Commit graph

168 commits

Author SHA1 Message Date
Alexander Neumann
eddb8549ef backup: By default, do not save the access time
This can be re-enabled with `--with-atime`.
2017-11-28 21:31:35 +01:00
Alexander Neumann
dc38265b54 Merge pull request #1436 from restic/remove-old-cache
Remove old cache directories
2017-11-27 21:37:05 +01:00
Alexander Neumann
0f5e38609f restore: Fix restore to relative path 2017-11-26 18:36:48 +01:00
Alexander Neumann
27d29b9853 restorer: Make sure node names are clean 2017-11-26 15:28:45 +01:00
Alexander Neumann
8a171731ba restorer: Add tests for invalid node names 2017-11-26 15:28:37 +01:00
Alexander Neumann
134abbd82b rest: Use client for creating the repository
Before, creating a new repo via REST would use the defaut HTTP client,
which is not a problem unless the server uses HTTPS and a TLS
certificate which isn't signed by a CA in the system's CA store. In this
case, all commands work except the 'init' command, which fails with a
message like "invalid certificate".
2017-11-25 20:56:40 +01:00
Alexander Neumann
1ebf0e8de8 Merge pull request #1437 from restic/fix-1292
s3: Document and remove default prefix
2017-11-25 11:34:26 +01:00
Alexander Neumann
47b326b7b5 Merge pull request #1423 from harshavardhana/creds
Fix chaining of credentials for minio-go
2017-11-24 21:57:52 +01:00
Alexander Neumann
f79698dcdd Merge pull request #1410 from armhold/deadlock2
unify behavior for max http connections across backends
2017-11-24 21:32:56 +01:00
Alexander Neumann
f4bab789b8 cache: Simplify cache dir creation 2017-11-24 20:53:26 +01:00
Alexander Neumann
fa893ee477 cache: Add detection code for old cache dirs 2017-11-24 20:53:26 +01:00
Alexander Neumann
262b0cd9d4 s3: Remove default prefix "/restic" 2017-11-21 21:33:09 +01:00
Alexander Neumann
e83ec17e95 s3: Correct comment 2017-11-20 22:21:39 +01:00
Harshavardhana
41c8c946ba Fix chaining of credentials for minio-go
chaining failed because chaining provider
was only looking for subsequent credentials
provider after an error. Writer a new
chaining provider which proceeds to fetch
new credentials also under situations where
providers do not return but instead return
no keys at all.

Fixes https://github.com/restic/restic/issues/1422
2017-11-18 02:51:12 -08:00
George Armhold
0268d0e7d6 swift backend: limit http concurrency in Save(), Stat(), Test(), Remove(),
List().

move comment regarding problematic List() backend api (it's s3's ListObjects
that has a problem, NOT swift's ObjectsWalk).

As per discussion in PR #1399.
2017-11-02 18:29:32 -04:00
George Armhold
8515d093e0 swift backend: fix premature release of semaphore in Load() & document
concurrency issue in List().

refactor wrapReader from b2 -> semaphore so it can be used elsewhere.

As per discussion in PR #1399.
2017-11-02 12:38:17 -04:00
George Armhold
99ac0da4bc s3 backend: limit http concurrency in Save(), Stat(), Test(), Remove()
NB: List() is NOT currently limited, as it would cause deadlock due to
be.client.ListObjects() implementation.

as per discussion in PR #1399
2017-11-01 09:40:54 -04:00
Alexander Neumann
6d46824fb0 Pass in a nil buffer to Open() 2017-11-01 10:30:40 +01:00
Alexander Neumann
bb435b39d9 crypto: Rework Seal/Open to use sliceForAppend 2017-11-01 10:30:40 +01:00
Alexander Neumann
2a67d7a6c2 crypto: Correct test function names 2017-11-01 10:30:40 +01:00
Alexander Neumann
ba43c8bab5 crypto: Fix nonce test, make it faster 2017-11-01 10:30:40 +01:00
Alexander Neumann
931e6ed2ac Use Seal/Open everywhere 2017-11-01 10:30:40 +01:00
Alexander Neumann
a5f0e9ab65 Remove custom Encrypt/Decrypt methods 2017-11-01 10:30:40 +01:00
Alexander Neumann
6fc133ad6a Run tests on Seal/Open 2017-11-01 10:30:40 +01:00
Alexander Neumann
e1b80859f2 Make crypto.Key implement cipher.AEAD 2017-11-01 10:30:39 +01:00
George Armhold
d069ee31b2 GS backend: limit http concurrency in Save(), Stat(), Test(), Remove(), List()
as per discussion in PR #1399
2017-10-31 08:01:43 -04:00
George Armhold
981752ade0 Azure backend: limit http concurrency in Stat(), Test(), Remove()
as per discussion in PR #1399
2017-10-31 07:32:30 -04:00
George Armhold
2f8147af59 log unexpected errs from b2 ListCurrentObject()
gh-1385
2017-10-29 08:53:39 -04:00
Alexander Neumann
f854a41ba9 Merge pull request #1399 from armhold/deadlock2
prevent deadlock in List() for B2 when b2.connections=1
2017-10-29 09:26:46 +01:00
George Armhold
3304b0fcf0 prevent deadlock in List() for B2 when b2.connections=1
This is a fix for the following situation (gh-1188):

List() grabs a semaphore token upon entry, starts a goroutine, and
does not release the token until the routine exits (via a defer).

The goroutine iterates over the results from ListCurrentObjects(),
sending them one at a time to a channel, where they are ultimately
processed by be.Load().

Since be.Load() also needs a token, this will result in deadlock if
b2.connections=1.

This fix changes List() so that the token is only held during the call
to ListCurrentObjects().
2017-10-28 18:46:47 -04:00
George Armhold
d8938e259a sftp ReadDir: add path to return error messages (gh-1323)
fix missing "Close" string in debug log fmt
2017-10-28 14:16:27 -04:00
Alexander Neumann
e71db01230 Improve crypto test 2017-10-28 12:09:33 +02:00
Alexander Neumann
178e946fc7 Rename KDFParams -> Params 2017-10-28 10:28:29 +02:00
Alexander Neumann
7a99418dc5 Merge pull request #1393 from armhold/lint-errcheck
detect errors from fs.Walk() in local backend List()
2017-10-28 09:56:11 +02:00
Alexander Neumann
c71ba466ea Merge pull request #1391 from armhold/b2-listmax
pass in defaultListMaxItems to b2Backend constructor
2017-10-28 09:54:57 +02:00
Alexander Neumann
8ce5d35543 Merge pull request #1390 from armhold/lint
small cleanup
2017-10-28 09:54:35 +02:00
George Armhold
8a37c07295 send errors from fs.Walk() to debug log
clarify non-err returns from Walk where err is already proved to be nil
2017-10-27 08:41:17 -04:00
George Armhold
bd0ada7842 go fmt 2017-10-26 16:37:11 -04:00
George Armhold
eea96f652d go fmt 2017-10-26 16:22:10 -04:00
George Armhold
38c3061df7 pass in defaultListMaxItems to b2Backend constructor
gh-1385
2017-10-26 14:22:16 -04:00
George Armhold
f5fa602482 detect and return error from file Close() in Node.createFileAt()
gh-1385
2017-10-26 13:53:31 -04:00
Konrad Wojas
5b96885c6d Control progress rate with RESTIC_PROGRESS_FPS env
Add a RESTIC_PROGRESS_FPS environment variable to limit the interval
at which the progress indicator updates (allowed values: 1-60).

The default rate of 60 FPS can cause high terminal CPU load on some
systems, like iTerm2 on macOS with font anti-aliasing enabled.

Usage:

    RESTIC_PROGRESS_FPS=1 restic ...
    RESTIC_PROGRESS_FPS=60 restic ...
2017-10-26 14:46:56 +08:00
George Armhold
bcdebfb84e small cleanup:
- be explicit when discarding returned errors from .Close(), etc.
- remove named return values from funcs when naked return not used
- fix some "err" shadowing when redeclaration not needed
2017-10-25 12:03:55 -04:00
Alexander Neumann
90b96d19cd Merge pull request #1365 from felix9/fix_1068
Fix failure to detect some legacy s3 repos
2017-10-21 12:19:58 +02:00
Alexander Neumann
d63ab4e9a4 Merge pull request #1358 from prattmic/chunk_size
gs: add option to set chunk size
2017-10-21 11:13:48 +02:00
Felix Lee
944fc857eb Fix failure to detect some legacy s3 repos
Sometimes s3 listobjects for a directory includes an entry for that
directory. The restic s3 backend doesn't expect that and returns
an error.

Symptom is:
  ReadDir: invalid key name restic/key/, removing prefix
     restic/key/ yielded empty string

I'm not sure when s3 does that; I'm unable to reproduce it myself.

But in any case, it seems correct to ignore that when it happens.

Fixes #1068
2017-10-18 13:45:31 -07:00
Michael Pratt
9fa4f5eb6b gs: disable resumable uploads
By default, the GCS Go packages have an internal "chunk size" of 8MB,
used for blob uploads.

Media().Do() will buffer a full 8MB from the io.Reader (or less if EOF
is reached) then write that full 8MB to the network all at once.

This behavior does not play nicely with --limit-upload, which only
limits the Reader passed to Media. While the long-term average upload
rate will be correctly limited, the actual network bandwidth will be
very spikey.

e.g., if an 8MB/s connection is limited to 1MB/s, Media().Do() will
spend 8s reading from the rate-limited reader (performing no network
requests), then 1s writing to the network at 8MB/s.

This is bad for network connections hurt by full-speed uploads,
particularly when writing 8MB will take several seconds.

Disable resumable uploads entirely by setting the chunk size to zero.
This causes the io.Reader to be passed further down the request stack,
where there is less (but still some) buffering.

My connection is around 1.5MB/s up, with nominal ~15ms ping times to
8.8.8.8.

Without this change, --limit-upload 1024 results in several seconds of
~200ms ping times (uploading), followed by several seconds of ~15ms ping
times (reading from rate-limited reader). A bandwidth monitor reports
this as several seconds of ~1.5MB/s followed by several seconds of
0.0MB/s.

With this change, --limit-upload 1024 results in ~20ms ping times and
the bandwidth monitor reports a constant ~1MB/s.

I've elected to make this change unconditional of --limit-upload because
the resumable uploads shouldn't be providing much benefit anyways, as
restic already uploads mostly small blobs and already has a retry
mechanism.

--limit-download is not affected by this problem, as Get().Download()
returns the real http.Response.Body without any internal buffering.

Updates #1216
2017-10-17 21:12:04 -07:00
Alexander Neumann
ce4d71d626 backend: Add partial read failure to error backend 2017-10-17 22:11:38 +02:00
Alexander Neumann
8e2ef3f38b cache: Store first, then store in the backend
Store the file in the backend, then rewind the reader and store it
in the cache afterwards.
2017-10-17 22:01:20 +02:00
Alexander Neumann
8dc952775e backend: Correctly retry Save() calls
Make sure the given reader is an io.Seeker and rewind it properly each
time.
2017-10-17 21:46:38 +02:00