Commit Graph

141 Commits

Author SHA1 Message Date
Michael Eischer eef0ee7a85 repository: trigger GC after loading the index
Loading the index requires some scratch space, thus make sure that this
memory does not factor into the targeted gc memory usage limit.
2023-06-02 21:56:14 +02:00
Michael Eischer 5773b86d02 repository: Push all usage of errors.Fatal out of the package
As the `Fatal` error type only includes a string, it becomes impossible
to inspect the contained error. This is for a example a problem for the
fuse implementation, which must be able to detect context.Canceled
errors.

Co-authored-by: greatroar <61184462+greatroar@users.noreply.github.com>
2023-05-18 17:27:41 +02:00
greatroar d129baba7a repository: Reuse buffers in Repository.LoadUnpacked
This method had a buffer argument, but that was nil at all call sites.
That's removed, and instead LoadUnpacked now reuses whatever it
allocates inside its retry loop.
2023-01-30 22:01:01 +01:00
Michael Eischer 1adf28a2b5 repository: properly return invalid data error in LoadUnpacked
The retry backend does not return the original error, if its execution
is interrupted by canceling the context. Thus, we have to manually
ensure that the invalid data error gets returned.

Additionally, use the retry backend for some of the repository tests, as
this is the configuration which will be used by restic.
2023-01-14 17:57:02 +01:00
Michael Eischer 6d9675c323 repository: cleanup error message on invalid data
The retry printed the filename twice:
```
Load(<lock/04804cba82>, 0, 0) returned error, retrying after 720.254544ms: load(<lock/04804cba82>): invalid data returned
```
now the warning has changed to
```
Load(<lock/04804cba82>, 0, 0) returned error, retrying after 720.254544ms: invalid data returned
```
2023-01-14 17:57:02 +01:00
Michael Eischer 40ac678252 backend: remove Test method
The Test method was only used in exactly one place, namely when trying
to create a new repository it was used to check whether a config file
already exists.

Use a combination of Stat() and IsNotExist() instead.
2022-12-03 11:28:10 +01:00
Michael Eischer 01f0db4e56 return error if RESTIC_COMPRESSION env variable is invalid 2022-10-29 22:03:39 +02:00
Michael Eischer 8d62a7adb4 identify keys by ID and not name 2022-10-15 16:07:43 +02:00
Michael Eischer 2e3f1c08c5 repository: split index into a separate package 2022-10-08 21:15:34 +02:00
Michael Eischer 5760ba6989
Merge pull request #3949 from MichaelEischer/simplify-mixedpacks
repository: remove IsMixedPack and add replacement for checker
2022-10-08 21:14:14 +02:00
Michael Eischer 4bb5240720 repository: remove unused PrefixLength 2022-10-03 12:15:53 +02:00
Michael Eischer 999fe29976 repository: hide prepareCache 2022-10-03 12:15:53 +02:00
Michael Eischer ddcf549eba repository: remove IsMixedPack and add replacement for checker
Repositories with mixed packs are probably quite rare by now. When
loading data blobs from a mixed pack file, this will no longer trigger
caching that file. However, usually tree blobs are accessed first such
that this shouldn't make much of a difference.

The checker gets a simpler replacement.
2022-10-03 12:03:59 +02:00
Michael Eischer 5c6b6edefe retry index, lock and snapshot loading on hash mismatch 2022-09-25 11:35:35 +02:00
Michael Eischer 78d2312ee9
Merge pull request #3854 from MichaelEischer/sparsefiles
restore: Add support for sparse files
2022-09-24 22:04:02 +02:00
Michael Eischer c147422ba5 repository: special case SaveBlob for all zero chunks
Sparse files contain large regions containing only zero bytes. Checking
that a blob only contains zeros is possible with over 100GB/s for modern
x86 CPUs. Calculating sha256 hashes is only possible with 500MB/s (or
2GB/s using hardware acceleration). Thus we can speed up the hash
calculation for all zero blobs (which always have length
chunker.MinSize) by checking for zero bytes and then using the
precomputed hash.

The all zeros check is only performed for blobs with the minimal chunk
size, and thus should add no overhead most of the time. For chunks which
are not all zero but have the minimal chunks size, the overhead will be
below 2% based on the above performance numbers.

This allows reading sparse sections of files as fast as the kernel can
return data to us. On my system using BTRFS this resulted in about
4GB/s.
2022-09-24 21:39:39 +02:00
Michael Eischer 1ebd57247a repository: optimize MasterIndex.Each
Sending data through a channel at very high frequency is extremely
inefficient. Thus use simple callbacks instead of channels.

> name                old time/op  new time/op  delta
> MasterIndexEach-16   6.68s ±24%   0.96s ± 2%  -85.64%  (p=0.008 n=5+5)
2022-09-24 12:21:59 +02:00
Michael Eischer 7266f07c87 repository: StreamPack in parts if there are too large gaps
For large pack sizes we might be only interested in the first and last
blob of a pack file. Thus stream a pack file in multiple parts if the
gaps between requested blobs grow too large.
2022-08-05 23:48:36 +02:00
Michael Eischer 1b076cda97 rename option to --pack-size 2022-08-05 23:47:43 +02:00
Michael Eischer 0a6fa602c8 add option for setting min pack size 2022-08-05 23:47:12 +02:00
MichaelEischer 443cc49afd
Merge pull request #3830 from MichaelEischer/cleanup-repo
Extract Load/SaveTree/JSONUnpacked from repository
2022-07-23 10:46:13 +02:00
Michael Eischer 9729e6d7ef backend: extract readerat from restic package 2022-07-17 15:29:09 +02:00
Michael Eischer 8c11fc3ec9 crypto: move crypto buffer helpers 2022-07-17 13:42:23 +02:00
Michael Eischer 89d3ce852b repository: extract Load/StoreJSONUnpacked
A Load/Store method for each data type is much clearer. As a result the
repository no longer needs a method to load / store json.
2022-07-17 13:22:00 +02:00
Michael Eischer fbcbd5318c repository: extract LoadTree/SaveTree
The repository has no real idea what a Tree is. So these methods never
belonged there.
2022-07-17 13:11:28 +02:00
Lorenz Bausch d6e3c7f28e
Wording: change repo to repository 2022-07-08 20:05:35 +02:00
Michael Eischer 6f53ecc1ae adapt workers based on whether an operation is CPU or IO-bound
Use runtime.GOMAXPROCS(0) as worker count for CPU-bound tasks,
repo.Connections() for IO-bound task and a combination if a task can be
both. Streaming packs is treated as IO-bound as adding more worker
cannot provide a speedup.

Typical IO-bound tasks are download / uploading / deleting files.
Decoding / Encoding / Verifying are usually CPU-bound. Several tasks are
a combination of both, e.g. for combined download and decode functions.
In the latter case add both limits together. As the backends have their
own concurrency limits restic still won't download more than
repo.Connections() files in parallel, but the additional workers can
decode already downloaded data in parallel.
2022-07-03 12:19:26 +02:00
Michael Eischer 753e56ee29 repository: Limit to a single pending pack file
Use only a single not completed pack file to keep the number of open and
active pack files low. The main change here is to defer hashing the pack
file to the upload step. This prevents the pack assembly step to become
a bottleneck as the only task is now to write data to the temporary pack
file.

The tests are cleaned up to no longer reimplement packer manager
functions.
2022-07-02 22:42:34 +02:00
Michael Eischer 120ccc8754 repository: Rework blob saving to use an async pack uploader
Previously, SaveAndEncrypt would assemble blobs into packs and either
return immediately if the pack is not yet full or upload the pack file
otherwise. The upload will block the current goroutine until it
finishes.

Now, the upload is done using separate goroutines. This requires changes
to the error handling. As uploads are no longer tied to a SaveAndEncrypt
call, failed uploads are signaled using an errgroup.

To count the uploaded amount of data, the pack header overhead is no
longer returned by `packer.Finalize` but rather by
`packer.HeaderOverhead`. This helper method is necessary to continue
returning the pack header overhead directly to the responsible call to
`repository.SaveBlob`. Without the method this would not be possible,
as packs are finalized asynchronously.
2022-07-02 22:42:34 +02:00
Michael Eischer a6e9e08034 Account for pack header overhead at each entry
This will miss the pack header crypto overhead and the length field,
which only amount to a few bytes per pack file.
2022-07-02 18:55:58 +02:00
Alexander Neumann 99634c0936 Return real size from SaveBlob 2022-07-02 18:55:12 +02:00
Michael Eischer ec7c9ce88b drop unused repository.Loader interface 2022-07-02 18:39:59 +02:00
Michael Eischer e68c3a4e62 repository: simplify CreateIndexFromPacks 2022-07-02 18:39:59 +02:00
Michael Eischer bf81bf0795 repository: Properly set id for finalized index
As MergeFinalIndex and index uploads can occur concurrently, it is
necessary for MergeFinalIndex to check whether the IDs for an index were
already set before merging it. Otherwise, we'd loose the ID of an index
which is set _after_ uploading it.
2022-07-02 18:39:59 +02:00
Michael Eischer 628ae799ca repository: make flushPacks private 2022-07-02 18:39:12 +02:00
Michael Eischer a77d5c4d11 repository: index saving belongs into the MasterIndex 2022-07-02 18:38:56 +02:00
greatroar c9557b2822 internal/repository: Fix LoadBlob + fuzz test
When given a buf that is big enough for a compressed blob but not its
decompressed contents, the copy at the end of LoadBlob would skip the
last part of the contents.

Fixes #3783.
2022-06-06 17:02:28 +02:00
greatroar 2e0f1f5113 repository: Remove RunWorkers, report ctx.Err()
This removes RunWorkers, which had become mere overhead by successive
refactors. It also ensures that each former user of that function
returns any context error that occurs, so failure to complete an
operation is always reported as an error.
2022-05-10 22:26:00 +02:00
Michael Eischer cf5cb673fb repository: Use existing method to collect pack ids 2022-04-30 19:14:21 +02:00
Michael Eischer b335cb6285 repository: Refactor index IDs collection 2022-04-30 19:14:21 +02:00
Michael Eischer abe5935693 repository: unify repository version-specific initialization
Mark the master index as compressed also when initializing a new
repository. This is only relevant for testing.
2022-04-30 11:34:10 +02:00
Alexander Neumann 8776031f96 Leave allocating slices to the decompress code 2022-04-30 11:34:10 +02:00
Alexander Neumann 5eb05a0afe Configure zstd encoder/decoder 2022-04-30 11:34:10 +02:00
Michael Eischer 2f36e044db Cleanup pack header check 2022-04-30 11:34:10 +02:00
Alexander Neumann 8b11b86383 Add option global --compression 2022-04-30 11:34:10 +02:00
Michael Eischer 7132df529e repository: Increase index size for repo version 2
A compressed index is only about one third the size of an uncompressed
one. Thus increase the number of entries in an index to avoid cluttering
the repository with small indexes.
2022-04-30 11:34:10 +02:00
Michael Eischer 66f9048bce repository: Alloc zstd encoder/decoder on demand 2022-04-30 11:34:10 +02:00
Michael Eischer 6fb408d90e repository: implement pack compression 2022-04-30 11:34:10 +02:00
Michael Eischer 362ab06023 init: Add flag to specify created repository version 2022-04-30 10:07:42 +02:00
Michael Eischer 4b957e7373 repository: Implement index/snapshot/lock compression
The config file is not compressed as it should remain readable by older
restic versions such that these can return a proper error.

As the old format for unpacked data does not include a version header,
make use of a trick: The old data is always encoded as JSON. Thus it can
only start with '{' or '['. For any other value the first byte indicates
a versioned format. The version is set to 2 for now. Then the zstd
compressed data follows.
2022-04-30 10:07:42 +02:00