached VG in place. So, we should only drop the
existing PVs in lvmetad when the VG is going to be updated.
David Teigland [Thu, 16 Jun 2016 17:04:05 +0000 (12:04 -0500)]
pvscan: don't activate LVs when use_lvmetad=0
commit 15da467b was meant to address the case where
use_lvmetad=1 in lvm.conf, and lvmetad is not available,
in which case, pvscan --cache -aay should activate LVs.
But the commit unintentionally also changed the case
where use_lvmetad=0 in lvm.conf, in which case
pvscan --cache -aay should not activate LVs, so fix
that here.
David Teigland [Wed, 15 Jun 2016 19:19:18 +0000 (14:19 -0500)]
pvscan: do activation when lvmetad is not running
When pvscan --cache -aay fails to connect to lvmetad it will
simply exit and do nothing. Change this so that it will
skip the lvmetad cache step and do the activation step from
disk.
Tony Asleson [Fri, 10 Jun 2016 17:11:17 +0000 (12:11 -0500)]
lvmdbusd: Create correct LV object type
We were initially looking to see if an LV was hidden and if it was we were
creating an instance of a LvCommon object to represent it. Thus if we
had a hidden cache pool for example we were missing the methods and
properties for the cache pool. However, when we create the object path,
any hidden LVs, regardless of type/functionality will be placed in the
hidden path.
Tony Asleson [Fri, 10 Jun 2016 17:05:52 +0000 (12:05 -0500)]
lvmdbusd: Add method get_object_path_by_lvm_id
The object manager method get_object_by_lvm_id was used in many cases for
the sole reason of getting the object path for the object. Instead of
retrieving the object and then calling 'dbus_object_path' on the object, we
are adding a method which returns the object path.
Tony Asleson [Fri, 10 Jun 2016 16:58:55 +0000 (11:58 -0500)]
lvmdbusd: Bug fix for missing LV properties
When we are processing the LVs we need to build up dbus objects from least
dependent to most dependent, so that we have information available when
constructing.
Zdenek Kabelac [Wed, 8 Jun 2016 18:52:14 +0000 (20:52 +0200)]
display: yes no prompt improvement
Original code missed to catch all apperances of SIGINT.
Also enhance logging when running in shell without tty.
Accept this regex as valid input:
'^[ ^t]*([Yy]([Ee]([Ss]|)|)|[Nn]([Oo]|))[ ^t]*$'
David Teigland [Mon, 6 Jun 2016 20:20:55 +0000 (15:20 -0500)]
lvmcache: fix duplicate handling with multiple scans
Some commands scan labels to populate lvmcache multiple
times, i.e. lvmcache_init, scan labels to fill lvmcache,
lvmcache_destroy, then later repeat
Each time labels are scanned, duplicates are detected,
and preferred devices are chosen. Each time this is done
within a single command, we want to choose the same
preferred devices. So, check for existing preferences
when choosing preferred devices.
This also fixes a problem with the list of unused duplicate
devs when run in an lvm shell. The devs had been allocated
from cmd memory, resulting in invalid list entries between
commands.
David Teigland [Mon, 6 Jun 2016 19:04:17 +0000 (14:04 -0500)]
lvmcache: add optional dev arg to lvmcache_info_from_pvid
A number of places are working on a specific dev when they
call lvmcache_info_from_pvid() to look up an info struct
based on a pvid. In those cases, pass the dev being used
to lvmcache_info_from_pvid(). When a dev is specified,
lvmcache_info_from_pvid() will verify that the cached
info it's using matches the dev being processed before
returning the info. Calling code will not mistakenly
get info for the wrong dev when duplicate devs exist.
This confusion was happening when scanning labels when
duplicate devs existed. label_read for the first dev
would add an info struct to lvmcache for that dev/pvid.
label_read for the second dev would see the pvid in
lvmcache from first dev, and mistakenly conclude that
the label_read from the second dev can be skipped
because it's already been done. By verifying that the
dev for the cached pvid matches the dev being read,
this mismatch is avoided and the label is actually read
from the second duplicate.
David Teigland [Fri, 20 May 2016 18:32:14 +0000 (13:32 -0500)]
lvmetad: handle update failures
If a command gets stuck during an lvmetad update, lvmetad
will cancel that update after the timeout. The next command
to check the lvmetad will see that lvmetad needs to be
populated because lvmetad will return token of "none" after
a timed out update (same as when lvmetad is not populated
at all after starting.)
If a command gets an error during an lvmetad update, it
will now just quit and leave its updating token in place.
That update will be cancelled after the timeout.
Peter Rajnoha [Mon, 6 Jun 2016 12:57:41 +0000 (14:57 +0200)]
blkdeactivate: fix regression in blkdeactivate causing dm and md devices to be skipped.
Commit #5b3a4a9 caused the "name" variable to be cleared if
declaration and assignment is on two lines so put it back
so it's on one line for it to work again.
David Teigland [Fri, 3 Jun 2016 14:56:48 +0000 (09:56 -0500)]
pvmove: disallow tag args
pvmove began processing tags unintentionally from commit, 6d7dc87cb pvmove: use toollib
pvmove works on a single PV, but tags can match multiple PVs.
If we allowed tags, but processed only the first matching PV,
then the resulting PV would be unpredictable.
Also, the current processing code does not allow us to simply
report an error and do nothing if more than one PV matches the tag,
because the command starts processing PVs as they are found,
so it's too late to do nothing if a second PV matches.
Peter Rajnoha [Thu, 2 Jun 2016 11:41:55 +0000 (13:41 +0200)]
lvmconfig: fix lvmconfig --type diff to display complete diff if config cascade used
If configuration consists of several sources in config cascade
("config cascade" defined in man lvmconfig(8)), lvmconfig displayed
only difference from defaults of the topmost config in the cascade.
Fix lvmconfig to display complete difference, considering all
the configuration in the cascade.
For example, before this patch:
(use_lvmetad=0 set in lvm.conf which differs from defaults)
$ lvmconfig --type diff
global {
use_lvmetad=0
}
(compact_output=1 set on cmd line)
$ lvmconfig --type diff --config report/compact_output=1
report {
compact_output=1
}
(headings=0 set in profile)
$ lvmconfig --type diff --commandprofile test
report {
headings=0
}
(difference in topmost configuration source is displayed)
$ lvmconfig --type diff --commandprofile test --config report/compact_output=1
report {
compact_output=1
}
With this patch applied (the config cascade is merged before looking for
difference from defaults in configuration):
Zdenek Kabelac [Wed, 1 Jun 2016 14:39:47 +0000 (16:39 +0200)]
devices: handle partscan loop devices
Treat loop device created with 'losetup -P' as regular
partitioned device - so if it has partition table,
prevent its usage in commands like 'pvcreate'.
Before 'pvcreate /dev/loop0' could have erased and formated as PV,
after this patch, device is filtered out and cannot be used.
Peter Rajnoha [Tue, 31 May 2016 08:06:19 +0000 (10:06 +0200)]
coverity: fix warnings about missing return value check for sscanf
All the variables for sscanf in lvmlockctl.c and lvmlockd-sanlock.c are
zeroed before sscanf call so the failure is detected by seeing the zero
value instead of proper one in subsequent code - so use (void) for
sscanf calls to ignore return value here.
Peter Rajnoha [Mon, 30 May 2016 14:37:05 +0000 (16:37 +0200)]
report: fix lvm devtypes internal error if -S is used with field name from pvs/vgs/lvs
Before this fix, when reporting 'lvm devtypes', the report was
initialized with incorrect reserved values - the ones used for
pvs/vgs/lvs report were used instead of NULL value (because devtypes
doesn't have any reserved values).
For example, trying to (incorrectly) use lv_name for the -S|--select
with lvm devtypes which doesn't have this field at all:
Before this patch (internal error issued):
$ lvm devtypes -S 'lv_name=lvol0'
Internal error: _check_reserved_values_supported: field-specific reserved value of type 0x0 for field not supported
Internal error: dm_report_init_with_selection: trying to register unsupported reserved value type, skipping report selection
DevType MaxParts Description
aoe 16 ATA over Ethernet
ataraid 16 ATA Raid
bcache 1 bcache block device cache
...
With this patch applied (correct error displayed about
unrecognized selection field):
$ lvm devtypes -S 'lv_name=lvol0'
Device Types Fields
-------------------
devtype_name - Name of Device Type exactly as it appears in /proc/devices. [string]
devtype_max_partitions - Maximum number of partitions. (How many device minor numbers get reserved for each device.) [number]
devtype_description - Description of Device Type. [string]
Special Fields
--------------
selected - Set if item passes selection criteria. [number]
help - Show help. [unselectable number]
? - Show help. [unselectable number]
Unrecognised selection field: lv_name
Selection syntax error at 'lv_name=lvol0'.
Use 'help' for selection to get more help.
Zdenek Kabelac [Wed, 25 May 2016 14:26:10 +0000 (16:26 +0200)]
report: convert more options to use single status
Convert fields into using a single status ioctl call per LV.
This is a bit tricky since when there are more complicated
stacks, at this moment its undefined which values should be shown.
It's clear we need to cache more then single ioctl per LV,
but also we need to define more explicitely relation between
reported values for snapshots.
This patch is not a final state, rather a transitional step.
It should not be giving more 'worst' values then previous
many-ioctl-calls-per-lv solution.
Add function to obtain percentage value for cache lv_seg_status.
This API is rather evolving 'middle' step as the ultimate goal
is segment API fuctionality.
But first we need to be clear at reporting level which values
are needed to be reported for which LVs and segments.
Zdenek Kabelac [Wed, 25 May 2016 14:14:46 +0000 (16:14 +0200)]
lvstatus: enhance seg_status to handle snapshot
Add more code to properly store status for snapshot segment
maintaining lvm2 fiction of COW and snapshot internal volumes.
The key issue here is however not though-through reporting
logic - as there is no single answer for whole line state.
It not counting with layer and we may need few more ioctl to
cover all reporting needs depending upon what is actually
needed.
In reality we need to 'cache' more ioctl status queries for
individual LVs and their segments (so they checked at most once).
The other 'hard' topic for conversion is mirror segment handling.
Also we definitelly need to relocate some logic into segment's methods,
yet it might be complex as we have not clear border between targets.
TODO: define more clearly how are reporting fields defined in case
we 'stack' volumes like - cache of stacked thin LV snapshot origin.
Zdenek Kabelac [Fri, 27 May 2016 11:49:30 +0000 (13:49 +0200)]
snapshot: check merging_cow is cow
Check first the LV is cow before even checking it's a merging COW.
Note: previosly merging_cow was also merging origin, so without
this explicit check it used to return '1' also when passed
LV has been merging origin.
Zdenek Kabelac [Fri, 27 May 2016 11:48:30 +0000 (13:48 +0200)]
report: fix report copy_percent value
When mirror/raid called copy_percent function to return,
when 100% was supposed to be returned, wrong float 100.0 value
could have been reported back instead of dm_percent_t DM_PERCENT_100.
There is broken API somewhere, since the function here rely on
actively being modifid VG content even when doing 'lvs' operation.
(extents_copies)
This was a deliberate omission because logging token-by-token metadata
parsing greatly increases the amount of logging for hardly any benefit.
In general, only LVM config file settings need to be logged, and in
places where it's considered important to log particular elements of
metadata that should be done using specific log_* lines.
David Teigland [Mon, 23 May 2016 18:42:17 +0000 (13:42 -0500)]
lvconvert: use process_each_lv
In the same way that process_each_vg() can be passed
a single VG name to process, also allow process_each_lv()
to be passed a single VG name and LV name to process.