data shards were wrong due to a healing bug
reported in #13803 mainly with unaligned object
sizes.
This PR is an attempt to automatically avoid
these shards, with available information about
the `xl.meta` and actually disk mtime.
FileInfo quorum shouldn't be passed down, instead
inferred after obtaining a maximally occurring FileInfo.
This PR also changes other functions that rely on
wrong quorum calculation.
Update tests as well to handle the proper requirement. All
these changes are needed when migrating from older deployments
where we used to set N/2 quorum for reads to EC:4 parity in
newer releases.
dataDir loosely based on maxima is incorrect and does not
work in all situations such as disks in the following order
- xl.json migration to xl.meta there may be partial xl.json's
leftover if some disks are not yet connected when the disk
is yet to come up, since xl.json mtime and xl.meta is
same the dataDir maxima doesn't work properly leading to
quorum issues.
- its also possible that XLV1 might be true among the disks
available, make sure to keep FileInfo based on common quorum
and skip unexpected disks with the older data format.
Also, this PR tests upgrade from older to a newer release if the
data is readable and matches the checksum.
NOTE: this is just initial work we can build on top of this to do further tests.
- add checks such that swapped disks are detected
and ignored - never used for normal operations.
- implement `unrecognizedDisk` to be ignored with
all operations returning `errDiskNotFound`.
- also add checks such that we do not load unexpected
disks while connecting automatically.
- additionally humanize the values when printing the errors.
Bonus: fixes handling of non-quorum situations in
getLatestFileInfo(), that does not work when 2 drives
are down, currently this function would return errors
incorrectly.
DeleteMarkers were unreadable if they had quorum based
guarantees, this PR tries to fix this behavior appropriately.
DeleteMarkers with sufficient should be allowed and the
return error should be accordingly with or without version-id.
This also allows for overwrites which may not be possible
in a multi-pool setup.
fixes#12787
Don't perform an independent evaluation of inlining, but mirror the decision made when uploading the object.
Leads to some objects being inlined or not based on new metrics. Instead respect previous decision.
we will allow situations such as
```
a/b/1.txt
a/b
```
and
```
a/b
a/b/1.txt
```
we are going to document that this usecase is
not supported and we will never support it, if
any application does this users have to delete
the top level parent to make sure namespace is
accessible at lower level.
rest of the situations where the prefixes get
created across sets are supported as is.
this healing optimization caused multiple
regressions in healing
- delete-markers incorrectly missing
heal and returning incorrect healing
results to client.
- missing individual 'parts' such
as for restored object or simply
for all objects just missing few parts.
This optimization is not necessary, we
should proceed to verify all cases possible
not just when metadata is inconsistent.
delete-markers missing on drives were
not healed due to few things
disksWithAllParts() does not know-how
to deal with delete markers, add support
for that.
fixes#12787
- delete-markers are incorrectly reported
as corrupt with wrong data sent to client
'mc admin heal -r' on objects with delete
marker will report as 'grey' incorrectly.
- do not heal delete-markers during HeadObject()
this can lead to inconsistent order of heals
on the object, although this is not an issue
in terms of order of versions it is rather
simpler to keep the same order on all drives.
- defaultHealResult() should handle 'err == nil'
case such that valid cases should be handled
as 'drive' status OK.
healing code was using incorrect buffers to heal older
objects with 10MiB erasure blockSize, incorrect calculation
of such buffers can lead to incorrect premature closure of
io.Pipe() during healing.
fixes#12410
This is to ensure that there are no projects
that try to import `minio/minio/pkg` into
their own repo. Any such common packages should
go to `https://github.com/minio/pkg`
A lot of healing is likely to be on non-existing objects and
locks are very expensive and will slow down scanning
significantly.
In cases where all are valid or, all are broken allow
rejection without locking.
Keep the existing behavior, but move the check for
dangling objects to after the lock has been acquired.
```
_, err = getLatestFileInfo(ctx, partsMetadata, errs)
if err != nil {
return er.purgeObjectDangling(ctx, bucket, object, versionID, partsMetadata, errs, []error{}, opts)
}
```
Revert "heal: Hold lock when reading xl.meta from disks (#12362)"
This reverts commit abd32065aa
This PR fixes two bugs
- Remove fi.Data upon overwrite of objects from inlined-data to non-inlined-data
- Workaround for an existing bug on disk with latest releases to ignore fi.Data
and instead read from the disk for non-inlined-data
- Addtionally add a reserved metadata header to indicate data is inlined for
a given version.
Lock is hold in healObject() after reading xl.meta from disks the first
time. This commit will held the lock since the beginning of HealObject()
Co-authored-by: Anis Elleuch <anis@min.io>
However, this slice is also used for closing the writers, so close is never called on these.
Furthermore when an error is returned from a write it is now reported to the reader.
bonus: remove unused heal param from `newBitrotWriter`.
* Remove copy, now that we don't mutate.
At some places bloom filter tracker was getting
updated for `.minio.sys/tmp` bucket, there is no
reason to update bloom filters for those.
And add a missing bloom filter update for MakeBucket()
Bonus: purge unused function deleteEmptyDir()
- GetObject() should always use a common dataDir to
read from when it starts reading, this allows the
code in erasure decoding to have sane expectations.
- Healing should always heal on the common dataDir, this
allows the code in dangling object detection to purge
dangling content.
These both situations can happen under certain types of
retries during PUT when server is restarting etc, some
namespace entries might be left over.
avoid time_wait build up with getObject requests if there are
pending callers and they timeout, can lead to time_wait states
Bonus share the same buffer pool with erasure healing logic,
additionally also fixes a race where parallel readers were
never cleanup during Encode() phase, because pipe.Reader end
was never closed().
Added closer right away upon an error during Encode to make
sure to avoid racy Close() while stream was still being
Read().
upon errors to acquire lock context would still leak,
since the cancel would never be called. since the lock
is never acquired - proactively clear it before returning.
* lock: Always cancel the returned Get(R)Lock context
There is a leak with cancel created inside the locking mechanism. The
cancel purpose was to cancel operations such erasure get/put that are
holding non-refreshable locks.
This PR will ensure the created context.Cancel is passed to the unlock
API so it will cleanup and avoid leaks.
* locks: Avoid returning nil cancel in local lockers
Since there is no Refresh mechanism in the local locking mechanism, we
do not generate a new context or cancel. Currently, a nil cancel
function is returned but this can cause a crash. Return a dummy function
instead.
https://github.com/minio/console takes over the functionality for the
future object browser development
Signed-off-by: Harshavardhana <harsha@minio.io>
Part ETags are not available after multipart finalizes, removing this
check as not useful.
Signed-off-by: Poorna Krishnamoorthy <poorna@minio.io>
Co-authored-by: Harshavardhana <harsha@minio.io>
With this change, MinIO's ILM supports transitioning objects to a remote tier.
This change includes support for Azure Blob Storage, AWS S3 compatible object
storage incl. MinIO and Google Cloud Storage as remote tier storage backends.
Some new additions include:
- Admin APIs remote tier configuration management
- Simple journal to track remote objects to be 'collected'
This is used by object API handlers which 'mutate' object versions by
overwriting/replacing content (Put/CopyObject) or removing the version
itself (e.g DeleteObjectVersion).
- Rework of previous ILM transition to fit the new model
In the new model, a storage class (a.k.a remote tier) is defined by the
'remote' object storage type (one of s3, azure, GCS), bucket name and a
prefix.
* Fixed bugs, review comments, and more unit-tests
- Leverage inline small object feature
- Migrate legacy objects to the latest object format before transitioning
- Fix restore to particular version if specified
- Extend SharedDataDirCount to handle transitioned and restored objects
- Restore-object should accept version-id for version-suspended bucket (#12091)
- Check if remote tier creds have sufficient permissions
- Bonus minor fixes to existing error messages
Co-authored-by: Poorna Krishnamoorthy <poorna@minio.io>
Co-authored-by: Krishna Srinivas <krishna@minio.io>
Signed-off-by: Harshavardhana <harsha@minio.io>
* fix: pick valid FileInfo additionally based on dataDir
historically we have always relied on modTime
to be consistent and same, we can now add additional
reference to look for the same dataDir value.
A dataDir is the same for an object at a given point in
time for a given version, let's say a `null` version
is overwritten in quorum we do not by mistake pick
up the fileInfo's incorrectly.
* make sure to not preserve fi.Data
Signed-off-by: Harshavardhana <harsha@minio.io>
This is an optimization by reducing one extra system call,
and many network operations. This reduction should increase
the performance for small file workloads.
current master breaks this important requirement
we need to preserve legacyXLv1 format, this is simply
ignored and overwritten causing a myriad of issues
by leaving stale files on the namespace etc.
for now lets still use the two-phase approach of
writing to `tmp` and then renaming the content to
the actual namespace.
upgrading from 2yr old releases is expected to work,
the issue was we were missing checksum info to be
passed down to newBitrotReader() for whole bitrot
calculation
This PR adds deadlines per Write() calls, such
that slow drives are timed-out appropriately and
the overall responsiveness for Writes() is always
up to a predefined threshold providing applications
sustained latency even if one of the drives is slow
to respond.
* Provide information on *actively* healing, buckets healed/queued, objects healed/failed.
* Add concurrent healing of multiple sets (typically on startup).
* Add bucket level resume, so restarts will only heal non-healed buckets.
* Print summary after healing a disk is done.
Delete marker can have `metaSys` set to nil, that
can lead to crashes after the delete marker has
been healed.
Additionally also fix isObjectDangling check
for transitioned objects, that do not have parts
should be treated similar to Delete marker.
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 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.
Harshavardhana
jiangfucheng
Anis Elleuch
Klaus Post
Krishnan Parthasarathi
Poorna Krishnamoorthy