extern/rclone
1
0
Fork 0
mirror of https://github.com/rclone/rclone.git synced 2024-12-22 23:22:08 +01:00
Code Issues Packages Projects Releases Wiki Activity

s3: add provider quirk --s3-might-gzip to fix corrupted on transfer: sizes differ

Browse Source 
Before this change, some files were giving this error when downloaded
from Cloudflare and other providers.

    ERROR corrupted on transfer: sizes differ NNN vs MMM

This is because these providers auto gzips the object when rclone
wasn't expecting it to. (AWS does not gzip objects without their being
uploaded gzipped).

This patch adds a quirk to for fix the problem and a flag to control
it. The quirk `might_gzip` is set to `true` for all providers except
AWS.

See: https://forum.rclone.org/t/s3-error-corrupted-on-transfer-sizes-differ-nnn-vs-mmm/33694/
Fixes: #6533
This commit is contained in:
Nick Craig-Wood 2022-10-28 23:23:29 +01:00
parent 0de74864b6
commit 5b5fdc6bc5
1 changed files with 35 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

36
backend/s3/s3.go
View File

@ -2174,6 +2174,31 @@ can't check the size and hash but the file contents will be decompressed.
`,
Advanced: true,
Default: false,
}, {
Name: "might_gzip",
Help: strings.ReplaceAll(`Set this if the backend might gzip objects.
Normally providers will not alter objects when they are downloaded. If
an object was not uploaded with |Content-Encoding: gzip| then it won't
be set on download.
However some providers may gzip objects even if they weren't uploaded
with |Content-Encoding: gzip| (eg Cloudflare).
A symptom of this would be receiving errors like
ERROR corrupted on transfer: sizes differ NNN vs MMM
If you set this flag and rclone downloads an object with
Content-Encoding: gzip set and chunked transfer encoding, then rclone
will decompress the object on the fly.
If this is set to unset (the default) then rclone will choose
according to the provider setting what to apply, but you can override
rclone's choice here.
`, "|", "`"),
Default: fs.Tristate{},
Advanced: true,
}, {
Name: "no_system_metadata",
Help: `Suppress setting and reading of system metadata`,
@ -2305,6 +2330,7 @@ type Options struct {
Versions bool `config:"versions"`
VersionAt fs.Time `config:"version_at"`
Decompress bool `config:"decompress"`
MightGzip fs.Tristate `config:"might_gzip"`
NoSystemMetadata bool `config:"no_system_metadata"`
}
@ -2665,10 +2691,12 @@ func setQuirks(opt *Options) {
virtualHostStyle = true
urlEncodeListings = true
useMultipartEtag = true
mightGzip = true // assume all providers might gzip until proven otherwise
)
switch opt.Provider {
case "AWS":
// No quirks
mightGzip = false // Never auto gzips objects
case "Alibaba":
useMultipartEtag = false // Alibaba seems to calculate multipart Etags differently from AWS
case "HuaweiOBS":
@ -2782,6 +2810,12 @@ func setQuirks(opt *Options) {
opt.UseMultipartEtag.Valid = true
opt.UseMultipartEtag.Value = useMultipartEtag
}
// set MightGzip if not manually set
if !opt.MightGzip.Valid {
opt.MightGzip.Valid = true
opt.MightGzip.Value = mightGzip
}
}
// setRoot changes the root of the Fs
@ -4867,7 +4901,7 @@ func (o *Object) Open(ctx context.Context, options ...fs.OpenOption) (in io.Read
// Decompress body if necessary
if aws.StringValue(resp.ContentEncoding) == "gzip" {
if o.fs.opt.Decompress {
if o.fs.opt.Decompress || (resp.ContentLength == nil && o.fs.opt.MightGzip.Value) {
return readers.NewGzipReader(resp.Body)
}
o.fs.warnCompressed.Do(func() {