For objects with `N` prefix depth, this PR reduces `N` such network
operations by converting `CheckFile` into a single bulk operation.
Reduction in chattiness here would allow disks to be utilized more
cleanly, while maintaining the same functionality along with one
extra volume check stat() call is removed.
Update tests to test multiple sets scenario
Fixes support for using multiple base DNs for user search in the LDAP directory
allowing users from different subtrees in the LDAP hierarchy to request
credentials.
- The username in the produced credentials is now the full DN of the LDAP user
to disambiguate users in different base DNs.
parentDirIsObject is not using set level understanding
to check for parent objects, without this it can lead to
objects that can actually reside on a separate set as
objects and would conflict.
Current implementation requires server pools to have
same erasure stripe sizes, to facilitate same SLA
and expectations.
This PR allows server pools to be variadic, i.e they
do not have to be same erasure stripe sizes - instead
they should have SLA for parity ratio.
If the parity ratio cannot be guaranteed by the new
server pool, the deployment is rejected i.e server
pool expansion is not allowed.
This ensures that all the prometheus monitoring and usage
trackers to avoid alerts configured, although we cannot
support v1 to v2 here - we can v2 to v3.
A lot of memory is consumed when uploading small files in parallel, use
the default upload parameters and add MINIO_AZURE_UPLOAD_CONCURRENCY for
users to tweak.
Synchronous replication can be enabled by setting the --sync
flag while adding a remote replication target.
This PR also adds proxying on GET/HEAD to another node in a
active-active replication setup in the event of a 404 on the current node.
This PR refactors the way we use buffers for O_DIRECT and
to re-use those buffers for messagepack reader writer.
After some extensive benchmarking found that not all objects
have this benefit, and only objects smaller than 64KiB see
this benefit overall.
Benefits are seen from almost all objects from
1KiB - 32KiB
Beyond this no objects see benefit with bulk call approach
as the latency of bytes sent over the wire v/s streaming
content directly from disk negate each other with no
remarkable benefits.
All other optimizations include reuse of msgp.Reader,
msgp.Writer using sync.Pool's for all internode calls.
30 seconds white spaces is long for some setups which time out when no
read activity in short time, reduce the subnet health white space ticker
to 5 seconds, since it has no cost at all.
Use separate sync.Pool for writes/reads
Avoid passing buffers for io.CopyBuffer()
if the writer or reader implement io.WriteTo or io.ReadFrom
respectively then its useless for sync.Pool to allocate
buffers on its own since that will be completely ignored
by the io.CopyBuffer Go implementation.
Improve this wherever we see this to be optimal.
This allows us to be more efficient on memory usage.
```
385 // copyBuffer is the actual implementation of Copy and CopyBuffer.
386 // if buf is nil, one is allocated.
387 func copyBuffer(dst Writer, src Reader, buf []byte) (written int64, err error) {
388 // If the reader has a WriteTo method, use it to do the copy.
389 // Avoids an allocation and a copy.
390 if wt, ok := src.(WriterTo); ok {
391 return wt.WriteTo(dst)
392 }
393 // Similarly, if the writer has a ReadFrom method, use it to do the copy.
394 if rt, ok := dst.(ReaderFrom); ok {
395 return rt.ReadFrom(src)
396 }
```
From readahead package
```
// WriteTo writes data to w until there's no more data to write or when an error occurs.
// The return value n is the number of bytes written.
// Any error encountered during the write is also returned.
func (a *reader) WriteTo(w io.Writer) (n int64, err error) {
if a.err != nil {
return 0, a.err
}
n = 0
for {
err = a.fill()
if err != nil {
return n, err
}
n2, err := w.Write(a.cur.buffer())
a.cur.inc(n2)
n += int64(n2)
if err != nil {
return n, err
}
```
with changes present to automatically throttle crawler
at runtime, there is no need to have an environment
value to disable crawling. crawling is a fundamental
piece for healing, lifecycle and many other features
there is no good reason anyone would need to disable
this on a production system.
* Apply suggestions from code review
globalSubscribers.NumSubscribers() is heavily used in S3 requests and it
uses mutex, use atomic.Load instead since it is faster
Co-authored-by: Anis Elleuch <anis@min.io>
Rewrite parentIsObject() function. Currently if a client uploads
a/b/c/d, we always check if c, b, a are actual objects or not.
The new code will check with the reverse order and quickly quit if
the segment doesn't exist.
So if a, b, c in 'a/b/c' does not exist in the first place, then returns
false quickly.
The only purpose of check-dir flag in
ReadVersion is to return 404 when
an object has xl.meta but without data.
This is causing an extract call to the disk
which can be penalizing in case of busy system
where disks receive many concurrent access.
mc admin trace does not show the correct handler name in the output: it
is printing `maxClients' for all handlers. The reason is that the wrong
order of handler wrapping.
Fixes two problems
- Double healing when bitrot is enabled, instead heal attempt
once in applyActions() before lifecycle is applied.
- If applyActions() is successful and getSize() returns proper
value, then object is accounted for and should be removed
from the oldCache namespace map to avoid double heal attempts.
main reason is that HealObjects starts a recursive listing
for each object, this can be a really really long time on
large namespaces instead avoid recursive listing just
perform HealObject() instead at the prefix.
delete's already handle purging dangling content, we
don't need to achieve this by doing recursive listing,
this in-turn can delay crawling significantly.
Optimizations include
- do not write the metacache block if the size of the
block is '0' and it is the first block - where listing
is attempted for a transient prefix, this helps to
avoid creating lots of empty metacache entries for
`minioMetaBucket`
- avoid the entire initialization sequence of cacheCh
, metacacheBlockWriter if we are simply going to skip
them when discardResults is set to true.
- No need to hold write locks while writing metacache
blocks - each block is unique, per bucket, per prefix
and also is written by a single node.
current implementation was incorrect, it in-fact
assumed only read quorum number of disks. in-fact
that value is only meant for read quorum good entries
from all online disks.
This PR fixes this behavior properly.
This commit refactors the code in `cmd/crypto`
and separates SSE-S3, SSE-C and SSE-KMS.
This commit should not cause any behavior change
except for:
- `IsRequested(http.Header)`
which now returns the requested type {SSE-C, SSE-S3,
SSE-KMS} and does not consider SSE-C copy headers.
However, SSE-C copy headers alone are anyway not valid.
Additional cases handled
- fix address situations where healing is not
triggered on failed writes and deletes.
- consider object exists during listing when
metadata can be successfully decoded.
always find the right set of online peers for remote listing,
this may have an effect on listing if the server is down - we
should do this to avoid always performing transient operations
on bucket->peerClient that is permanently or down for a long
period.
additionally also configure http2 healthcheck
values to quickly detect unstable connections
and let them timeout.
also use single transport for proxying requests
with missing nextMarker with delimiter based listing,
top level prefixes beyond 4500 or max-keys value
wouldn't be sent back for client to ask for the next
batch.
reproduced at a customer deployment, create prefixes
as shown below
```
for year in $(seq 2017 2020)
do
for month in {01..12}
do for day in {01..31}
do
mc -q cp file myminio/testbucket/dir/day_id=$year-$month-$day/;
done
done
done
```
Then perform
```
aws s3api --profile minio --endpoint-url http://localhost:9000 list-objects \
--bucket testbucket --prefix dir/ --delimiter / --max-keys 1000
```
You shall see missing NextMarker, this would disallow listing beyond max-keys
requested and also disallow beyond 4500 (maxKeyObjectList) prefixes being listed
because client wouldn't know the NextMarker available.
This PR addresses this situation properly by making the implementation
more spec compatible. i.e NextMarker in-fact can be either an object, a prefix
with delimiter depending on the input operation.
This issue was introduced after the list caching changes and has been present
for a while.
issue was introduced in #11106 the following
pattern
<-t.C // timer fired
if !t.Stop() {
<-t.C // timer hangs
}
Seems to hang at the last `t.C` line, this
issue happens because a fired timer cannot be
Stopped() anymore and t.Stop() returns `false`
leading to confusing state of usage.
Refactor the code such that use timers appropriately
with exact requirements in place.
Both Azure & S3 gateways call for object information before returning
the stream of the object, however, the object content/length could be
modified meanwhile, which means it can return a corrupted object.
Use ETag to ensure that the object was not modified during the GET call
crawler should only ListBuckets once not for each serverPool,
buckets are same across all pools, across sets and ListBuckets
always returns an unified view, once list buckets returns
sort it by create time to scan the latest buckets earlier
with the assumption that latest buckets would have lesser
content than older buckets allowing them to be scanned faster
and also to be able to provide more closer to latest view.
When searching the caches don't copy the ids, instead inline the loop.
```
Benchmark_bucketMetacache_findCache-32 19200 63490 ns/op 8303 B/op 5 allocs/op
Benchmark_bucketMetacache_findCache-32 20338 58609 ns/op 111 B/op 4 allocs/op
```
Add a reasonable, but still the simplistic benchmark.
Bonus - make nicer zero alloc logging
With new refactor of bucket healing, healing bucket happens
automatically including its metadata, there is no need to
redundant heal buckets also in ListBucketsHeal remove
it.
optimization mainly to avoid listing the entire
`.minio.sys/buckets/.minio.sys` directory, this
can get really huge and comes in the way of startup
routines, contents inside `.minio.sys/buckets/.minio.sys`
are rather transient and not necessary to be healed.
Tests environments (go test or manual testing) should always consider
the passed disks are root disks and should not rely on disk.IsRootDisk()
function. The reason is that this latter can return a false negative
when called in a busy system. However, returning a false negative will
only occur in a testing environment and not in a production, so we can
accept this trade-off for now.
Perform cleanup operations on copied data. Avoids read locking
data while determining which caches to keep.
Also, reduce the log(N*N) operation to log(N*M) where M caches
with the same root or below when checking potential replacements.
This refactor is done for few reasons below
- to avoid deadlocks in scenarios when number
of nodes are smaller < actual erasure stripe
count where in N participating local lockers
can lead to deadlocks across systems.
- avoids expiry routines to run 1000 of separate
network operations and routes per disk where
as each of them are still accessing one single
local entity.
- it is ideal to have since globalLockServer
per instance.
- In a 32node deployment however, each server
group is still concentrated towards the
same set of lockers that partipicate during
the write/read phase, unlike previous minio/dsync
implementation - this potentially avoids send
32 requests instead we will still send at max
requests of unique nodes participating in a
write/read phase.
- reduces overall chattiness on smaller setups.
The bucket forwarder handler considers MakeBucket to be always local but
it mistakenly thinks that PUT bucket lifecycle to be a MakeBucket call.
Fix the check of the MakeBucket call by ensuring that the query is empty
in the PUT url.
till now we used to match the inode number of the root
drive and the drive path minio would use, if they match
we knew that its a root disk.
this may not be true in all situations such as running
inside a container environment where the container might
be mounted from a different partition altogether, root
disk detection might fail.
partNumber was miscalculting the start and end of parts when partNumber
query is specified in the GET request. This commit fixes it and also
fixes the ContentRange header in that case.
PR 038bcd9079 introduced
version '3', we need to make sure that we do not
print an unexpected error instead log a message to
indicate we will auto update the version.
Due to botched upstream renames of project repositories
and incomplete migration to go.mod support, our current
dependency version of `go.mod` had bugs i.e it was
using commits from master branch which didn't have
the required fixes present in release-3.4 branches
which leads to some rare bugs
https://github.com/etcd-io/etcd/pull/11477 provides
a workaround for now and we should migrate to this.
release-3.5 eventually claims to fix all of this
properly until then we cannot use /v3 import right now
supports `mc admin config set <alias> heal sleep=100ms` to
enable more aggressive healing under certain times.
also optimize some areas that were doing extra checks than
necessary when bitrotscan was enabled, avoid double sleeps
make healing more predictable.
fixes#10497
- accountInfo API that returns information about
user, access to buckets and the size per bucket
- addUser - user is allowed to change their secretKey
- getUserInfo - returns user info if the incoming
is the same user requesting their information
In some cases a writer could be left behind unclosed, leaking compression blocks.
Always close and set compression concurrency to 2 which should be fine to keep up.
Due to https://github.com/philhofer/fwd/issues/20 when skipping a metadata entry that is >2048 bytes and the buffer is full (2048 bytes) the skip will fail with `io.ErrNoProgress`.
Enlarge the buffer so we temporarily make this much more unlikely.
If it still happens we will have to rewrite the skips to reads.
Fixes#10959
dangling object when deleted means object doesn't exist
anymore, so we should return appropriate errors, this
allows crawler heal to ensure that it removes the tracker
for dangling objects.
AZURE_STORAGE_ACCOUNT and AZURE_STORAGE_KEY are used in
azure CLI to specify the azure blob storage access & secret keys. With this commit,
it is possible to set them if you want the gateway's own credentials to be
different from the Azure blob credentials.
Co-authored-by: Harshavardhana <harsha@minio.io>
dangling objects when removed `mc admin heal -r` or crawler
auto heal would incorrectly return error - this can interfere
with usage calculation as the entry size for this would be
returned as `0`, instead upon success use the resultant
object size to calculate the final size for the object
and avoid reporting this in the log messages
Also do not set ObjectSize in healResultItem to be '-1'
this has an effect on crawler metrics calculating 1 byte
less for objects which seem to be missing their `xl.meta`
X-Minio-Replication-Delete-Status header shows the
status of the replication of a permanent delete of a version.
All GETs are disallowed and return 405 on this object version.
In the case of replicating delete markers.
X-Minio-Replication-DeleteMarker-Status shows the status
of replication, and would similarly return 405.
Additionally, this PR adds reporting of delete marker event completion
and updates documentation
Alternative to #10927
Instead of having an upstream fix, do unwrap when checking network errors.
'As' will also work when destination is an interface as checked by the tests.
This PR adds transition support for ILM
to transition data to another MinIO target
represented by a storage class ARN. Subsequent
GET or HEAD for that object will be streamed from
the transition tier. If PostRestoreObject API is
invoked, the transitioned object can be restored for
duration specified to the source cluster.
allow directories to be replicated as well, along with
their delete markers in replication.
Bonus fix to fix bloom filter updates for directories
to be preserved.
fixes a regression introduced in #10859, due
to the error returned by rest.Client being typed
i.e *rest.NetworkError - IsNetworkHostDown function
didn't work as expected to detect network issues.
This in-turn aggravated the situations when nodes
are disconnected leading to performance loss.
Do listings with prefix filter when bloom filter is dirty.
This will forward the prefix filter to the lister which will make it
only scan the folders/objects with the specified prefix.
If we have a clean bloom filter we try to build a more generally
useful cache so in that case, we will list all objects/folders.
Add shortcut for `APN/1.0 Veeam/1.0 Backup/10.0`
It requests unique blocks with a specific prefix. We skip
scanning the parent directory for more objects matching the prefix.
Allow each crawler operation to sleep up to 10 seconds on very heavily loaded systems.
This will of course make minimum crawler speed less, but should be more effective at stopping.
Delete marker replication is implemented for V2
configuration specified in AWS spec (though AWS
allows it only in the V1 configuration).
This PR also brings in a MinIO only extension of
replicating permanent deletes, i.e. deletes specifying
version id are replicated to target cluster.
This will make the health check clients 'silent'.
Use `IsNetworkOrHostDown` determine if network is ok so it mimics the functionality in the actual client.
this is needed such that we make sure to heal the
users, policies and bucket metadata right away as
we do listing based on list cache which only lists
'3' sufficiently good drives, to avoid possibly
losing access to these users upon upgrade make
sure to heal them.
If a scanning server shuts down unexpectedly we may have "successful" caches that are incomplete on a set.
In this case mark the cache with an error so it will no longer be handed out.
Add `MINIO_API_EXTEND_LIST_CACHE_LIFE` that will extend
the life of generated caches for a while.
This changes caches to remain valid until no updates have been
received for the specified time plus a fixed margin.
This also changes the caches from being invalidated when the *first*
set finishes until the *last* set has finished plus the specified time
has passed.
Similar to #10775 for fewer memory allocations, since we use
getOnlineDisks() extensively for listing we should optimize it
further.
Additionally, remove all unused walkers from the storage layer
A new field called AccessKey is added to the ReqInfo struct and populated.
Because ReqInfo is added to the context, this allows the AccessKey to be
accessed from 3rd-party code, such as a custom ObjectLayer.
Co-authored-by: Harshavardhana <harsha@minio.io>
Co-authored-by: Kaloyan Raev <kaloyan@storj.io>
On extremely long running listings keep the transient list 15 minutes after last update instead of using start time.
Also don't do overlap checks on transient lists.
Add trashcan that keeps recently updated lists after bucket deletion.
All caches were deleted once a bucket was deleted, so caches still running would report errors. Now they are canceled.
Fix `.minio.sys` not being transient.
Bonus fixes, remove package retry it is harder to get it
right, also manage context remove it such that we don't have
to rely on it anymore instead use a simple Jitter retry.
WriteAll saw 127GB allocs in a 5 minute timeframe for 4MiB buffers
used by `io.CopyBuffer` even if they are pooled.
Since all writers appear to write byte buffers, just send those
instead and write directly. The files are opened through the `os`
package so they have no special properties anyway.
This removes the alloc and copy for each operation.
REST sends content length so a precise alloc can be made.
this reduces allocations in order of magnitude
Also, revert "erasure: delete dangling objects automatically (#10765)"
affects list caching should be investigated.
Add store and a forward option for a single part
uploads when an async mode is enabled with env
MINIO_CACHE_COMMIT=writeback
It defaults to `writethrough` if unspecified.
Bonus fixes, we do not need reload format anymore
as the replaced drive is healed locally we only need
to ensure that drive heal reloads the drive properly.
We preserve the UUID of the original order, this means
that the replacement in `format.json` doesn't mean that
the drive needs to be reloaded into memory anymore.
fixes#10791
when server is booting up there is a possibility
that users might see '503' because object layer
when not initialized, then the request is proxied
to neighboring peers first one which is online.
* Fix caches having EOF marked as a failure.
* Simplify cache updates.
* Provide context for checkMetacacheState failures.
* Log 499 when the client disconnects.
`decryptObjectInfo` is a significant bottleneck when listing objects.
Reduce the allocations for a significant speedup.
https://github.com/minio/sio/pull/40
```
λ benchcmp before.txt after.txt
benchmark old ns/op new ns/op delta
Benchmark_decryptObjectInfo-32 24260928 808656 -96.67%
benchmark old MB/s new MB/s speedup
Benchmark_decryptObjectInfo-32 0.04 1.24 31.00x
benchmark old allocs new allocs delta
Benchmark_decryptObjectInfo-32 75112 48996 -34.77%
benchmark old bytes new bytes delta
Benchmark_decryptObjectInfo-32 287694772 4228076 -98.53%
```
Design: https://gist.github.com/klauspost/025c09b48ed4a1293c917cecfabdf21c
Gist of improvements:
* Cross-server caching and listing will use the same data across servers and requests.
* Lists can be arbitrarily resumed at a constant speed.
* Metadata for all files scanned is stored for streaming retrieval.
* The existing bloom filters controlled by the crawler is used for validating caches.
* Concurrent requests for the same data (or parts of it) will not spawn additional walkers.
* Listing a subdirectory of an existing recursive cache will use the cache.
* All listing operations are fully streamable so the number of objects in a bucket no
longer dictates the amount of memory.
* Listings can be handled by any server within the cluster.
* Caches are cleaned up when out of date or superseded by a more recent one.
only newly replaced drives get the new `format.json`,
this avoids disks reloading their in-memory reference
format, ensures that drives are online without
reloading the in-memory reference format.
keeping reference format in-tact means UUIDs
never change once they are formatted.
lockers currently might leave stale lockers,
in unknown ways waiting for downed lockers.
locker check interval is high enough to safely
cleanup stale locks.
reference format should be source of truth
for inconsistent drives which reconnect,
add them back to their original position
remove automatic fix for existing offline
disk uuids
Bonus fixes
- logging improvements to ensure that we don't use
`go logger.LogIf` to avoid runtime.Caller missing
the function name. log where necessary.
- remove unused code at erasure sets
Test TestDialContextWithDNSCacheRand was failing sometimes because it depends
on a random selection of addresses when testing random DNS resolution from cache.
Lower addr selection exception to 10%
Allow requests to come in for users as soon as object
layer and config are initialized, this allows users
to be authenticated sooner and would succeed automatically
on servers which are yet to fully initialize.
Go stdlib resolver doesn't support caching DNS
resolutions, since we compile with CGO disabled
we are more probe to DNS flooding for all network
calls to resolve for DNS from the DNS server.
Under various containerized environments such as
VMWare this becomes a problem because there are
no DNS caches available and we may end up overloading
the kube-dns resolver under concurrent I/O.
To circumvent this issue implement a DNSCache resolver
which resolves DNS and caches them for around 10secs
with every 3sec invalidation attempted.
connect disks pre-emptively upon startup, to ensure we have
enough disks are connected at startup rather than wait
for them.
we need to do this to avoid long wait times for server to
be online when we have servers come up in rolling upgrade
fashion
Only use dynamic delays for the crawler. Even though the max wait was 1 second the number
of waits could severely impact crawler speed.
Instead of relying on a global metric, we use the stateless local delays to keep the crawler
running at a speed more adjusted to current conditions.
The only case we keep it is before bitrot checks when enabled.
This PR fixes a hang which occurs quite commonly at higher concurrency
by allowing following changes
- allowing lower connections in time_wait allows faster socket open's
- lower idle connection timeout to ensure that we let kernel
reclaim the time_wait connections quickly
- increase somaxconn to 4096 instead of 2048 to allow larger tcp
syn backlogs.
fixes#10413
This change tracks bandwidth for a bucket and object
- [x] Add Admin API
- [x] Add Peer API
- [x] Add BW throttling
- [x] Admin APIs to set replication limit
- [x] Admin APIs for fetch bandwidth
In almost all scenarios MinIO now is
mostly ready for all sub-systems
independently, safe-mode is not useful
anymore and do not serve its original
intended purpose.
allow server to be fully functional
even with config partially configured,
this is to cater for availability of actual
I/O v/s manually fixing the server.
In k8s like environments it will never make
sense to take pod into safe-mode state,
because there is no real access to perform
any remote operation on them.
- select lockers which are non-local and online to have
affinity towards remote servers for lock contention
- optimize lock retry interval to avoid sending too many
messages during lock contention, reduces average CPU
usage as well
- if bucket is not set, when deleteObject fails make sure
setPutObjHeaders() honors lifecycle only if bucket name
is set.
- fix top locks to list out always the oldest lockers always,
avoid getting bogged down into map's unordered nature.
This is to allow remote targets to be generalized
for replication/ILM transition
Also adding a field in BucketTarget to identify
a remote target with a label.
This commit fixes a misuse of the `http.ResponseWriter.WriteHeader`.
A caller should **either** call `WriteHeader` exactly once **or**
write to the response writer and causing an implicit 200 OK.
Writing the response headers more than once causes a `http: superfluous
response.WriteHeader call` log message. This commit fixes this
by preventing a 2nd `WriteHeader` call being forwarded to the underlying
`ResponseWriter`.
Updates #10587
* add NVMe drive info [model num, serial num, drive temp. etc.]
* Ignore fuse partitions
* Add the nvme logic only for linux
* Move smart/nvme structs to a separate file
Co-authored-by: wlan0 <sidharthamn@gmail.com>
throw proper error when port is not accessible
for the regular user, this is possibly a regression.
```
ERROR Unable to start the server: Insufficient permissions to use specified port
> Please ensure MinIO binary has 'cap_net_bind_service=+ep' permissions
HINT:
Use 'sudo setcap cap_net_bind_service=+ep /path/to/minio' to provide sufficient permissions
```
After #10594 let's invalidate the bloom filters to force the next cycles to go through all data.
There is a small chance that the linked PR could have caused missing bloom filter data.
This will invalidate the current bloom filters and make the crawler go through everything.
Routing using on source IP if found. This should distribute
the listing load for V1 and versioning on multiple nodes
evenly between different clients.
If source IP is not found from the http request header, then falls back
to bucket name instead.
Disallow versioning suspension on a bucket with
pre-existing replication configuration
If versioning is suspended on the target,replication
should fail.
`mc admin info` on busy setups will not move HDD
heads unnecessarily for repeated calls, provides
a better responsiveness for the call overall.
Bonus change allow listTolerancePerSet be N-1
for good entries, to avoid skipping entries
for some reason one of the disk went offline.
add a hint on the disk to allow for tracking fresh disk
being healed, to allow for restartable heals, and also
use this as a way to track and remove disks.
There are more pending changes where we should move
all the disk formatting logic to backend drives, this
PR doesn't deal with this refactor instead makes it
easier to track healing in the future.
- Add owner information for expiry, locking, unlocking a resource
- TopLocks returns now locks in quorum by default, provides
a way to capture stale locks as well with `?stale=true`
- Simplify the quorum handling for locks to avoid from storage
class, because there were challenges to make it consistent
across all situations.
- And other tiny simplifications to reset locks.
context canceled errors bubbling up from the network
layer has the potential to be misconstrued as network
errors, taking prematurely a server offline and triggering
a health check routine avoid this potential occurrence.
isEnded() was incorrectly calculating if the current healing sequence is
ended or not. h.currentStatus.Items could be empty if healing is very
slow and mc admin heal consumed all items.
As the bulk/recursive delete will require multiple connections to open at an instance,
The default open connections limit will be reached which results in the following error
```FATAL: sorry, too many clients already```
By setting the open connections to a reasonable value - `2`, We ensure that the max open connections
will not be exhausted and lie under bounds.
The queries are simple inserts/updates/deletes which is operational and sufficient with the
the maximum open connection limit is 2.
Fixes#10553
Allow user configuration for MaxOpenConnections
It is possible the heal drives are not reported from
the maintenance check because the background heal
state simply relied on the `format.json` for capturing
unformatted drives. It is possible that drives might
be still healing - make sure that applications which
rely on cluster health check respond back this detail.
Also, revamp the way ListBuckets work make few portions
of the healing logic parallel
- walk objects for healing disks in parallel
- collect the list of buckets in parallel across drives
- provide consistent view for listBuckets()
Healing was not working correctly in the distributed mode because
errFileVersionNotFound was not properly converted in storage rest
client.
Besides, fixing the healing delete marker is not working as expected.
performance improves by around 100x or more
```
go test -v -run NONE -bench BenchmarkGetPartFile
goos: linux
goarch: amd64
pkg: github.com/minio/minio/cmd
BenchmarkGetPartFileWithTrie
BenchmarkGetPartFileWithTrie-4 1000000000 0.140 ns/op 0 B/op 0 allocs/op
PASS
ok github.com/minio/minio/cmd 1.737s
```
fixes#10520
* Fix cases where minimum timeout > default timeout.
* Add defensive code for too small/negative timeouts.
* Never set timeout below the maximum value of a request.
* Protect against (unlikely) int64 wraps.
* Decrease timeout slower.
* Don't re-lock before copying.
This bug was introduced in 14f0047295
almost 3yrs ago, as a side affect of removing stale `fs.json`
but we in-fact end up removing existing good `fs.json` for an
existing object, leading to some form of a data loss.
fixes#10496
from 20s for 10000 parts to less than 1sec
Without the patch
```
~ time aws --endpoint-url=http://localhost:9000 --profile minio s3api \
list-parts --bucket testbucket --key test \
--upload-id c1cd1f50-ea9a-4824-881c-63b5de95315a
real 0m20.394s
user 0m0.589s
sys 0m0.174s
```
With the patch
```
~ time aws --endpoint-url=http://localhost:9000 --profile minio s3api \
list-parts --bucket testbucket --key test \
--upload-id c1cd1f50-ea9a-4824-881c-63b5de95315a
real 0m0.891s
user 0m0.624s
sys 0m0.182s
```
fixes#10503
It was observed in VMware vsphere environment during a
pod replacement, `mc admin info` might report incorrect
offline nodes for the replaced drive. This issue eventually
goes away but requires quite a lot of time for all servers
to be in sync.
This PR fixes this behavior properly.
If the ILM document requires removing noncurrent versions, the
the server should be able to remove 'null' versions as well.
'null' versions are created when versioning is not enabled
or suspended.
The entire encryption layer is dependent on the fact that
KMS should be configured for S3 encryption to work properly
and we only support passing the headers as is to the backend
for encryption only if KMS is configured.
Make sure that this predictability is maintained, currently
the code was allowing encryption to go through and fail
at later to indicate that KMS was not configured. We should
simply reply "NotImplemented" if KMS is not configured, this
allows clients to simply proceed with their tests.
This is to ensure that Go contexts work properly, after some
interesting experiments I found that Go net/http doesn't
cancel the context when Body is non-zero and hasn't been
read till EOF.
The following gist explains this, this can lead to pile up
of go-routines on the server which will never be canceled
and will die at a really later point in time, which can
simply overwhelm the server.
https://gist.github.com/harshavardhana/c51dcfd055780eaeb71db54f9c589150
To avoid this refactor the locking such that we take locks after we
have started reading from the body and only take locks when needed.
Also, remove contextReader as it's not useful, doesn't work as expected
context is not canceled until the body reaches EOF so there is no point
in wrapping it with context and putting a `select {` on it which
can unnecessarily increase the CPU overhead.
We will still use the context to cancel the lockers etc.
Additional simplification in the locker code to avoid timers
as re-using them is a complicated ordeal avoid them in
the hot path, since locking is very common this may avoid
lots of allocations.
configurable remote transport timeouts for some special cases
where this value needs to be bumped to a higher value when
transferring large data between federated instances.
In `(*cacheObjects).GetObjectNInfo` copy the metadata before spawning a goroutine.
Clean up a few map[string]string copies as well, reducing allocs and simplifying the code.
Fixes#10426
From https://docs.aws.amazon.com/AmazonS3/latest/dev/intro-lifecycle-rules.html#intro-lifecycle-rules-actions
```
When specifying the number of days in the NoncurrentVersionTransition
and NoncurrentVersionExpiration actions in a Lifecycle configuration,
note the following:
It is the number of days from when the version of the object becomes
noncurrent (that is, when the object is overwritten or deleted), that
Amazon S3 will perform the action on the specified object or objects.
Amazon S3 calculates the time by adding the number of days specified in
the rule to the time when the new successor version of the object is
created and rounding the resulting time to the next day midnight UTC.
For example, in your bucket, suppose that you have a current version of
an object that was created at 1/1/2014 10:30 AM UTC. If the new version
of the object that replaces the current version is created at 1/15/2014
10:30 AM UTC, and you specify 3 days in a transition rule, the
transition date of the object is calculated as 1/19/2014 00:00 UTC.
```
This PR adds a DNS target that ensures to update an entry
into Kubernetes operator when a bucket is created or deleted.
See minio/operator#264 for details.
Co-authored-by: Harshavardhana <harsha@minio.io>
MaxConnsPerHost can potentially hang a call without any
way to timeout, we do not need this setting for our proxy
and gateway implementations instead IdleConn settings are
good enough.
Also ensure to use NewRequestWithContext and make sure to
take the disks offline only for network errors.
Fixes#10304
inconsistent drive healing when one of the drive is offline
while a new drive was replaced, this change is to ensure
that we can add the offline drive back into the mix by
healing it again.
Add context to all (non-trivial) calls to the storage layer.
Contexts are propagated through the REST client.
- `context.TODO()` is left in place for the places where it needs to be added to the caller.
- `endWalkCh` could probably be removed from the walkers, but no changes so far.
The "dangerous" part is that now a caller disconnecting *will* propagate down, so a
"delete" operation will now be interrupted. In some cases we might want to disconnect
this functionality so the operation completes if it has started, leaving the system in a cleaner state.
This commit refactors the certificate management implementation
in the `certs` package such that multiple certificates can be
specified at the same time. Therefore, the following layout of
the `certs/` directory is expected:
```
certs/
│
├─ public.crt
├─ private.key
├─ CAs/ // CAs directory is ignored
│ │
│ ...
│
├─ example.com/
│ │
│ ├─ public.crt
│ └─ private.key
└─ foobar.org/
│
├─ public.crt
└─ private.key
...
```
However, directory names like `example.com` are just for human
readability/organization and don't have any meaning w.r.t whether
a particular certificate is served or not. This decision is made based
on the SNI sent by the client and the SAN of the certificate.
***
The `Manager` will pick a certificate based on the client trying
to establish a TLS connection. In particular, it looks at the client
hello (i.e. SNI) to determine which host the client tries to access.
If the manager can find a certificate that matches the SNI it
returns this certificate to the client.
However, the client may choose to not send an SNI or tries to access
a server directly via IP (`https://<ip>:<port>`). In this case, we
cannot use the SNI to determine which certificate to serve. However,
we also should not pick "the first" certificate that would be accepted
by the client (based on crypto. parameters - like a signature algorithm)
because it may be an internal certificate that contains internal hostnames.
We would disclose internal infrastructure details doing so.
Therefore, the `Manager` returns the "default" certificate when the
client does not specify an SNI. The default certificate the top-level
`public.crt` - i.e. `certs/public.crt`.
This approach has some consequences:
- It's the operator's responsibility to ensure that the top-level
`public.crt` does not disclose any information (i.e. hostnames)
that are not publicly visible. However, this was the case in the
past already.
- Any other `public.crt` - except for the top-level one - must not
contain any IP SAN. The reason for this restriction is that the
Manager cannot match a SNI to an IP b/c the SNI is the server host
name. The entire purpose of SNI is to indicate which host the client
tries to connect to when multiple hosts run on the same IP. So, a
client will not set the SNI to an IP.
If we would allow IP SANs in a lower-level `public.crt` a user would
expect that it is possible to connect to MinIO directly via IP address
and that the MinIO server would pick "the right" certificate. However,
the MinIO server cannot determine which certificate to serve, and
therefore always picks the "default" one. This may lead to all sorts
of confusing errors like:
"It works if I use `https:instance.minio.local` but not when I use
`https://10.0.2.1`.
These consequences/limitations should be pointed out / explained in our
docs in an appropriate way. However, the support for multiple
certificates should not have any impact on how deployment with a single
certificate function today.
Co-authored-by: Harshavardhana <harsha@minio.io>
- do not fail the healthcheck if heal status
was not obtained from one of the nodes,
if many nodes fail then report this as a
catastrophic error.
- add "x-minio-write-quorum" value to match
the write tolerance supported by server.
- admin info now states if a drive is healing
where madmin.Disk.Healing is set to true
and madmin.Disk.State is "ok"
Currently, cache purges are triggered as soon as the low watermark is exceeded.
To reduce IO this should only be done when reaching the high watermark.
This simplifies checks and reduces all calls for a GC to go through
`dcache.diskSpaceAvailable(size)`. While a comment claims that
`dcache.triggerGC <- struct{}{}` was non-blocking I don't see how
that was possible. Instead, we add a 1 size to the queue channel
and use channel semantics to avoid blocking when a GC has
already been requested.
`bytesToClear` now takes the high watermark into account to it will
not request any bytes to be cleared until that is reached.
This commit reduces the retry delay when retrying a request
to a KES server by:
- reducing the max. jitter delay from 3s to 1.5s
- skipping the random delay when there are more KES endpoints
available.
If there are more KES endpoints we can directly retry to the request
by sending it to the next endpoint - as pointed out by @krishnasrinivas
- delete-marker should be created on a suspended bucket as `null`
- delete-marker should delete any pre-existing `null` versioned
object and create an entry `null`
When checking parts we already do a stat for each part.
Since we have the on disk size check if it is at least what we expect.
When checking metadata check if metadata is 0 bytes.
The `getNonLoopBackIP` may grab an IP from an interface that
doesn't allow binding (on Windows), so this test consistently fails.
We exclude that specific error.
* readDirN: Check if file is directory
`syscall.FindNextFile` crashes if the handle is a file.
`errFileNotFound` matches 'unix' functionality: d19b434ffc/cmd/os-readdir_unix.go (L106)Fixes#10384
This commit addresses a maintenance / automation problem when MinIO-KES
is deployed on bare-metal. In orchestrated env. the orchestrator (K8S)
will make sure that `n` KES servers (IPs) are available via the same DNS
name. There it is sufficient to provide just one endpoint.
ListObjectsV1 requests are actually redirected to a specific node,
depending on the bucket name. The purpose of this behavior was
to optimize listing.
However, the current code sends a Bad Gateway error if the
target node is offline, which is a bad behavior because it means
that the list request will fail, although this is unnecessary since
we can still use the current node to list as well (the default behavior
without using proxying optimization)
Currently, you can see mint fails when there is one offline node, after
this PR, mint will always succeed.
We can reduce this further in the future, but this is a good
value to keep around. With the advent of continuous healing,
we can be assured that namespace will eventually be
consistent so we are okay to avoid the necessity to
a list across all drives on all sets.
Bonus Pop()'s in parallel seem to have the potential to
wait too on large drive setups and cause more slowness
instead of gaining any performance remove it for now.
Also, implement load balanced reply for local disks,
ensuring that local disks have an affinity for
- cleanupStaleMultipartUploads()
If DiskInfo calls failed the information returned was used anyway
resulting in no endpoint being set.
This would make the drive be attributed to the local system since
`disk.Endpoint == disk.DrivePath` in that case.
Instead, if the call fails record the endpoint and the error only.
bonus make sure to ignore objectNotFound, and versionNotFound
errors properly at all layers, since HealObjects() returns
objectNotFound error if the bucket or prefix is empty.
When crawling never use a disk we know is healing.
Most of the change involves keeping track of the original endpoint on xlStorage
and this also fixes DiskInfo.Endpoint never being populated.
Heal master will print `data-crawl: Disk "http://localhost:9001/data/mindev/data2/xl1" is
Healing, skipping` once on a cycle (no more often than every 5m).
We should only enforce quotas if no error has been returned.
firstErr is safe to access since all goroutines have exited at this point.
If `firstErr` hasn't been set by something else return the context error if cancelled.
Keep dataUpdateTracker while goroutine is starting.
This will ensure the object is updated one `start` returns
Tested with
```
λ go test -cpu=1,2,4,8 -test.run TestDataUpdateTracker -count=1000
PASS
ok github.com/minio/minio/cmd 8.913s
```
Fixes#10295
fresh drive setups when one of the drive is
a root drive, we should ignore such a root
drive and not proceed to format.
This PR handles this properly by marking
the disks which are root disk and they are
taken offline.
newDynamicTimeout should be allocated once, in-case
of temporary locks in config and IAM we should
have allocated timeout once before the `for loop`
This PR doesn't fix any issue as such, but provides
enough dynamism for the timeout as per expectation.
Based on our previous conversations I assume we should send the version
id when healing an object.
Maybe we should even list object versions and heal all?
In a non recursive mode, issuing a list request where prefix
is an existing object with a slash and delimiter is a slash will
return entries in the object directory (data dir IDs)
```
$ aws s3api --profile minioadmin --endpoint-url http://localhost:9000 \
list-objects-v2 --bucket testbucket --prefix code_of_conduct.md/ --delimiter '/'
{
"CommonPrefixes": [
{
"Prefix":
"code_of_conduct.md/ec750fe0-ea7e-4b87-bbec-1e32407e5e47/"
}
]
}
```
This commit adds a fast exit track in Walk() in this specific case.
use `/etc/hosts` instead of `/` to check for common
device id, if the device is same for `/etc/hosts`
and the --bind mount to detect root disks.
Bonus enhance healthcheck logging by adding maintenance
tags, for all messages.
adds a feature where we can fetch the MinIO
command-line remotely, this
is primarily meant to add some stateless
nature to the MinIO deployment in k8s
environments, MinIO operator would run a
webhook service endpoint
which can be used to fetch any environment
value in a generalized approach.
It is possible in situations when server was deployed
in asymmetric configuration in the past such as
```
minio server ~/fs{1...4}/disk{1...5}
```
Results in setDriveCount of 10 in older releases
but with fairly recent releases we have moved to
having server affinity which means that a set drive
count ascertained from above config will be now '4'
While the object layer make sure that we honor
`format.json` the storageClass configuration however
was by mistake was using the global value obtained
by heuristics. Which leads to prematurely using
lower parity without being requested by the an
administrator.
This PR fixes this behavior.
Bonus also fix a bug where we did not purge relevant
service accounts generated by rotating credentials
appropriately, service accounts should become invalid
as soon as its corresponding parent user becomes invalid.
Since service account themselves carry parent claim always
we would never reach this problem, as the access get
rejected at IAM policy layer.
when source and destination are same and versioning is enabled
on the destination bucket - we do not need to re-create the entire
object once again to optimize on space utilization.
Cases this PR is not supporting
- any pre-existing legacy object will not
be preserved in this manner, meaning a new
dataDir will be created.
- key-rotation and storage class changes
of course will never re-use the dataDir
conflicting files can exist on FS at
`.minio.sys/buckets/testbucket/policy.json/`, this is an
expected valid scenario for FS mode allow it to work,
i.e ignore and move forward
With reduced parity our write quorum should be same
as read quorum, but code was still assuming
```
readQuorum+1
```
In all situations which is not necessary.
Generalize replication target management so
that remote targets for a bucket can be
managed with ARNs. `mc admin bucket remote`
command will be used to manage targets.
Context timeout might race on each other when timeouts are lower
i.e when two lock attempts happened very quickly on the same resource
and the servers were yet trying to establish quorum.
This situation can lead to locks held which wouldn't be unlocked
and subsequent lock attempts would fail.
This would require a complete server restart. A potential of this
issue happening is when server is booting up and we are trying
to hold a 'transaction.lock' in quick bursts of timeout.
replace dummy buffer with nullReader{} instead,
to avoid large memory allocations in memory
constrainted environments. allows running
obd tests in such environments.
Currently, listing directories on HDFS incurs a per-entry remote Stat() call
penalty, the cost of which can really blow up on directories with many
entries (+1,000) especially when considered in addition to peripheral
calls (such as validation) and the fact that minio is an intermediary to the
client (whereas other clients listed below can query HDFS directly).
Because listing directories this way is expensive, the Golang HDFS library
provides the [`Client.Open()`] function which creates a [`FileReader`] that is
able to batch multiple calls together through the [`Readdir()`] function.
This is substantially more efficient for very large directories.
In one case we were witnessing about +20 seconds to list a directory with 1,500
entries, admittedly large, but the Java hdfs ls utility as well as the HDFS
library sample ls utility were much faster.
Hadoop HDFS DFS (4.02s):
λ ~/code/minio → use-readdir
» time hdfs dfs -ls /directory/with/1500/entries/
…
hdfs dfs -ls 5.81s user 0.49s system 156% cpu 4.020 total
Golang HDFS library (0.47s):
λ ~/code/hdfs → master
» time ./hdfs ls -lh /directory/with/1500/entries/
…
./hdfs ls -lh 0.13s user 0.14s system 56% cpu 0.478 total
mc and minio **without** optimization (16.96s):
λ ~/code/minio → master
» time mc ls myhdfs/directory/with/1500/entries/
…
./mc ls 0.22s user 0.29s system 3% cpu 16.968 total
mc and minio **with** optimization (0.40s):
λ ~/code/minio → use-readdir
» time mc ls myhdfs/directory/with/1500/entries/
…
./mc ls 0.13s user 0.28s system 102% cpu 0.403 total
[`Client.Open()`]: https://godoc.org/github.com/colinmarc/hdfs#Client.Open
[`FileReader`]: https://godoc.org/github.com/colinmarc/hdfs#FileReader
[`Readdir()`]: https://godoc.org/github.com/colinmarc/hdfs#FileReader.Readdir
If there are many listeners to bucket notifications or to the trace
subsystem, healing fails to work properly since it suspends itself when
the number of concurrent connections is above a certain threshold.
These connections are also continuous and not costly (*no disk access*),
it is okay to just ignore them in waitForLowHTTPReq().
this is to detect situations of corruption disk
format etc errors quickly and keep the disk online
in such scenarios for requests to fail appropriately.
Fixes two different types of problems
- continuation of the problem seen in FS #9992
as not fixed for erasure coded deployments,
reproduced this issue with spark and its fixed now
- another issue was leaking walk go-routines which
would lead to high memory usage and crash the system
this is simply because all the walks which were purged
at the top limit had leaking end walkers which would
consume memory endlessly.
closes#9966closes#10088
not all claims need to be present for
the JWT claim, let the policies not
exist and only apply which are present
when generating the credentials
once credentials are generated then
those policies should exist, otherwise
the request will fail.
- copyObject in-place decryption failed
due to incorrect verification of headers
- do not decode ETag when object is encrypted
with SSE-C, so that pre-conditions don't fail
prematurely.
This PR adds support for healing older
content i.e from 2yrs, 1yr. Also handles
other situations where our config was
not encrypted yet.
This PR also ensures that our Listing
is consistent and quorum friendly,
such that we don't list partial objects
In federated NAS gateway setups, multiple hosts in srvRecords
was picked at random which could mean that if one of the
host was down the request can indeed fail and if client
retries it would succeed. Instead allow server to figure
out the current online host quickly such that we can
exclude the host which is down.
At the max the attempt to look for a downed node is to
300 millisecond, if the node is taking longer to respond
than this value we simply ignore and move to the node,
total attempts are equal to number of srvRecords if no
server is online we simply fallback to last dialed host.
healing was not working properly when drives were
replaced, due to the error check in root disk
calculation this PR fixes this behavior
This PR also adds additional fix for missing
metadata entries from .minio.sys as part of
disk healing as well.
Added code to ignore and print more context
sensitive errors for better debugging.
This PR is continuation of fix in 7b14e9b660
Enforce bucket quotas when crawling has finished.
This ensures that we will not do quota enforcement on old data.
Additionally, delete less if we are closer to quota than we thought.
for users who don't have access to HDFS rootPath '/'
can optionally specify `minio gateway hdfs hdfs://namenode:8200/path`
for which they have access to, allowing all writes to be
performed at `/path`.
NOTE: once configured in this manner you need to make
sure command line is correctly specified, otherwise
your data might not be visible
closes#10011
this PR to allow legacy support for big-data
applications which run older Java versions
which do not support the secure ciphers
currently defaulted in MinIO. This option
allows optionally to turn them off such
that client and server can negotiate the
best ciphers themselves.
This env is purposefully not documented,
meant as a last resort when client
application cannot be changed easily.
- admin info node offline check is now quicker
- admin info now doesn't duplicate the code
across doing the same checks for disks
- rely on StorageInfo to return appropriate errors
instead of calling locally.
- diskID checks now return proper errors when
disk not found v/s format.json missing.
- add more disk states for more clarity on the
underlying disk errors.
while we handle all situations for writes and reads
on older format, what we didn't cater for properly
yet was delete where we only ended up deleting
just `xl.meta` - instead we should allow all the
deletes to go through for older format without
versioning enabled buckets.
CORS is notorious requires specific headers to be
handled appropriately in request and response,
using cors package as part of handlerFunc() for
options method lacks the necessary control this
package needs to add headers.
Without instantiating a new rest client we can
have a recursive error which can lead to
healthcheck returning always offline, this can
prematurely take the servers offline.
gorilla/mux broke their recent release 1.7.4 which we
upgraded to, we need the current workaround to ensure
that our regex matches appropriately.
An upstream PR is sent, we should remove the
workaround once we have a new release.
- reduce locker timeout for early transaction lock
for more eagerness to timeout
- reduce leader lock timeout to range from 30sec to 1minute
- add additional log message during bootstrap phase
Main issue is that `t.pool[params]` should be `t.pool[oldest]`.
We add a bit more safety features for the code.
* Make writes to the endTimerCh non-blocking in all cases
so multiple releases cannot lock up.
* Double check expectations.
* Shift down deletes with copy instead of truncating slice.
* Actually delete the oldest if we are above total limit.
* Actually delete the oldest found and not the current.
* Unexport the mutex so nobody from the outside can meddle with it.
This commit adds a new admin API for creating master keys.
An admin client can send a POST request to:
```
/minio/admin/v3/kms/key/create?key-id=<keyID>
```
The name / ID of the new key is specified as request
query parameter `key-id=<ID>`.
Creating new master keys requires KES - it does not work with
the native Vault KMS (deprecated) nor with a static master key
(deprecated).
Further, this commit removes the `UpdateKey` method from the `KMS`
interface. This method is not needed and not used anymore.
with the merge of https://github.com/etcd-io/etcd/pull/11823
etcd v3.5.0 will now have a properly imported versioned path
this fixes our pending migration to newer repo
Use a separate client for these calls that can take a long time.
Add request context to these so they are canceled when the client
disconnects instead except for ListObject which doesn't have any equivalent.
Healing an object which has multiple versions was not working because
the healing code forgot to consider errFileVersionNotFound error as a
use case that needs healing
Currently, lifecycle expiry is deleting all object versions which is not
correct, unless noncurrent versions field is specified.
Also, only delete the delete marker if it is the only version of the
given object.
- additionally upgrade to msgp@v1.1.2
- change StatModTime,StatSize fields as
simple Size/ModTime
- reduce 50000 entries per List batch to 10000
as client needs to wait too long to see the
first batch some times which is not desired
and it is worth we write the data as soon
as we have it.
object KMS is configured with auto-encryption,
there were issues when using docker registry -
this has been left unnoticed for a while.
This PR fixes an issue with compatibility.
Additionally also fix the continuation-token implementation
infinite loop issue which was missed as part of #9939
Also fix the heal token to be generated as a client
facing value instead of what is remembered by the
server, this allows for the server to be stateless
regarding the token's behavior.
When manual healing is triggered, one node in a cluster will
become the authority to heal. mc regularly sends new requests
to fetch the status of the ongoing healing process, but a load
balancer could land the healing request to a node that is not
doing the healing request.
This PR will redirect a request to the node based on the node
index found described as part of the client token. A similar
technique is also used to proxy ListObjectsV2 requests
by encoding this information in continuation-token
The S3 specification says that versions are ordered in the response of
list object versions.
mc snapshot needs this to know which version comes first especially when
two versions have the same exact last-modified field.
Ritesh H Shukla
Harshavardhana
Aditya Manthramurthy
Poorna Krishnamoorthy
Anis Elleuch
Klaus Post
Andreas Auernhammer
ebozduman
Shireesh Anjal
Poorna Krishnamoorthy
Rafael Bodill
cksac
Steven Reitsma
Bill Thorp
Krishna Srinivas
Kaloyan Raev
飞雪无情
Lenin Alevski
Praveen raj Mani
Nitish Tiwari
Gaige B Paulsen
Jorge Israel Peña
Daniel Valdivia
kannappanr
KevinSmile
Adam Brown
Benjamin Sodenkamp