Commit graph

1039 commits

Author SHA1 Message Date
47dcfa20f3 [#895] test: Use t.Cleanup only for external resources
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2024-01-11 12:32:09 +00:00
f1b2b8bffa [#895] test: Fix NewLogger arguments list
`debug` is always true.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2024-01-11 12:32:09 +00:00
4b8b4da681 [#864] engine: Drop container count metric if container removed
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2024-01-10 10:45:32 +03:00
d75e7e9a21 [#864] engine: Drop container size metric if container deleted
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2024-01-10 10:44:54 +03:00
dfd62ca6b1 [#864] metabase: Refactor delete/inhume
Available -> Logic, Raw -> Phy for delete/inhume results.
Use single counter instead of vectors.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2024-01-09 09:59:42 +03:00
225fe2d4d5 [#894] blobovniczatree: Speedup rebuild test
Down from 3s to 300ms.

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-12-29 16:28:54 +00:00
581887148a [#569] cli: Add control shards writecache seal command
It does the same as `control shards flush-writecache --seal`, but
has better name.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-29 16:05:37 +03:00
7a9db5bcdd [#569] writecache: Do not wait modeMtx if mode changes
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-29 16:05:37 +03:00
32c282ca10 [#569] writecache: Refactor flush
Make single RUnlock call instead of two.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-29 16:05:37 +03:00
0cb0fc1735 [#569] writecache: Allow to seal writecache after flush
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-29 16:05:37 +03:00
8180a0664f [#887] node: Drop badger writecache implementation
Badger implementation isn't tested and works not well,
but requires human resources to maintain.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-22 13:00:54 +03:00
d9cbb16bd3 [#866] Use TTL for blobovnicza tree cache
Signed-off-by: Anton Nikiforov <an.nikiforov@yadro.com>
2023-12-19 16:36:28 +00:00
7eb46404a1 [#863] blobovnicza: Fix counters
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-13 13:34:29 +03:00
94ffe8bb45 [#857] golangci: Add testifylint linter
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-12 16:27:02 +03:00
3b7c0362a8 [#861] shard: Fix Delete object
It is possible that object doesn't exist in metabase.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-12 14:25:40 +03:00
681b2c5fd4 [#825] policer: Do not drop required linking objects
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-12 11:04:03 +00:00
db49ad16cc [#826] blobovniczatree: Do not create DB's on init
Blobovniczas will be created on write requests.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-07 15:37:33 +03:00
ad0697adc4 [#661] blobovnicza: Compute size with record size
To get more accurate size of blobovnicza use record
size (lenght of key + lenght of data).

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-07 15:37:33 +03:00
e54dc3dc7c [#698] blobovnicza: Store counter values
Blobovnicza initialization take a long time because of bucket
Stat() call. So now blobovnicza stores counters in META bucket.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-07 15:37:33 +03:00
5e8c08da3e [#661] blobstore: Add address to error logs
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-07 15:37:33 +03:00
8911656b1a [#661] metrcis: Add rebuild percent metric
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-07 15:37:33 +03:00
2407e5f5ff [#661] blobovniczatree: Do not sort DB's and indicies
Put stores object to next active DB, so there is no need to sort DBs.
In addition, it adds unnecessary DB openings.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-07 15:37:33 +03:00
c6a739e746 [#661] blobovniczatree: Make Rebuild concurrent for objects
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-07 15:37:33 +03:00
f1c7905263 [#661] blobovniczatree: Make Rebuild concurrent
Different DBs can be rebuild concurrently.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-07 15:37:33 +03:00
d4d905ecc6 [#661] metrics: Add blobovniczatree rebuild metrics
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-07 15:37:33 +03:00
b2769ca3de [#661] blobovniczatree: Make Rebuild failover safe
Now move info stores in blobovnicza, so in case of failover
rebuild completes previous operation first.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-07 15:37:33 +03:00
da4fee2d0b [#698] blobovniczatree: Init blobovniczas concurrently
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-07 15:37:33 +03:00
422226da18 [#661] blobovniczatree: Add Rebuild implementation
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-07 15:37:32 +03:00
a531eaf8bc [#661] blobstor: Add Rebuild implementation
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-07 15:37:32 +03:00
c1667a11d2 [#661] blobovniczatree: Allow to change depth or width
Now it is possible to change depth or with of blobovniczatree.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-07 15:37:32 +03:00
484eb59893 [#661] blobovniczatree: Use .db extension for db files
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-07 15:37:32 +03:00
44552a849b [#661] shard: Add blobstor rebuilder
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-07 15:37:32 +03:00
a478050639 [#838] metabase: Resolve funlen linter
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-06 15:44:21 +03:00
d30ab5f29e [#838] metabase: Count user objects
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-06 15:44:21 +03:00
f314da4af3 [#838] metabase: Add user object type counter
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-06 15:44:21 +03:00
29550fe600 [#838] shard: Refactor updateMetrics method
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-12-06 15:44:21 +03:00
b2c63e57ba [#651] engine/test: Speedup StorageEngine_Inhume
Signed-off-by: Alexander Chuprov <a.chuprov@yadro.com>
2023-11-30 13:19:43 +00:00
445ebcc0e7 [#651] shard/test: Speedup Shard_Delete
Signed-off-by: Alexander Chuprov <a.chuprov@yadro.com>
2023-11-30 13:19:43 +00:00
2302e5d342 [#651] shard/test: Refactor Shard_Delete
Signed-off-by: Alexander Chuprov <a.chuprov@yadro.com>
2023-11-30 13:19:43 +00:00
07390ad4e3 [#715] node: Unify config parameter names
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-11-22 17:13:50 +03:00
8d18fa159e [#667] writecache: Fix flush test
Allow to disable background flusher for testing purposes.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-11-17 17:45:43 +03:00
76ff26039c [#96] node: Drop neo-go's slices package
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-11-17 13:24:04 +03:00
47286ebf32 [#805] pilorama: Fix TreeDrop
* If treeID is empty then deleting buckets for cursor may get
  invalidated. So, buckets should be gathered before deleting.

Signed-off-by: Airat Arifullin <a.arifullin@yadro.com>
2023-11-17 10:21:35 +00:00
29fe8c41f3 [#655] storage: Drop ErrorHandler
The only one usage was for logging.
Now logging performed by storage anyway.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-11-16 17:27:38 +03:00
137e987a4e [#655] storage: Drop LazyHandler
LazyHandler is implemented and used incorrectly.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-11-16 17:27:38 +03:00
1b22801eed [#788] engine: Fix flaky tests
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-11-09 10:25:46 +03:00
70ab1ebd54 [#763] metrics: Add container_objects_total metric
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-11-08 12:30:57 +03:00
9c98fa6152 [#763] metabase: Add container objects counter
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-11-08 12:30:57 +03:00
226e84d782 [#684] node: Add skipped objects count to evacuation result
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-11-07 12:17:11 +00:00
c80b46fad3 [#754] blobstor: Estimate compressability
Now it is possible to enable compressability estimation.
If data is likely uncompressable, it should reduce CPU time and memory.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-11-01 11:24:32 +03:00
79088baa06 [#772] node: Apply gofumpt
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-10-31 17:03:03 +03:00
00aa6d9749 [#633] shard/test: Fix TestCounters()
Introduced in 362f24953a, forgotten to be changed because test
generator didn't provide payload size.

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-10-31 12:53:28 +00:00
b8f79f4227 [#633] shard/test: Fix race conditions in TestCounters()
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-10-31 12:53:28 +00:00
869518be0a [#728] writecache: Fix Badger writecache race.
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-10-30 18:36:41 +03:00
d4b6ebe7e7 [#725] writecache: Fix metric values
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-10-27 12:22:29 +03:00
559ad58ab1 [#642] writecache: Remove usage of close channel in bbolt
Signed-off-by: Anton Nikiforov <an.nikiforov@yadro.com>
2023-10-24 15:57:50 +00:00
c0b86f2d93 [#642] writecache: Remove usage of close channel in badger
Signed-off-by: Anton Nikiforov <an.nikiforov@yadro.com>
2023-10-24 15:57:50 +00:00
f2437f7ae9 [#734] shard: Fix Delete method
Due to the flushing data from the writecache to the storage
and simultaneous deletion, a partial deletion situation is possible.
So as a solution, deletion is allowed only when the object is in storage,
because object will be deleted from writecache by flush goroutine.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-10-16 17:00:18 +03:00
f26233b47a [#734] metabase: Include UpdateStorageID in metrics and traces
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-10-16 17:00:18 +03:00
7e0c5a55de [#734] writecache: Fix flush
Now UpdateStorageID doesn't return error in case of logical error.
If object is in graveyard or GC market, it is still required to
update storage ID.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-10-16 17:00:17 +03:00
d07afd803c [#726] writecache: Fix small object flush for Badger
Do not marshal object twice.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-10-06 11:32:50 +03:00
997ac7cd8d [#726] writecache: Fix small object flush for BBolt
Do not marshal object twice.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-10-06 11:32:44 +03:00
c1e4130020 [#146] node: Add trace_id to logs
Signed-off-by: Alexander Chuprov <a.chuprov@yadro.com>
2023-09-27 11:05:27 +03:00
3889e829e6 [#667] writecache: Add logs for report error func in tests
Signed-off-by: Airat Arifullin a.arifullin@yadro.com
2023-09-14 17:02:54 +00:00
c6af4a3ec8 [#679] engine: Do not increase error counter on meta mismatch
It was introduced in 69e1e6ca to help node determine faulty shards.
However, the situation is possible in a real-life scenario:
1. Object O is evacuated from shard A to B.
2. Shard A is unmounted because of lower-level errors.
3. We now have object in meta on A and in blobstor on B. Technically we
   have it in meta on shard B too, but we still got the error if B goes
   to a degraded mode.

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-09-14 10:39:18 +03:00
8a81af5a3b [#653] Add context parameter to Open functions
Signed-off-by: Ekaterina Lebedeva <ekaterina.lebedeva@yadro.com>
2023-09-07 18:03:29 +03:00
a716db99db [#668] shard/test: Do not alter rootPath option
Supposedly, this was added to allow creating 2 different shards without
subtest. Now we use t.TempDir() everywhere, so this should not be a
problem.

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-09-07 07:39:39 +00:00
36759f8434 [#668] shard/test: Properly check event processing
See https://git.frostfs.info/TrueCloudLab/frostfs-node/actions/runs/1594/jobs/2

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-09-07 07:39:39 +00:00
39879fa868 [#668] shard/test: Add dontRelease options
Most of the time we would like to close shard with minor exceptions.

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-09-07 07:39:39 +00:00
c661ba1312 [#668] shard/test: Use sane defaults in the test constructor
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-09-07 07:39:39 +00:00
268adb79cb [#668] shard/test: Simplify shard construction
newCustomShard() has many parameters but only the first is obligatory.
`enableWriteCache` is left as-is, because it directly affects the
functionality.

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-09-07 07:39:39 +00:00
429f941cda [#668] shard/test: Release shard in t.Cleanup()
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-09-07 07:39:39 +00:00
382eb8a485 [#668] shard/test: Disable GC where it is not needed
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-09-07 07:39:39 +00:00
42696016de [#668] shard: Close stopChannel in GC
It is done once, but now we could read it from multiple places.

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-09-07 07:39:39 +00:00
bdecfbc1be [#668] shard/test: Move tests to the main package
Semantic patch (also, duplicate definitions are removed):
```
@@
var e identifier
@@
-import "git.frostfs.info/TrueCloudLab/frostfs-node/pkg/local_object_storage/shard"

-shard.e
+e
```

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-09-07 07:39:39 +00:00
aa23c6a83a [#668] shard/test: Remove subtest from TestCounters
Otherwise, individual tests cannot be run.

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-09-07 07:39:39 +00:00
da8f384324 [#668] shard/test: Fix typo in existence
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-09-07 07:39:39 +00:00
7456c8556a [#536] blobovnicza: Add blobovniczatree DB cache
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-09-01 13:53:11 +03:00
c672f59ab8 [#536] blobovnicza: Drop cache
Each blobovnicza instance is opened
while is in use.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-09-01 13:51:26 +03:00
b9b86d2ec8 [#666] shard/test: Fix data race in metrics tests
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-08-31 08:39:42 +00:00
4dff9555f1 [#568] writecache: Improve flushing scheme for badger
Signed-off-by: Anton Nikiforov <an.nikiforov@yadro.com>
2023-08-30 17:22:28 +00:00
1daef2ceeb [#660] writecache: Fix remaining addr2key uses
Signed-off-by: Alejandro Lopez <a.lopez@yadro.com>
2023-08-30 17:12:33 +00:00
9072772a09 [#649] shard/test: Increase GC remover interval
This was set in #348 to speed up tests.
It seems 100ms doesn't increase overall test time,
but it reduces the amount of logs by 100x factor.

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-08-25 10:10:25 +00:00
c4db8e7690 [#637] shard/test: Fix data race
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-08-25 10:10:25 +00:00
96e690883f [#638] Unify test loggers
In some places we have debug=false, in others debug=true.
Let's be consistent.

Semantic patch:
```
@@
@@
-test.NewLogger(..., false)
+test.NewLogger(..., true)
```

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-08-23 11:21:05 +00:00
322c1dc273 [#638] Use test.NewLogger() in tests
Semantic patch (restricted to **/*_test.go):
```
@@
@@
+import "git.frostfs.info/TrueCloudLab/frostfs-node/pkg/util/logger/test"
-import "go.uber.org/zap"
-import "git.frostfs.info/TrueCloudLab/frostfs-node/pkg/util/logger"

-&logger.Logger{Logger: zap.L()}
+test.NewLogger(t, false)
```

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-08-23 11:21:05 +00:00
82cc453be9 [#xx] shard: Fix data race in metrics tests
Protect test metric store fields with a mutex. Probably, not every field
should be protected, but better safe than sorry.

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-08-23 10:26:12 +00:00
345a1a69a2 [#635] Use internal key type when deleting from badger wc
Signed-off-by: Alejandro Lopez <a.lopez@yadro.com>
2023-08-22 10:53:19 +03:00
354a92ea2c [#602] blobovnicza: Add leaf width implementation
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-08-21 10:27:32 +03:00
4d9a6c07fb [#618] core: Replace fmt.Sprintf with strconv.FormatUint
Signed-off-by: Alexander Chuprov <a.chuprov@yadro.com>
2023-08-21 07:14:50 +00:00
f2811f8585 [#602] metrics: Add blobovnicza items counter
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-08-18 13:01:27 +03:00
c4e1d8eb07 [#602] node: Fix blobovnicza typos
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-08-18 11:14:10 +03:00
10e63537b2 [#602] metrics: Rename blobovnicza size metric
`Size` is not size, but open db size.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-08-17 19:17:35 +00:00
809e97626b [#602] blobovnicza: Fix size counter
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-08-17 19:17:35 +00:00
2e49d7ea7e [#602] blobovnicza: Init before using
Fix blobovnicza size: after restart size metric resets.

Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-08-17 19:17:35 +00:00
f7042c5a6f [#609] Replace zaptest.NewLogger() with zap.L()
Semantic patch:
```
@@
@@
-import "go.uber.org/zap/zaptest"
+import "go.uber.org/zap"

-zaptest.NewLogger(t)
+zap.L()
```

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-08-17 16:51:10 +00:00
e604a3d749 [#607] *: Use zap.Stringer() where possible
Semantic patch:
```
@@
var f expression
var t expression
var a expression
@@
 f(
    ...,
-    zap.String(t, a.String()),
+    zap.Stringer(t, a),
    ...,
)
```

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-08-17 16:50:02 +00:00
a8de37c8a2 [#607] *: Remove redundant if on error returns
Semantic patch:
```
@@
@@
-if err != nil { return err }
-return nil
+return err
```

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-08-17 16:50:02 +00:00
5335e7089e [#615] pilorama: Speedup TestForest_ApplyRandom()
Some of our pilorama tests fail on CI.
The reasons are not obvious, but one possible improvement
is using `WithNoSync` option for these. It should have much effect,
because we are writing on the tmpfs, but doesn't hurt anyway.

If I replace `t.TempDir()` with a local directory, test execution time
goes down from 5s (sync) to 0.4s (nosync), which is the same time as
with `t.TempDir()`. Maybe we have some strange CI configuration.

```
panic: test timed out after 10m0s
running tests:
	TestForest_ApplyRandom (8m22s)
	TestForest_ApplyRandom/bbolt (8m21s)
...
goroutine 170 [syscall]:
syscall.Syscall(0xc000100000?, 0xc00047b758?, 0x6aff9a?, 0xc00041c1b0?)
	/opt/hostedtoolcache/go/1.20.7/x64/src/syscall/syscall_linux.go:69 +0x27
syscall.Fdatasync(0x9e35c0?)
	/opt/hostedtoolcache/go/1.20.7/x64/src/syscall/zsyscall_linux_amd64.go:418 +0x2a
go.etcd.io/bbolt.fdatasync(0xc000189000?)
```

Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
2023-08-16 19:09:08 +00:00
2efe9cc1be [#585] writecache: Fix DB counter
Signed-off-by: Dmitrii Stepanov <d.stepanov@yadro.com>
2023-08-16 14:47:44 +03:00