While adding support for pending_objects table one case was missed.
When we are uploading object to location where old objects exists
we are not removing old object segments at all. Old object is
overwritten with new object metadata but segments remains without
corresponding object. This fix removes all existing committed objects
(with it's segments) before committing new object.
https://github.com/storj/storj/issues/6255
Change-Id: Id657840edf763fd6aec8191788d819191b074fb7
If MaxObjectTTL is set in the API key, BeginObject will use it for the
object expiration time, unless an explicit ExpireAt is available in the
request.
Context: https://github.com/storj/storj/issues/6249
Change-Id: I2adf57d979a9c68eec3a787f3739d2f1dbec1f7e
Some errors were returned as metabase errors, not pure drpc
errors because of how rpcstatus.Code method is working. Status
code was returned for errors like metabase context canceled but
we would like to not leak our internals to the client.
Change-Id: I3f0194755f8d7359b1e3d342fa3be3d984019ecb
This small feature will give us ability to test pending_objects table
without enabling it globally.
Change-Id: I802f45987ad329f94adfc0f02957c802b21d8251
table
New method IteratePendingObjectsByKeyNew is used to provide results for
metainfo.ListPendingObjectStreams. This endpoint is used to list
pending objects with the same object key. In this case to support
both tables (objects, pending_objects) we need to do one query per table
and merge results.
Because existing metainfo protobuf API is missing some fields to have
proper listing cursor we are not able to make ListPendingObjectStreams
correct for returning more than single page. We need to fix it
separately.
With this change also turns out that approach to merge results from
listing objects for ListObjects method was wrong and this change is also
fixing this problem.
Handling both tables will be removed at some point and only
pending_objects will be used to look for results.
Part of https://github.com/storj/storj/issues/6047
Change-Id: I8a88a6f885ad529704e6c032f1d97926123c2909
Adjust metainfo.ListObjects method to use IteratePendingObjects to
support new pending_objects table. New method will be used only when
we are listing pending objects.
Because until objects table will be free from pending objects we can
have results in both tables we are merging listing results. This also
means that in some (rare?) cases we may return more results than
specified listing limit. This situation is temporary.
Part of https://github.com/storj/storj/issues/6047
Change-Id: I06389145e5d916c532dfdbd3dcc9ef68ef70e515
In some rare cases when two entities are trying to create the same
bucket at the same time it's possible that we will return internal
error instead of `bucket already exists`. It's because we are not
handling correctly DB error about constraint error. This change checks
if while inserting bucket into DB we got constraint error and propagate
correct error to metainfo API.
Change-Id: Ie6fd2c943b864b4ea7d71e4a162e74dc3510e386
Allow a longer encrypted key length to reduce 'key length is too big'
errors in gateway-mt. Gateway is enforcing an unencrypted key length
of 1024 bytes but when encrypted some keys are exceeding the current
limit.
Updates https://github.com/storj/gateway-mt/issues/335
Change-Id: I38a0fbb0843fd782aeadca85f9a202821421b5a2
We are deleting pending objects while aborting multipart upload. We are
using metainfo BeginDeleteObject to do that. This change starts using
DeletePendingObjectNew to delete entry from pending_objects table when
request indicates that object is in this table.
Part of https://github.com/storj/storj/issues/6048
Change-Id: I4478a9c13c8e3db48dc5de3087ef03d1b4c47a5c
This patch fixes the node tag based placement of rangedloop/repairchecker + repair process.
The main change is just adding the node tags for Reliable and KnownReliabel database calls + adding new tests to prove, it works.
https://github.com/storj/storj/issues/6126
Change-Id: I245d654a18c1d61b2c72df49afa0718d0de76da1
We don't need to support segment copies with references anymore.
We migrated to copies where all metadata are copied from original
segment to copy.
https://github.com/storj/storj/issues/5891
Change-Id: Ic91dc21b0386ddf5c51aea45530024cd463e8ba9
With this change we are removing code responsible to handle deleting
bucket and supporting server side copies created with references. In practice we are restoring delete queries that we had before server side copy implementation (with small exception, see bellow).
From deletion queries we are also removing parts with segment metadata
as result because we are not longer sending explicit delete requests to
storage nodes.
https://github.com/storj/storj/issues/5891
Change-Id: If866d9f3a1b01e9ebd9b49c4740a6425ba06dd43
Change is adjusting CommitObject to use `pending_objects` table to
commit object.
Satellite stream id is used to determine if we need to use
`pending_objects` or `objects` table during commit.
General goal is to support both tables until `objects` table will be
free from pending objects.
Part of https://github.com/storj/storj/issues/6046
Change-Id: I2ebe0cd6b446727c98c8e210d4d00504dd0dacb6
Change is adjusting CommitSegment to check pending object existence in
`pending_objects` or `objects` table.
Satellite stream id is used to determine if we need to use
`pending_objects` or `objects` table.
General goal is to support both tables until `objects` table will be
free from pending objects. Whenever it will be needed code will be
supporting both tables at once.
Part of https://github.com/storj/storj/issues/6046
Change-Id: I954444a53b4733ae6fc909420573242b02746787
Change is adjusting BeginSegment to check pending object existence in
`pending_objects` or `objects` table.
Satellite stream id is used to determine if we need to use
`pending_objects` or `objects` table.
General goal is to support both tables until `objects` table will be
free from pending objects. Whenever it will be needed code will be
supporting both tables at once.
Part of https://github.com/storj/storj/issues/6046
Change-Id: I08aaa605c23d82695fde352fdbd0a7fd11f46bb5
Change is adjusting BeginObjectNextVersion to create pending object in
`pending_objects` or `objects` table depends on configuration. This is
first change to move pending objects from objects table.
General goal is to support both tables until `objects` table will be
free from pending objects. Whenever it will be needed code will be
supporting both tables at once.
To be able to decide if we need to use `pending_objects` table or
`objects` table we extend satellite stream id to keep that information
for later use.
BeginObjectExactVersion will be not adjusted because at the moment it's
used only in tests.
Part of https://github.com/storj/storj/issues/6046
Change-Id: Ibf21965f63cca5e1775469994a29f1fd1261af4e
We are doing full bucket name validation for many requests but
we should do this only while creating bucket. Other requests will be
covered only by basic name length validation. Less strict validation for
other requests will make bucket usable in case of invalid bucket names
in DB (we have such cases from the past).
https://github.com/storj/storj/issues/6044
Change-Id: I3a41050e3637787f788705ef15b5dc4df4d01fc6
Before this change, if a user creates a bucket with a user_agent attributed then deletes and recreates it, the row in bucket_metainfos
will not have the user_agent. This is because we skip setting the field
in bucket_metainfos if the bucket already exists in value_attributions.
This can be problematic, as we return the bucket's user agent during the
ListBuckets operation, and the client may be expecting this value to be
populated.
This change ensures the bucket table user_agent is set when (re)creating a bucket. To avoid decreasing BeginObject performance, which also
updates attribution, a flag has been added to determine whether to
make sure the buckets table is updated: `forceBucketUpdate`.
Change-Id: Iada2f233b327b292ad9f98c73ea76a1b0113c926
placement.AllowedCountry is the old way to specify placement, with the new approach we can use a more generic (dynamic method), which can check full node information instead of just the country code.
The 90% of this patch is just search and replace:
* we need to use NodeFilters instead of placement.AllowedCountry
* which means, we need an initialized PlacementRules available everywhere
* which means we need to configure the placement rules
The remaining 10% is the placement.go, where we introduced a new type of configuration (lightweight expression language) to define any kind of placement without code change.
Change-Id: Ie644b0b1840871b0e6bbcf80c6b50a947503d7df
..instead of using segment_copies and ancestor_stream_id, etc.
This bypasses reference counting entirely, depending on our mark+sweep+
bloomfilter garbage collection strategy to get rid of pieces once they
are no longer part of a segment.
This is only safe to do after we have stopped passing delete requests on
to storage nodes.
Refs: https://github.com/storj/storj/issues/5889
Change-Id: I37bdcffaa752f84fd85045235d6875b3526b5ecc
Allow a longer encrypted key length to reduce 'key length is too big'
errors in gateway-mt. Gateway is enforcing an unencrypted key length
of 1024 bytes but when encrypted some keys are exceeding the current
limit.
Updates https://github.com/storj/gateway-mt/issues/335
Change-Id: Ib02e2064c42e96b9d59936905832d8dd6068d2c7
We decided that we will stop sending explicit delete requests to nodes
and we will cleanup deleted with GC instead.
https://github.com/storj/storj/issues/5888
Change-Id: I65a308cca6fb17e97e3ba85eb3212584c96a32cd
* optimize SQL for zombie objects deletion query by reducing some direct
selects to segments table
* set AOST for expired/zombie object deletion (was 0 since now)
https://github.com/storj/storj/issues/5881
Change-Id: I50482151d056a86fe0e31678a463f413d410759d
For now we will use bucket placement to determine if we should exclude
some node IPs from metainfo.GetObjectIPs results. Bucket placement is
retrieved directly from DB in parallel to metabase
GetStreamPieceCountByNodeID request.
GetObjectIPs is not heavily used so additional request to DB shouldn't
be a problem for now.
https://github.com/storj/storj/issues/5950
Change-Id: Idf58b1cfbcd1afff5f23868ba2f71ce239f42439
this change adds code to CreateGetOrderLimits to filter
out any nodes that are not in the placement specified
by the segment. notably, it does not change the audit
or repair order limits. the list segments code had to be
changed to include getting the placement field from the
database.
Change-Id: Ice3e42a327811bb20928c619a72ed94e0c1464ac
Allow a longer encrypted key length to reduce 'key length is too big'
errors in gateway-mt. Gateway is enforcing an unencrypted key length
of 1024 bytes but when encrypted some keys are exceeding the current
limit.
updates https://github.com/storj/gateway-mt/issues/335
Change-Id: I861a2313e558c9f7d39569d21c7a3d429c83575c
This field is deprecated in favor of UserAgent; Removing these
references is the final step necessary before dropping the columns from
the database.
https: //github.com/storj/storj/issues/5432
Change-Id: I3a6619170dcf382f82dc8eddb73b6547eaf636f0
lrucache is now using time2 package and we can make expiration
test without using time.Sleep.
https://github.com/storj/storj/issues/5788
Change-Id: I48f2693c3db78fcf4e30e618bb3304be3625100c
We would like to have ability to limit burst uploads to the single
object (the same location). This change we are limiting such upload to
one per second.
Change-Id: Ib9351df1017cbc07d7fc2f846c2dbdbfcd3a360c
The blobstore implementation is entirely related to storagenode, so the
rightful place is together with the storagenode implementation.
Fixes https://github.com/storj/storj/issues/5754
Change-Id: Ie6637b0262cf37af6c3e558556c7604d9dc3613d
This combines the ListStreamPositions and GetSegmentByPosition
calls with a ListSegments call that now knows how to return
only the segments within a Range, just like ListStreamPositions.
It would theoretically be possible to also include the
GetObjectLastCommitted call by having it do one of three
queries based on the incoming request range, but that would
mean duplicating the data for the object in every single
row that is returned for each segment in the range.
One gross thing that ListSegments has to do now is update the
first segment returned with the information from any ancestor
segment because GetSegmentByPosition used to do that. It only
updates the first segment so that it doesn't do O(N) database
queries. It seems difficult to have it do a single query to
update all of the segments at once. I'm not certain this change
should be merged on this basis alone.
This change has made me think a couple of things should happen:
1. Server side copy with ancestor segments strikes again
making the code less clear and potentially more buggy
or inefficient for a rare case (empirically <0.1%)
2. The download code requests individual segments from
the satellite lazily as part of its download which
requires the satellite telling it the locations of
all of the segments which requires the satellite
querying the locations of all of the segments. Instead
the download RPC could return the orders for all of
the segments for a range and the download code could
issue N download calls rather than 1 download call and
N get segment calls. I believe both sides of the code
paths would be simpler and more efficient this way.
3. In looking at the timing information for downloads when
testing this, we really need to focus on getting the
auth key and bandwidth limit verification times down.
Here's the timing I saw:
- 42ms: validate auth
- 52ms: bandwidth usage checking
- 14ms: get object info
- 26ms: get segment position info
- 26ms: getting the first segment full info
- 20ms: unaccounted for by spans
- 6ms: creating the orders
This change will remove 26ms, but there's a good 90ms
in just validation. With improved semantics hitting the
database only once and improved validation, a download
rpc taking ~30ms seems doable compared to our current
~200ms.
Change-Id: I4109dba082eaedb79e634c61dbf86efa93ab1222
