e801cc607b
[block_manager] Hard code block size to 4k.
...
We're never going to use anything other than 4k, and by hard coding it we
avoid making block_manager a template.
2020-04-30 14:30:01 +01:00
3e5de399a7
[functional tests] Remove dependency on the ThunderChez library.
...
I've just moved the relevant code into the functional-tests dir.
2020-04-30 12:07:42 +01:00
ad79b627a4
Merge branch 'master' of github.com:jthornber/thin-provisioning-tools
2020-04-08 12:30:26 +01:00
0098803345
[thinp_*] Make mtree_ref_counter derive from ref_counter.
...
Slight performance hit, but I'd rather everything was derived
from ref_counter.
Also implement it! Not sure why this wasn't done before.
2020-04-08 12:24:37 +01:00
ed22bfcdda
[space-map-disk] Add virtual destructor to index_store.
...
Bug. Cleaning up index stores would not be correct.
2020-04-08 12:23:28 +01:00
1481ed0f8a
[ref_counter] Make ref_counter an abstract base class.
...
Forces people to override correctly.
Also make no_op_ref_counter final.
2020-04-08 12:22:28 +01:00
501299a4ee
[cache_*] array ref_counter wasn't overloaded correctly.
...
Major bug, explains some space map issues I've seen when
restoring cache metadata.
2020-04-08 12:19:55 +01:00
8db3800e2c
[base/application] One of the run methods wasn't overriding
2020-04-08 12:19:06 +01:00
0ccee1759e
[various] Squash a couple of clang warning wrt friend declarations.
2020-04-08 12:18:24 +01:00
5e347dddbf
[various] remove dead code.
...
Squashes warnings from clang++
2020-04-08 12:12:59 +01:00
ed44a1128f
Merge pull request #134 from Arusekk/issue-12
...
Fix compilation with uClibc
2020-03-25 15:00:28 +00:00
3ed3925370
Fix compilation with uClibc
...
Closes #12
2020-03-25 15:58:02 +01:00
47da81b6b7
Merge pull request #133 from mingnus/master
...
Improve performance, and remove unused code
2020-03-02 11:22:49 +00:00
7a1c6dc4bf
[space-map-disk] improve performance of finding a free bitmap entry
...
* Simply test the raw bitmap entries against zero, to avoid
time-consumed reference-count value extraction.
* Improve the way of iterating entries inside a bitmap:
Extract 64-bit of bitmap entries at once, and use bitwise shift
to iterate through the entries.
2020-02-29 17:39:17 +08:00
cb0a77e2ae
[block-manager] remove unused copy-assignment operator
...
block_cache::block is non-copyable, and so are the containing structures.
2020-02-29 00:29:01 +08:00
70f11e0866
Merge pull request #132 from mingnus/master
...
Fix potential file descriptor leak and compile flags
2020-02-25 16:08:36 +00:00
57cae67ff1
[unit-tests] fix googletest compile flags
...
googletest now requires c++11 compiler
2020-02-23 15:41:20 +08:00
955e11bc28
[block-cache] fix potential file descriptor leak
...
Encapsulate file descriptor into an object, to ensure that an fd will be
closed properly while exception raised, e.g., the block_cache throws
exception during the block_manager's construction.
2020-02-23 15:41:16 +08:00
c85ea5ef76
[thin_check] factor out metadata_checker
2020-02-17 11:42:24 +00:00
d6a8c03aa2
[btree_damage_visitor] factor out non-template functions
2020-02-17 11:42:24 +00:00
e6a4ba53f0
[build] add options for gprof
2020-02-17 11:42:24 +00:00
5f2c3bed69
Add some ftests, and fixup whitespace from Nikhil's work
2019-10-28 11:52:21 +00:00
cb055c90e5
Man page changes for corrupt superblock repair
2019-10-25 15:57:51 +05:30
91b4714e38
Corrupt superblock repair
2019-10-25 15:02:56 +05:30
0fc7529c01
[cache_repair, thin_repair] fix bug introduced in recent patch
...
I hadn't realised that check_file_exists() also checked that it was
a regular file, which we don't want for the couple of uses I recently
added.
This patch adds an optional arg must_be_regular_file, and defaults
it to true, preserving the original behaviour. The recent additions
have this set to false.
2019-10-14 09:21:38 +01:00
1dd7b454bb
[thin_repair, cache_repair] Check input file exists earlier
...
and zero superblock if we fail part way through a repair.
bz1499781
2019-10-08 14:34:24 +01:00
a3b8e9db78
[thin_trim] remove all mention of --pool-inactive from man page.
...
This option is obsolete now that we open O_EXCL. If used a warning message
will be output.
2019-10-07 13:32:52 +01:00
6332962ee8
[ft-lib/bcache] rename raise() -> raise_()
...
Name clash with signal.h on Debian and Gentoo.
2019-06-05 15:02:05 +01:00
5e5409f48b
[space-map/noop] Be explicit about returning an optional
...
Needed for older version of g++
2019-06-05 11:07:28 +01:00
f5acd0426d
bump version
2019-06-04 10:39:59 +01:00
7be9be1494
[space-maps/noop] Add the noops space map
...
Used when we know we're npt modifying the metadata.
2019-06-04 10:38:58 +01:00
c19a05f28f
bump version
2019-06-03 14:11:37 +01:00
733c7df798
[thin_repair/thin_dump] Fix some more spurious error messages when doing repair.
...
Repair was falling back to non-repair behaviour if it thought the roots
were ok. Now if --repair is specified the same dumping code is always
executed.
2019-06-03 14:03:24 +01:00
e45a211902
[thin/metadata_dumper] Change a couple of confusing class names.
...
metadata_emitter is actually a visitor that passes on it's data
to an encapsulated emitter object.
metadata_emitter -> metadata_emit_visitor
metadata_tree_emitter -> metadata_tree_emit_visitor
2019-06-03 11:07:32 +01:00
a13178ffd1
Bump version
2019-05-28 09:24:22 +01:00
0624ded67f
[thin_repair, thin_dump] Don't look for better roots if the ones in the sb are ok.
...
Calculating the most recent trees is inexact, so trees other than the ones
in the superblock, which presumably are the most recent, could be chosen.
2019-05-27 16:43:32 +01:00
5e6ddbf2da
[thin_dump, thin_repair] Fix bad metadata_damage exception
...
The first pass of the repair process scans all metadata working out the
largest orphan btrees. This scan doesn't use as much validation as
the btree_walk function which subsequently gets called.
This patch catches any exceptions thrown by the btree walk function
and removes that btree from consideration.
2019-05-27 16:06:59 +01:00
d299cc66d1
[functional-tests/thin_dump] check no error msg with clean metadata
2019-05-27 15:38:45 +01:00
acf2e6e625
[unit-tests] squash some warnings
...
ASSERT_THAT(), defined in google-test needs to be wrapped in braces if
it's the only statement in an if block (dangling else).
2019-05-27 09:19:14 +01:00
65264911c1
[thin_repair, thin_dump] --repair falg wasn't being passed through properly.
...
Recent regression.
2019-05-27 09:04:55 +01:00
9e87ad8298
bump version to 0.8.1
2019-05-03 15:19:06 +01:00
595ddfc50f
[thin_repair/thin_dump] repair flag wasn't being set internally.
...
Introduced by recent changes
2019-05-03 15:15:01 +01:00
4bca324c9f
Bump version to 0.8.0
2019-04-17 12:55:15 +01:00
9e20465fd1
[thin-repair, thin_dump] When repairing we now hunt for the best btree roots.
...
We've had a trickle of users who accidentally activate the same pool on a
VM and host at the same time. Typically the host doesn't do any IO, but
the kernel will still rewrite the superblock on shutdown. This leaves
the superblock pointing to very out of date btree roots and so we get
massive metadata loss.
This patch changes thin_repair, and thin_dump --repair. They now hunt
for the most recent, undamaged and consistent roots of the device and
mapping trees, and use that as the starting point of the repair.
2019-04-17 12:17:13 +01:00
b027a1039f
[thin-provisioning] Add a new constructor for the metadata class
...
This one takes a superblock. Used when repairing.
2019-04-17 12:15:46 +01:00
b777a77023
[thin-provisioning] pass the tm by reference to the metadata object
2019-04-04 15:06:41 +01:00
aace49cdd6
[thin_repair] work in progress
2019-03-29 12:56:31 +00:00
8abac422b2
[space-map-disk] fix the search range ( #117 )
2019-01-03 10:46:35 +00:00
95ae4fef6a
[functional tests] Use dd to create zeroed files
...
Previously I was using fallocate, which has issues on some filesystems.
2018-12-12 13:14:09 +00:00
e03b878df3
[functional tests] Move test output from /tmp to .
...
tmpfs has trouble with O_DIRECT.
2018-12-12 13:13:11 +00:00
Joe Thornber
Arusekk
Ming-Hung Tsai
Nikhil Kshirsagar