Commit Graph

141 Commits

Author SHA1 Message Date
Anis Elleuch
cf52691959
Save resync status in the backend using a last update timestamp (#15638)
Currently, there is a short time window where the code is allowed 
to save the status of a replication resync. Currently, the window is
`now.Sub(st.EndTime) <= resyncTimeInterval`. Also, any failure to 
write in the backend disks is not retried.

Refactor the code a little bit to rely on the last timestamp of a
successful write of the resync status of any given bucket in the 
backend disks.
2022-09-01 16:53:36 -07:00
Anis Elleuch
10e75116ef
Avoid replicating dirs in listing with replication enabled (#15641)
When replication is enabled in a particular bucket, the listing will send
objects to bucket replication, but it is also sending prefixes for non
recursive listing which is useless and shows a lot of error logs.

This commit will ignore prefixes.
2022-09-01 15:22:11 -07:00
Harshavardhana
433b6fa8fe
upgrade golang-lint to the latest (#15600) 2022-08-26 12:52:29 -07:00
Harshavardhana
edba7c987b
fix: objects matching prefixes should not leave delete markers (#15586)
This is needed to ensure that we do not leave prefixes where
version is suspended, instead we never leave versions on
these paths.
2022-08-24 13:46:29 -07:00
Poorna
4155c5b695
replication: improve MRF healing. (#15556)
This PR improves the replication failure healing by persisting
most recent failures to disk and re-queuing them until the replication
is successful.

While this does not eliminate the need for healing during a full scan, 
queuing MRF vastly improves the ETA to keeping replicated buckets 
in sync as it does not wait for the scanner visit to detect unreplicated 
object versions.
2022-08-22 16:53:06 -07:00
Harshavardhana
ae4ee95d25
change default lock retry interval to 50ms (#15560)
competing calls on the same object on versioned bucket
mutating calls on the same object may unexpected have
higher delays.

This can be reproduced with a replicated bucket
overwriting the same object writes, deletes repeatedly.

For longer locks like scanner keep the 1sec interval
2022-08-19 16:21:05 -07:00
Harshavardhana
e9055e9ef7
fix: walk() should cancel itself upon context cancellation (#15553)
This PR fixes possible leaks that may emanate from not
listening on context cancelation or timeouts.

```
goroutine 60957610 [chan send, 16 minutes]:
github.com/minio/minio/cmd.(*erasureServerPools).Walk.func1.1.1(...)
        github.com/minio/minio/cmd/erasure-server-pool.go:1724 +0x368
github.com/minio/minio/cmd.listPathRaw({0x4a9a740, 0xc0666dffc0},...
        github.com/minio/minio/cmd/metacache-set.go:1022 +0xfc4
github.com/minio/minio/cmd.(*erasureServerPools).Walk.func1.1()
        github.com/minio/minio/cmd/erasure-server-pool.go:1764 +0x528
created by github.com/minio/minio/cmd.(*erasureServerPools).Walk.func1
        github.com/minio/minio/cmd/erasure-server-pool.go:1697 +0x1b7
```
2022-08-18 17:49:08 -07:00
Poorna
21fe14201f
replication: centralize healthcheck for remote targets (#15516)
This PR moves health check from minio-go client to being
managed on the server.

Additionally integrating health check into site replication
2022-08-16 17:46:22 -07:00
Poorna
21bf5b4db7
replication: heal proactively upon access (#15501)
Queue failed/pending replication for healing during listing and GET/HEAD
API calls. This includes healing of existing objects that were never
replicated or those in the middle of a resync operation.

This PR also fixes a bug in ListObjectVersions where lifecycle filtering
should be done.
2022-08-09 15:00:24 -07:00
ebozduman
b57e7321e7
Replaces 'disk'=>'drive' visible to end user (#15464) 2022-08-04 16:10:08 -07:00
Poorna
5e0776e96a
replication: Include replica object versions for resync (#15427) 2022-07-28 13:43:02 -07:00
Harshavardhana
5e763b71dc
use logger.LogOnce to reduce printing disconnection logs (#15408)
fixes #15334

- re-use net/url parsed value for http.Request{}
- remove gosimple, structcheck and unusued due to https://github.com/golangci/golangci-lint/issues/2649
- unwrapErrs upto leafErr to ensure that we store exactly the correct errors
2022-07-27 09:44:59 -07:00
Poorna
cab8d3d568
feat: add API to return list of objects waiting to be replicated (#15091) 2022-07-21 11:05:44 -07:00
Poorna
b4f6901903
resync: Avoid concurrent access/write on map (#15286)
fixes a crash

```
fatal error: concurrent map iteration and map write
minio[19309]: goroutine 18640 [running]:
minio[19309]: runtime.throw({0x27a3399?, 0x1785?})
minio[19309]: runtime/panic.go:992 +0x71 fp=0xc0062f1c80 sp=0xc0062f1c50 pc=0x438671
minio[19309]: runtime.mapiternext(0xc0062f1e90?)
minio[19309]: runtime/map.go:871 +0x4eb fp=0xc0062f1cf0 sp=0xc0062f1c80 pc=0x41002b
minio[19309]: github.com/minio/minio/cmd.(*ReplicationPool).periodicResyncMetaSave(0xc0056c00c0, {0x4d06a48, 0xc0005b2480}, {0x4d22fc0, 0xc0015ea0
```
2022-07-13 16:29:10 -07:00
Harshavardhana
0a8b78cb84
fix: simplify passing auditLog eventType (#15278)
Rename Trigger -> Event to be a more appropriate
name for the audit event.

Bonus: fixes a bug in AddMRFWorker() it did not
cancel the waitgroup, leading to waitgroup leaks.
2022-07-12 10:43:32 -07:00
Harshavardhana
31c4fdbf79
fix: resyncing 'null' version on pre-existing content (#15043)
PR #15041 fixed replicating 'null' version however
due to a regression from #14994 caused the target
versions for these 'null' versioned objects to have
different 'versions', this may cause confusion with
bi-directional replication and cause double replication.

This PR fixes this properly by making sure we replicate
the correct versions on the objects.
2022-06-06 15:14:56 -07:00
Harshavardhana
48e367ff7d
reject resync start on misconfigured replication rules (#15041)
we expect resync to start on buckets with replication
rule ExistingObjects enabled, if not we reject such
calls.
2022-06-06 02:54:39 -07:00
Harshavardhana
52221db7ef
fix: for unexpected errors in reading versioning config panic (#14994)
We need to make sure if we cannot read bucket metadata
for some reason, and bucket metadata is not missing and
returning corrupted information we should panic such
handlers to disallow I/O to protect the overall state
on the system.

In-case of such corruption we have a mechanism now
to force recreate the metadata on the bucket, using
`x-minio-force-create` header with `PUT /bucket` API
call.

Additionally fix the versioning config updated state
to be set properly for the site replication healing
to trigger correctly.
2022-05-31 02:57:57 -07:00
Harshavardhana
f1abb92f0c
feat: Single drive XL implementation (#14970)
Main motivation is move towards a common backend format
for all different types of modes in MinIO, allowing for
a simpler code and predictable behavior across all features.

This PR also brings features such as versioning, replication,
transitioning to single drive setups.
2022-05-30 10:58:37 -07:00
Poorna
5c81d0d89a
site replication: heal missing/invalid replication config (#14979)
Validate remote target ARNs and heal any stale rules in
the replication config
2022-05-26 17:57:23 -07:00
Harshavardhana
f8650a3493
fetch bucket replication stats across peers in single call (#14956)
current implementation relied on recursively calling one bucket
at a time across all peers, this would be very slow and chatty
when there are 100's of buckets which would mean 100*peerCount
amount of network operations.

This PR attempts to reduce this entire call into `peerCount`
amount of network calls only. This functionality addresses also a
concern where the Prometheus metrics would significantly slow
down when one of the peers is offline.
2022-05-23 09:15:30 -07:00
Harshavardhana
6cfb1cb6fd
fix: timer usage across codebase (#14935)
it seems in some places we have been wrongly using the
timer.Reset() function, nicely exposed by an example
shared by @donatello https://go.dev/play/p/qoF71_D1oXD

this PR fixes all the usage comprehensively
2022-05-17 22:42:59 -07:00
Harshavardhana
62aa42cccf
avoid replication proxy on version excluded paths (#14878)
no need to attempt proxying objects that were
never replicated, but do have local `null`
versions on them.
2022-05-08 16:50:31 -07:00
Harshavardhana
5cffd3780a
fix: multiple fixes in prefix exclude implementation (#14877)
- do not need to restrict prefix exclusions that do not
  have `/` as suffix, relax this requirement as spark may
  have staging folders with other autogenerated characters
  , so we are better off doing full prefix March and skip. 

- multiple delete objects was incorrectly creating a
  null delete marker on a versioned bucket instead of
  creating a proper versioned delete marker.

- do not suspend paths on the excluded prefixes during
  delete operations to avoid creating `null` delete markers,
  honor suspension of versioning only at bucket level for
  delete markers.
2022-05-07 22:06:44 -07:00
Krishnan Parthasarathi
ad8e611098
feat: implement prefix-level versioning exclusion (#14828)
Spark/Hadoop workloads which use Hadoop MR 
Committer v1/v2 algorithm upload objects to a 
temporary prefix in a bucket. These objects are 
'renamed' to a different prefix on Job commit. 
Object storage admins are forced to configure 
separate ILM policies to expire these objects 
and their versions to reclaim space.

Our solution:

This can be avoided by simply marking objects 
under these prefixes to be excluded from versioning, 
as shown below. Consequently, these objects are 
excluded from replication, and don't require ILM 
policies to prune unnecessary versions.

-  MinIO Extension to Bucket Version Configuration
```xml
<VersioningConfiguration xmlns="http://s3.amazonaws.com/doc/2006-03-01/"> 
        <Status>Enabled</Status>
        <ExcludeFolders>true</ExcludeFolders>
        <ExcludedPrefixes>
          <Prefix>app1-jobs/*/_temporary/</Prefix>
        </ExcludedPrefixes>
        <ExcludedPrefixes>
          <Prefix>app2-jobs/*/__magic/</Prefix>
        </ExcludedPrefixes>

        <!-- .. up to 10 prefixes in all -->     
</VersioningConfiguration>
```
Note: `ExcludeFolders` excludes all folders in a bucket 
from versioning. This is required to prevent the parent 
folders from accumulating delete markers, especially
those which are shared across spark workloads 
spanning projects/teams.

- To enable version exclusion on a list of prefixes

```
mc version enable --excluded-prefixes "app1-jobs/*/_temporary/,app2-jobs/*/_magic," --exclude-prefix-marker myminio/test
```
2022-05-06 19:05:28 -07:00
Poorna
3a64580663
Add support for site replication healing (#14572)
heal bucket metadata and IAM entries for
sites participating in site replication from
the site with the most updated entry.

Co-authored-by: Harshavardhana <harsha@minio.io>
Co-authored-by: Aditya Manthramurthy <aditya@minio.io>
2022-04-24 02:36:31 -07:00
Krishnan Parthasarathi
7b81967a3c
Fix handling of object versions pending purge (#14555)
- GetObject() with vid should return 405
- GetObject() without vid should return 404
- ListObjects() should ignore this object if this is the "latest" version of the object
- ListObjectVersions() should list this object as "DELETE marker"
- Remove data parts before sync'ing the version pending purge
2022-03-16 16:59:43 -07:00
Poorna
1e39ca39c3
fix: consistent replies for incorrect range requests on replicated buckets (#14345)
Propagate error from replication proxy target correctly to the client if range GET is unsatisfiable.
2022-03-08 13:58:55 -08:00
Poorna
ed3418c046
Refactor replication resync to be an active process (#14266)
When resync is triggered, walk the bucket namespace and
resync objects that are unreplicated. This PR also adds
an API to report resync progress.
2022-02-10 10:16:52 -08:00
Poorna
288e276abe
Specify tags in options while selecting replication targets (#14126)
When the replication rule is based on tag matches, the replication process
should pick up targets matching the tags specified in the replication
rule.

Fixing regression due to #12880
2022-01-19 10:45:42 -08:00
Harshavardhana
cc3f139d1f
replication: attempt abort multipart-upload at max 3 times on remote (#14087)
this is mainly an attempt to relinquish space on the remote
site, if this still doesn't do it we give and let the admin
know with a log message.
2022-01-11 22:32:29 -08:00
Poorna
54a98773f8
fix: replication of tag removal (#14056)
Currently tag removal leaves replication state as `PENDING` 
because the `HEAD` api returns just a tag count but not the 
actual tags, and this is treated as a no-op
2022-01-10 19:06:10 -08:00
Harshavardhana
f527c708f2
run gofumpt cleanup across code-base (#14015) 2022-01-02 09:15:06 -08:00
Aditya Manthramurthy
997e808088
fix; race in bucket replication stats (#13942)
- r.ulock was not locked when r.UsageCache was being modified

Bonus:

- simplify code by removing some unnecessary clone methods - we can 
do this because go arrays are values (not pointers/references) that are 
automatically copied on assignment.

- remove some unnecessary map allocation calls
2021-12-17 15:33:13 -08:00
Poorna K
e270ab65b3
fix: healing of replication delete markers (#13933)
A corner case can occur where the delete-marker was propagated 
but the metadata could not be updated on the primary. Sending 
a RemoveObject call with the Delete marker version would end 
up permanently deleting the version on target. Instead, perform 
a Stat on the delete-marker version on target and redo replication 
only if the delete-marker is missing on target.
2021-12-16 15:34:55 -08:00
Poorna K
d422d24278
replication: warn if insufficient workers (#13899)
This should give an early warning if configured replication 
workers are insufficient to meet application workload.
2021-12-13 18:22:56 -08:00
Harshavardhana
914bfb2d9c
fix: allow compaction on replicated buckets (#13711)
currently getReplicationConfig() failure incorrectly
returns error on unexpected buckets upon upgrade, we
should always calculate usage as much as possible.
2021-11-19 14:46:14 -08:00
Klaus Post
faf013ec84
Improve performance on multiple versions (#13573)
Existing:

```go
type xlMetaV2 struct {
    Versions []xlMetaV2Version `json:"Versions" msg:"Versions"`
}
```

Serialized as regular MessagePack.

```go
//msgp:tuple xlMetaV2VersionHeader
type xlMetaV2VersionHeader struct {
	VersionID [16]byte
	ModTime   int64
	Type      VersionType
	Flags     xlFlags
}
```

Serialize as streaming MessagePack, format:

```
int(headerVersion)
int(xlmetaVersion)
int(nVersions)
for each version {
    binary blob, xlMetaV2VersionHeader, serialized
    binary blob, xlMetaV2Version, serialized.
}
```

xlMetaV2VersionHeader is <= 30 bytes serialized. Deserialized struct 
can easily be reused and does not contain pointers, so efficient as a 
slice (single allocation)

This allows quickly parsing everything as slices of bytes (no copy).

Versions are always *saved* sorted by modTime, newest *first*. 
No more need to sort on load.

* Allows checking if a version exists.
* Allows reading single version without unmarshal all.
* Allows reading latest version of type without unmarshal all.
* Allows reading latest version without unmarshal of all.
* Allows checking if the latest is deleteMarker by reading first entry.
* Allows adding/updating/deleting a version with only header deserialization.
* Reduces allocations on conversion to FileInfo(s).
2021-11-18 12:15:22 -08:00
Anis Elleuch
4caed7cc0d
metrics: Add replication latency metrics (#13515)
Add a new Prometheus metric for bucket replication latency

e.g.:
minio_bucket_replication_latency_ns{
    bucket="testbucket",
    operation="upload",
    range="LESS_THAN_1_MiB",
    server="127.0.0.1:9001",
    targetArn="arn:minio:replication::45da043c-14f5-4da4-9316-aba5f77bf730:testbucket"} 2.2015663e+07

Co-authored-by: Klaus Post <klauspost@gmail.com>
2021-11-17 12:10:57 -08:00
Harshavardhana
661b263e77
add gocritic/ruleguard checks back again, cleanup code. (#13665)
- remove some duplicated code
- reported a bug, separately fixed in #13664
- using strings.ReplaceAll() when needed
- using filepath.ToSlash() use when needed
- remove all non-Go style comments from the codebase

Co-authored-by: Aditya Manthramurthy <donatello@users.noreply.github.com>
2021-11-16 09:28:29 -08:00
Harshavardhana
4ed0eb7012
remove double reads updating object metadata (#13542)
Removes RLock/RUnlock for updating metadata,
since we already take a write lock to update
metadata, this change removes reading of xl.meta
as well as an additional lock, the performance gain
should increase 3x theoretically for

- PutObjectRetention
- PutObjectLegalHold

This optimization is mainly for Veeam like
workloads that require a certain level of iops
from these API calls, we were losing iops.
2021-10-30 08:22:04 -07:00
Poorna K
e7f559c582
Fixes to replication metrics (#13493)
For reporting ReplicaSize and loading initial
replication metrics correctly.
2021-10-21 18:52:55 -07:00
Poorna Krishnamoorthy
7f6ed35347
Allow null versions to be replicated (#13310)
for pre-existing objects present in a bucket
prior to enabling existing object replication.

Co-authored-by: Poorna Krishnamoorthy <poorna@minio.io>
2021-09-28 10:26:12 -07:00
Poorna Krishnamoorthy
19ecdc75a8
replication: Simplify metrics calculation (#13274)
Also doing some code cleanup
2021-09-22 10:48:45 -07:00
Poorna Krishnamoorthy
806b10b934
fix: improve error messages returned during replication setup (#13261) 2021-09-21 13:03:20 -07:00
Poorna Krishnamoorthy
c4373ef290
Add support for multi site replication (#12880) 2021-09-18 13:31:35 -07:00
Harshavardhana
0892f1e406
fix: multipart replication and encrypted etag for sse-s3 (#13171)
Replication was not working properly for encrypted
objects in single PUT object for preserving etag,

We need to make sure to preserve etag such that replication
works properly and not gets into infinite loops of copying
due to ETag mismatches.
2021-09-08 22:25:23 -07:00
Poorna Krishnamoorthy
9af4e7b1da
Add healthcheck back for replication targets (#13168)
This will allow objects to relinquish read lock held during
replication earlier if the target is known to be down
without waiting for connection timeout when replication 
is attempted.
2021-09-08 15:34:50 -07:00
Poorna Krishnamoorthy
a366143c5b
Remove replication permission check (#13135)
Fixes #13105
2021-09-02 09:31:13 -07:00
Poorna Krishnamoorthy
6a7e22386e
Use part sizes correctly in multipart replication (#13061)
fixes #13057
2021-08-24 14:41:05 -07:00