Commit Graph

29 Commits

Author SHA1 Message Date
Michael Eischer 94cbc6392d restore: slightly reduce memory usage while restoring files
The information which target files are contained in a pack file is no
longer necessary after processing a pack.
2023-04-23 11:33:21 +02:00
Mark Herrmann f875a8843d restore: Add progress bar
Co-authored-by: Mark Herrmann <mark.herrmann@mailbox.org>
2023-04-07 12:08:23 +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 5b6a77058a Enable sparseness only conditionally
We can either preallocate storage for a file or sparsify it. This
detects a pack file as sparse if it contains an all zero block or
consists of only one block. As the file sparsification is just an
approximation, hide it behind a `--sparse` parameter.
2022-09-24 21:20:00 +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 6fb408d90e repository: implement pack compression 2022-04-30 11:34:10 +02:00
Michael Eischer 243698680a crypto: Use helpers for size calculations 2022-03-28 22:09:49 +02:00
Michael Eischer 27524979e8 restorer: Remove dead code 2022-02-13 11:43:09 +01:00
Michael Eischer f40abd92fa restorer: convert to use StreamPack 2022-02-12 20:18:25 +01:00
Alexander Weiss 81876d5c1b Simplify cache logic 2021-09-03 21:01:00 +02:00
Michael Eischer e8bbb05328 restore: Correctly handle partial pack download errors
Failed pack/blob downloads should be retried. For blobs that fail
decryption assume that the pack file is really damaged and try to
restore the remaining blobs.
2021-06-29 20:54:16 +02:00
Michael Eischer 47faf69230 restorer: Check dropped error 2021-01-31 18:06:28 +01:00
Alexander Weiss 5e22ae10f1 Add error handling for fileRestorer 2021-01-31 14:22:57 +01:00
Michael Eischer 69d5b4c36b restorer: lower-case variable name 2021-01-03 13:55:59 +01:00
Alexander Weiss 34a33565c8 Fix loadBlob in filerestorer 2021-01-01 08:06:04 +01:00
Alexander Weiss 07b3f65a6f filesrestorer: Re-use buffer 2021-01-01 07:24:46 +01:00
Alexander Weiss 3e0acf1395 restore: Don't save (part of) pack in memory 2021-01-01 07:24:46 +01:00
Alexander Weiss aa7a5f19c2 Use BlobHandle in index methods 2020-11-22 20:41:12 +01:00
Michael Eischer 8cc9514879 restorer: pre-allocate files before loading chunks 2020-09-07 21:41:47 +02:00
Michael Eischer 2e7d475029 Process packs in order of first appearance 2020-09-06 21:25:35 +02:00
aawsome 0fed6a8dfc
Use "pack file" instead of "data file" (#2885)
- changed variable names, especially changed DataFile into PackFile
- changed in some comments
- always use "pack file" in docu
2020-08-16 11:16:38 +02:00
Alexander Weiss 9d1fb94c6c make Lookup() return all blobs
+ simplify syntax
2020-07-25 21:18:34 +02:00
greatroar de5516a90e Fix sync.WaitGroup usage in restorer.fileRestorer 2020-02-27 00:07:49 +01:00
Igor Fedorenko f17ffa0283 restorer: Allow writing target file blobs out of order
Much simpler implementation that guarantees each required pack
is downloaded only once (and hence does not need to manage
pack cache). Also improves large file restore performance.

Signed-off-by: Igor Fedorenko <igor@ifedorenko.com>
2020-02-26 16:14:45 +01:00
Igor Fedorenko bf9a507148 Fix rare 'file already closed' during restore
Fixes #2183

Signed-off-by: Igor Fedorenko <igor@ifedorenko.com>
2019-03-16 17:01:20 +01:00
Igor Fedorenko bda8d7722e restorer: Optimize empty file restore
don't create fileInfo structs for empty files. this saves memory.
this also avoids extra serial scan of all fileInfo, which should
make restore faster and more consistent.

Signed-off-by: Igor Fedorenko <igor@ifedorenko.com>
2018-10-14 17:39:42 +02:00
Igor Fedorenko 9f3ca97ee8 restore: Chang fileInfo to use snapshot location instead of target path
* uses less memory as common prefix is only stored once
* stepping stone for simpler error callback api, which
  will allow further memory footprint reduction

Signed-off-by: Igor Fedorenko <igor@ifedorenko.com>
2018-10-14 17:39:42 +02:00
Igor Fedorenko da57302fca restore: Removed legacy restore implementation
Signed-off-by: Igor Fedorenko <igor@ifedorenko.com>
2018-10-14 17:37:30 +02:00
Igor Fedorenko 1869930d95 restore: New optimized multithreaded implementation
Signed-off-by: Igor Fedorenko <igor@ifedorenko.com>
2018-10-14 17:37:30 +02:00