After deploying to most of production satellites we want to enable
it by default. No issues found after few weeks of running.
Change-Id: I4d83c65edaa95b0e50eeab6ac5e2c6cbb7206c1e
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
We will remove segments loop soon so we need first to move
Segment definition to rangedloop package.
https://github.com/storj/storj/issues/5237
Change-Id: Ibe6aad316ffb7073cc4de166f1f17b87aac07363
Current observer used with ranged loop is using massive amount of
memory because each range is generating separate set of bloom filters.
Each bloom filter can be up to 2MB of memory. That's a lot.
This change is initial change to reduce used memory by sharing bloom
filters between ranges and just synchronize access to them. This
implementation is rather simple and even naive but maybe it will be
enough without doing something more complex.
https://github.com/storj/storj/issues/5803
Change-Id: Ie62d19276aa9023076b1c97f712b788bce963cbe
This is refactor/cleanup change before I will start working on adding
separate GC observer with optimized memory consumption.
https://github.com/storj/storj/issues/5803
Change-Id: I854cb3797802a32942c25f2765dbb72be88bacbd
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
implemented observer and partial, created new structures to keep mon
metrics remain in same way as in segment loop
Change-Id: I209c126096c84b94d4717332e56238266f6cd004
Add option AsOfSystemTime to segment provider to make it equivalent to
the old segment loop.
There's no comment on what it does because it's pretty complex and
makes no sense, but we can improve it later.
closes https://github.com/storj/storj/issues/5434
Change-Id: I8f09b03803e681e2fd41008c5dba67804b0f37a1
Change he bloomfilter generation process to prefix the objects with a date and update the LATEST object with the prefix. The sender will read the LATEST file to get the prefix to process.
Change-Id: Iae0d3c49015d57f391d87789fb799a7d774066bf
The current deployment strategy requires that the GC bloomfilter generation process executes only once and exits.
Change-Id: I952991f126596aa165d1f2e9fce6f8548c21bdba
Restore functionality where retain filters can be sent out to multiple
storage nodes simultaneously.
Fixes https://github.com/storj/team-metainfo/issues/121
Change-Id: I2bf86a166b09c6a277c1cb455cdca0165ce6b8af
Restore previously existing end-to-end garbage collection test using
the new separate services for bloom filter generation and storage node
communication.
Original tests can be found under:
https://github.com/storj/storj/blob/v1.63.1/satellite/gc/gc_test.go
Change-Id: I42d1ab0f9981dfe183140da4d08087f4a6cd9296
Pieces count in DB are stored as int64 and we would like to align bloom
filter processing with this type.
Change-Id: Iaec767e609a40d802077ae057520541805a7c44f
* service was running wrong RunOnce method
* after doing integration with GC sender we concluded that we don't
need special flag "gc-sender" to be uploaded as its safe to consume
partial results by GC senger. This part was removed.
* prefix format for moving data after error was unified with GC
sender
https: //github.com/storj/team-metainfo/issues/120
Change-Id: I204b696b9c2def4874ad1d17d0e84231cc98d583
Implement a new service to read retain filter from a bucket and
send them out to storagenodes.
This allows the retain filters to be generated by a separate command on
a backup of the database.
Paralellism (setting ConcurrentSends) and end-to-end garbage collection
tests will be restored in a subsequent commit.
Solves https://github.com/storj/team-metainfo/issues/121
Change-Id: Iaf8a33fbf6987676cc3cf74a18a8078916fe673d
Bloom filter CreationDate is used to avoid deleting pieces that
where not processed by GC. Every piece created after that timestamp
won't be deleted. Current GC process is taking CreationDate as a
beginning of bloom filter creation. This is appraoch allows to avoid
issues with inconsistent view on DB as currently we are using live DB
to create bloom filters.
With appraoch were we will be using DB snaphot with segment loop
we can get CreationDate from latest created segment in DB. Every piece
created after latest created segment won't be touched by GC on storage
node.
Updates https://github.com/storj/team-metainfo/issues/120
Change-Id: I6aaf64948ab7f60cfea62195689ad77c25ea772e
We would like to have separate process/command to collect bloom
filters from source different than production DBs. Such process will
use segment loop to build bloom filters for all storage nodes and
will send it to Storj bucket.
This change add main logic to new service. After collecting all bloom
filters with segment loop and piece tracker all filters are marshaled
and packed into zip files. Each zip contains up to "ZipBatchSize" bloom
filters and it's uploaded to specified in configuration bucket.
All uploaded objects have specified expiration time to not delete them
manually.
Updates https://github.com/storj/team-metainfo/issues/120
Change-Id: I2b6bc02a7dd7c3a639e75810fd013ae4afdc80a2
We would like to have separate process/command to collect bloom
filters from source different than production DBs. Such process will
use segment loop to build bloom filters for all storage nodes and
will send it to Storj bucket.
This change adds integration with testplanet which makes writing
unit tests possible.
Updates https://github.com/storj/team-metainfo/issues/120
Change-Id: I7b335c5dafa8cffe265c56b75d8c8f8567580893
We would like to have separate process/command to collect bloom
filters from source different than production DBs. Such process will
use segment loop to build bloom filters for all storage nodes and
will send it to Storj bucket. This this initial change to add such
service. Added service is joining segment loop and collects all
bloom filters.
Sending bloom filters to the bucket will be added as a subsequent
change.
Updates https://github.com/storj/team-metainfo/issues/120
Change-Id: I2551723605afa41bec84826b0c647cd1f61f3b14
We made optimization for segment loop observers to avoid
heavy monkit initialization on each call. It was applied to very
often executed methods. Unfortunately we used wrong monkit
method to track function times. Instead mon.Task we used
mon.Func().
https://github.com/spacemonkeygo/monkit#how-it-works
Change-Id: I9ca454dbd828c6b43ba09ca75c341991d2fd73a8
Recently we applied this optimization to metrics observer and time
used by its method dropped from 12m to 3m for us1 (220m segments).
It looks that it make sense to apply the same code to all observers.
Change-Id: I05898aaacbd9bcdf21babc7be9955da1db57bdf2
We can use PieceIDDeriver in all places where we are deriving id from
the same id multiple times. We have serveral such places: gc, segment
deletion, segment validation, order limit creation. Using it should
save some resources.
Change-Id: I24668d516c0f7cea4aec6470614067734149501d
Initial space used for pieces is calcualted, not retrieved
from storage nodes and at the end of test we are deleting
also copies that become ancestors to verify that all data
was removed from storage nodes.
Change-Id: I9804adb9fa488dc0094a67a6e258c144977e7f5d
We implemented server-side copy feature and we would like to
confirm that it is not affecting GC.
Fixes https://github.com/storj/storj/issues/4696
Change-Id: Id391f0badf5fce51f9910f0df732d477b07fa7ac
At some point we moved metabase package outside Metainfo
but we didn't do that for satellite structure. This change
refactors only tests.
When uplink will be adjusted we can remove old entries in
Metainfo struct.
Change-Id: I2b66ed29f539b0ec0f490cad42c72840e0351bcb
We used this to reduce initial load on the core to avoid OOM. However,
this is not a problem anymore with garbage collection running
separately.
Change-Id: Ifd62c822a74974bc21a5913199334469a4bc0130
Satellites set their configuration values to default values using
cfgstruct, however, it turns out our tests don't test these values
at all! Instead, they have a completely separate definition system
that is easy to forget about.
As is to be expected, these values have drifted, and it appears
in a few cases test planet is testing unreasonable values that we
won't see in production, or perhaps worse, features enabled in
production were missed and weren't enabled in testplanet.
This change makes it so all values are configured the same,
systematic way, so it's easy to see when test values are different
than dev values or release values, and it's less hard to forget
to enable features in testplanet.
In terms of reviewing, this change should be actually fairly
easy to review, considering private/testplanet/satellite.go keeps
the current config system and the new one and confirms that they
result in identical configurations, so you can be certain that
nothing was missed and the config is all correct.
You can also check the config lock to see what actual config
values changed.
Change-Id: I6715d0794887f577e21742afcf56fd2b9d12170e
errs.Class should not contain "error" in the name, since that causes a
lot of stutter in the error logs. As an example a log line could end up
looking like:
ERROR node stats service error: satellitedbs error: node stats database error: no rows
Whereas something like:
ERROR nodestats service: satellitedbs: nodestatsdb: no rows
Would contain all the necessary information without the stutter.
Change-Id: I7b7cb7e592ebab4bcfadc1eef11122584d2b20e0
Currently the loop handling is heavily related to the metabase rather
than metainfo.
metainfo over time has become related to the "public API" for accessing
the metabase data.
Currently updates monkit.lock, because monkit monitoring does not handle
ScopeNamed correctly. Needs a followup change to monitoring check.
Change-Id: Ie50519991d718dfb872ec9a0176a82e732c97584
metabase has become a central concept and it's more suitable for it to
be directly nested under satellite rather than being part of metainfo.
metainfo is going to be the "endpoint" logic for handling requests.
Change-Id: I53770d6761ac1e9a1283b5aa68f471b21e784198
Check that the bloom filter creation date is earlier than the
metainfo loop system time used for db scanning.
Change-Id: Ib0f47c124f5651deae0fd7e7996abcdcaac98fb4
