common/pb moved grpc to a separate package common/pb/pbgrpc.
This updates this repository to use it.
Change-Id: I2de2a190688871cf9cb61f7ea511f8a01e264e4e
Now that we are trying to identify the root cause of the satellite load limitations (i.e. currently the satellite has a max ability of 400 rps for uploads and we need this to be higher), we are using the golang diagnostic tools to collect insight into what the bottlenecks are. We currently have a debug endpoint to gather some cpu and mem data, but it could be useful to have continuous profiling. GCP stackdriver has support for continuous profiling so lets set that up and see if it is helpful to gather more data.
This PR adds support for [GCP continuous profiler](https://cloud.google.com/profiler) which allows enabling continuous cpu/mem profiling and the stats are sent to stackdriver in google cloud console.
To enable the continuous profiling for a storj component, do the following:
- prereq: the workload must be running in GKE and have Stackdriver Profiling IAM role permissions
- provide the config flag `debug.profilename` in the config.yaml file for the workload (i.e. satellite api process, etc). The profilename should be the workload name, for example "satellite-api".
- once the above config flag is provided, the profiler will be initialized and profiling stats will automatically be sent to GCP project where the workload is running and viewable in the Stackdriver Profile page in the console
The current implementation assumes the workload is running in GKE, however if we find if useful we can add support to enable this from anywhere. But for simplicity, its configured this way assuming the main goal is to enable in production systems.
Change-Id: Ibf8ebe2df7bf06fdd4951ee6a1e48854dd36ad47
With commit: 3331b443e7, satellite will
start calling `DeletePieces`. Therefore, we can remove the old endpoint
once the above commit is deployed with all satellites
Change-Id: I0124bc00a7cb808d119eb59f8fcd7fadf68158bb
we want to return back to the user as quick as possible but also keep
deleting remaining pieces on the storagenodes
Change-Id: I04e9e7a80b17a8c474c841cceae02bb21d2e796f
this commit updates our monkit dependency to the v3 version where
it outputs in an influx style. this makes discovery much easier
as many tools are built to look at it this way.
graphite and rothko will suffer some due to no longer being a tree
based on dots. hopefully time will exist to update rothko to
index based on the new metric format.
it adds an influx output for the statreceiver so that we can
write to influxdb v1 or v2 directly.
Change-Id: Iae9f9494a6d29cfbd1f932a5e71a891b490415ff
it was noticed that if you had a long lived transaction A that
was blocking some other transaction B and A was being aborted
due to retriable errors, then transaction B was never given
priority. this was due to using savepoints to do lightweight
retries.
this behavior was problematic becaue we had some queries blocked
for over 16 hours, so this commit addresses the issue with two
prongs:
1. bound the amount of time we will retry a transaction
2. create new transactions when a retry is needed
the first ensures that we never wait for 16 hours, and the value
chosen is 10 minutes. that should be long enough for an ample
amount of retries for small queries, and huge queries probably
shouldn't be retried, even if possible: it's more preferrable to
find a way to make them smaller.
the second ensures that even in the case of retries, queries that
are blocked on the aborted transaction gain priority to run.
between those two changes, the maximum stall time due to retries
should be bounded to around 10 minutes.
Change-Id: Icf898501ef505a89738820a3fae2580988f9f5f4
We move PathCipher to encryption.Store and we need to adjust
storj/uplink for those changes. Uplink repo is also using libuplink to
run tests so we need first adjust storj/storj libuplink and later
storj/uplink.
Change-Id: I84f23e6bad18ac139f72c19939dc526f9f46d88b
this is to help protect against intentional or unintentional
slowloris style problems where a client keeps a tcp connection
alive but never sends any data. because grpc is great, we have
to spawn a separate goroutine for every read/write to the stream
so that we can return from the server handler to cancel it if
necessary. yep. really.
additionally, we update the rpcstatus package to do some stack
trace capture and add a Wrap method for the times where we want
to just use the existing error.
also fixes a number of TODOs where we attach status codes to the
returned errors in the endpoints.
Change-Id: Id8bb8ff84aa34e0f711b0cf9bce3908b36a1d3c1
This reverts commit 8e242cd012.
Revert because lib/pq has known issues with context cancellation.
These issues need to be resolved before these changes can be merged.
Change-Id: I160af51dbc2d67c5449aafa406a403e5367bb555
this will allow for some nice runtime analysis down the road.
also, this allows for wrapping database handles in a way that
can interact with these contexts
requires https://review.dev.storj.io/c/storj/dbx/+/514
Change-Id: Ib087b7cd73296dd2c1e0331314da34d861f61d2b
Adds check to see if storage nodes are eligible to initiate
graceful exit, by checking their CreatedAt date and seeing if
their "age" is greater than the new config value:
NodeMinAgeInMonths
The default for this value is 6 months for now.
https://storjlabs.atlassian.net/browse/V3-3357
Change-Id: Ib807ab8987ddb5a38a27a83886490f73fe8c5816
these may not be optimal but they're probably better based on
our previous testing. we can tune better in the future now that
the groundwork is there.
Change-Id: Iafaee86d3181287c33eadf6b7eceb307dda566a6
* separate sadb migration, add version check
* update checkversion to do same validation as migration
* changes per CR
* add sa migration to storj-sim
* add different debug port in storj-sim for migration
* add wait for exit for storj-sim migration
* update sa docker entrypoint to support migration
* storj-sim satellite parts all wait for migration
* upgrade golang-migrate/migrate to v4 because bug
* fix go mod tidy
keep a pool of connections open when dialing for drpc. this
makes it so that long lived clients (like lib/uplink's Project)
don't continue to use a bad connection forever. it also allows
for concurrent rpcs.
Change-Id: If649b286050e4f09c413fadc3e1ce88f5fc6e600
all of the packages and tests work with both grpc and
drpc. we'll probably need to do some jenkins pipelines
to run the tests with drpc as well.
most of the changes are really due to a bit of cleanup
of the pkg/transport.Client api into an rpc.Dialer in
the spirit of a net.Dialer. now that we don't need
observers, we can pass around stateless configuration
to everything rather than stateful things that issue
observations. it also adds a DialAddressID for the
case where we don't have a pb.Node, but we do have an
address and want to assert some ID. this happened
pretty frequently, and now there's no more weird
contortions creating custom tls options, etc.
a lot of the other changes are being consistent/using
the abstractions in the rpc package to do rpc style
things like finding peer information, or checking
status codes.
Change-Id: Ief62875e21d80a21b3c56a5a37f45887679f9412
The fundamental problem is that both drpc and grpc servers
want to close the listener and they both want to ignore the
error from Accept after the listener is closed. There's no
way to do this in a race free way. Fortunately, the mux
hands out listeners that can be independently closed. That
means they can both do their own shutdown logic where they
ignore the error, and then after they're closed, the code
orchestrating the servers can close the listeners.
The final weird bit is that the server's Close method is
required to wait until the Run method has exited (or at
least enough for the listeners to definitely be closed)
because tests depend on that behavior, so we have to add
some channels/mutexes/onces to ensure that Run has exited
and that a new call can't start after Close is called.
Change-Id: I7c4ef293f7963f83138815f51824fd5b8d09ce15
What: Change cmd/uplink to use scopes
It moves the fields that will be subsumed by scopes into an explicit legacy section and hides their configuration flags.
Why: So that it can read scopes in from files and stuff
* update UI to reflect final mockups
* implement create handler and render offers table data to UI
* fix line-height unit and remove important from selectors
* update file names and ids for clarity
* shorten 'label' in ids
* localize global vars, fix endpoint names, remove unnecessary receiver, fix comments
* fix unnecessary initialization of pointer
* correct file-naming conventions
* register timeConverter in an init func for safety and remove unnecessary important from css
* consolidate create endpoints and add comments
* register timeConverter in init func
* add copyright to files
* introduce require pkg
* add proper http server unit test
* update linting and create offers concurrently in unit test
* fix getOffers comment
* add copy-right to unit-test
* fix data-races
* fix linting
* remove converter in NewServer
* requested changes in progress
* add require for checking status code
* renamed template file
* fix 400 handler
* fix missing copyright and remove extra line
* fix build
* run goroutine for testing parallel
* evaluate reqType with switch stmt and promp for credit amount in cents
* fix lint issue
* add default case
* remove unnecessary var
* fix range scope error
* remove empty lines and use long form for struct field
* fix merge conflicts
* fix template reference
* fix modal id
* not delete package
* add currency formatting and requested changes
* markup formatting
* lean out currency logic and move wait outside loop
* pass ToDollars func to home template
* fix lint
* update UI to reflect final mockups
* fix line-height unit and remove important from selectors
* update file names and ids for clarity
* shorten 'label' in ids
* correct file-naming conventions
* add copyright to files
* check if break in imports is causing lint error
* resolve lint
* tidy go mod
* fix shorthands
* Set up new port 8090 for in offers
Clean up commented code
Rename offers to offersweb
Remove unused code
Add todos for adding front-end templates
Add middleware for only allow local access
Add comment
Fix linting error
Remove commented code
Update storj-sim
Check request IP against Host IP
Use net pakcage to retrieve IP address
Rename service to marketing
* Add wrapper for all errors
* fix conflicts
* update the config file
* fix linting error
* remove unused packages
* remove global runtime var and add flag to storj-sim for mar static dir
* remove debugging lines
* add new config for test data and check if static dir flag is set before passing to mux
* change 'console' to 'marketing' for test data config
* fix linting errors
* update config flag
* Trigger Jenkins
* Trigger CLA
* change BindSetup to be an option to Bind
* add process.Bind to allow composite structures
* hack fix for noprefix flags
* used tagged version of structs
Before this PR, some flags were created by calling `cfgstruct.Bind` and having their fields create a flag. Once the flags were parsed, `viper` was used to acquire all the values from them and config files, and the fields in the struct were set through the flag interface.
This doesn't work for slices of things on config structs very well, since it can only set strings, and for a string slice, it turns out that the implementation in `pflag` appends an entry rather than setting it.
This changes three things:
1. Only have a `Bind` call instead of `Bind` and `BindSetup`, and make `BindSetup` an option instead.
2. Add a `process.Bind` call that takes in a `*cobra.Cmd`, binds the struct to the command's flags, and keeps track of that struct in a global map keyed by the command.
3. Use `viper` to get the values and load them into the bound configuration structs instead of using the flags to propagate the changes.
In this way, we can support whatever rich configuration we want in the config yaml files, while still getting command like flags when important.
this change removes the cryptopasta dependency.
a couple possible sources of problem with this change:
* the encoding used for ECDSA signatures on SignedMessage has changed.
the encoding employed by cryptopasta was workable, but not the same
as the encoding used for such signatures in the rest of the world
(most particularly, on ECDSA signatures in X.509 certificates). I
think we'll be best served by using one ECDSA signature encoding from
here on, but if we need to use the old encoding for backwards
compatibility with existing nodes, that can be arranged.
* since there's already a breaking change in SignedMessage, I changed
it to send and receive public keys in raw PKIX format, instead of
PEM. PEM just adds unhelpful overhead for this case.
* cmd/statreceiver: lua-scriptable stat receiver
Change-Id: I3ce0fe3f1ef4b1f4f27eed90bac0e91cfecf22d7
* some updates
Change-Id: I7c3485adcda1278fce01ae077b4761b3ddb9fb7a
* more comments
Change-Id: I0bb22993cd934c3d40fc1da80d07e49e686b80dd
* linter fixes
Change-Id: Ied014304ecb9aadcf00a6b66ad28f856a428d150
* catch errors
Change-Id: I6e1920f1fd941e66199b30bc427285c19769fc70
* review feedback
Change-Id: I9d4051851eab18970c5f5ddcf4ff265508e541d3
* errorgroup improvements
Change-Id: I4699dda3022f0485fbb50c9dafe692d3921734ff
* too tricky
the previous thing was better for memory with lots of errors at a time
but https://play.golang.org/p/RweTMRjoSCt is too much of a foot gun
Change-Id: I23f0b3d77dd4288fcc20b3756a7110359576bf44
* adds channel for getting node out of lookup
* WIP adding the channels to lookups
* WIP adding channel to node lookups
* Wires up FindNodes method with channels
* WIP adds a test suite for lookup - tests are still failing
* WIP wires up use of testplanet for kademlia lookup tests
* WIP merging in node id changes
* Merges in pkg/storj node type changes
* Tests passing
* Lookup node working via Inspector now
* updates
* WIP working on getting tests passing
* WIP getting tests passing
* FindNode works
* Linter fix
* Adds copyrights to lookup_test
* removes a fmt.Printf I missed
* Removes commented out lines
* preparing for use of `customtype` gogo extension with `NodeID` type
* review changes
* preparing for use of `customtype` gogo extension with `NodeID` type
* review changes
* wip
* tests passing
* wip fixing tests
* more wip test fixing
* remove NodeIDList from proto files
* linter fixes
* linter fixes
* linter/review fixes
* more freaking linter fixes
* omg just kill me - linterrrrrrrr
* travis linter, i will muder you and your family in your sleep
* goimports everything - burn in hell travis
* goimports update
* go mod tidy
* protobuf for sending bandwidth agreements to satellite from storage nodes
* Setup process for sending agreements
* Add payer_id to db with bandwidth agreements for better sorting
* Linter errors
* Read agreements from PSDB
* Try writing message to server
* Cleanup
* Basic functionality
* Better error handelling
* Fix test
* setup config and server structure for receiving bandwidth agreements
* Resolve linter issues
* Optional commit for if we want to handle deletes all at once
* add identity to Server, add logic for receiving bandwidth messsages
* Bandwidth agreement DBX creation and integration with bw agreement endpoint
Co-authored-by: Kishore <kishore@storj.io>
Co-authored-by: Cam <cameron@storj.io>
* protobuf for sending bandwidth agreements to satellite from storage nodes
* Setup process for sending agreements
* Add payer_id to db with bandwidth agreements for better sorting
* Linter errors
* Read agreements from PSDB
* Try writing message to server
* Cleanup
* Basic functionality
* Better error handelling
* Fix test
* setup config and server structure for receiving bandwidth agreements
* Resolve linter issues
* Optional commit for if we want to handle deletes all at once
* add identity to Server, add logic for receiving bandwidth messsages
* Bandwidth agreement DBX creation and integration with bw agreement endpoint
Co-authored-by: Kishore <kishore@storj.io>
Co-authored-by: Cam <cameron@storj.io>
* added postgres create/read/delete test function
Co-authored-by: kishore <kishore@storj.io
Co-authored-by: cam <cameron@storj.io>
* edit comment
* removed sqlite3 driver from dbx
* remove generated sqlite code, add dbx read limitoffset
* remove getServerAndDB function, rename getDBPath to getPSQLInfo
* WIP writing server endpoint test
* code review changes
..although it ought to work for other storage.KeyValueStore needs as
well. it's just optimized to work pretty well for a largish hierarchy of
paths.
This includes the addition of "long benchmarks" for KeyValueStore
testing. These will only be run when -test-bench-long is added to the
test flags. In these benchmarks, a large corpus of paths matching a
natural ("real-life") hierarchy is read from paths.data.gz (which you
can get from https://github.com/storj/path-test-corpus) and imported
into a particular KeyValueStore. Recursive and non-recursive queries are
run on it to detect performance problems that arise only at scale.
This also includes alternate implementation of the postgreskv client,
which works in a less-bizarre way for non-recursive queries, but suffers
from poor performance in tests such as the long benchmarks. Once this
alternate impl is committed to the tree, we can remove it again; I just
want it to be available for future reference.
* begin adding path encryption
* do not encrypt/decrypt first element of path (bucket)
* add path encryption for delete and list
* use encrypted paths in streamstore.Meta
* fix listing with encrypted paths
* move encrypt/decryptAfterBucket to streamstore
* fix listing with no prefix
* remove duplicate logic for listing with no prefix
* initial commit- wip, working on testing and library
* wip working on testing library functioon to get poointer
* working on nil reference for testing
* tests wip
* wip-working on getting tests to work
* working on tests
* put test passes
* working on test- need to export
* created pdclient, and now working on testing function
* tests working for list- getting object back
* wip - got derived piece id
* fixed making grpc public
* fixed linter errors and minor added method for size
* need to work on testing, added random integer function
* got psc server working for testing
* working on ranger test and ranger method
* testing creds for new computer
* working on getting segment metadata
* get random stripe
* added caveat to random fn
* fixed data types
* modified library to have one public function that returns a random stripe
* removed extra comments
* added commons.go file for audit
* added last path to be remembered
* changed random function to cryto/rand/ & worked on tests passing
* working on testing to get analysis of randomness
* changed to track last item in pagination
* finished testing randomness, cleaned up code
* fixed rebase errors
* removed error, kept common file
* fixed travis errors
* attempt to fix overlay issue
* fixed travis error
* updated pointer parameters
* made smaller functions, renamed audit
* made changes per suggestions
* removed gosum
* fixed pr per suggestions
* removed comment
* Loads cache from context for PointerDB access
* WIP adds overlay lookups to pointerdb requests
* Pointer lookup code is added for Get
* adds feature flag for pointerdb return
* refactors pointerdb code
* removes some unnecessary debug logs
* Fixes indent in config
* adds early return for non-remote pointers
* formats code, removes some comments
* Fixes tests broken by pointer proto changes
* adds error check and merges variable declaration
* removes commented out proto import
* adds error check to pdbclient
* begin adding encryption for remote pieces
* begin adding decryption
* add encryption key as arg to Put and Get
* move encryption/decryption to object store
* Add encryption key to object store constructor
* Add the erasure scheme to object store constructor
* Ensure decrypter is initialized with the stripe size used by encrypter
* Revert "Ensure decrypter is initialized with the stripe size used by encrypter"
This reverts commit 07272333f461606edfb43ad106cc152f37a3bd46.
* Revert "Add the erasure scheme to object store constructor"
This reverts commit ea5e793b536159d993b96e3db69a37c1656a193c.
* move encryption to stream store
* move decryption stuff to stream store
* revert changes in object store
* add encryptedBlockSize and close rangers on error during Get
* calculate padding sizes correctly
* encryptedBlockSize -> encryptionBlockSize
* pass encryption key and block size into stream store
* remove encryption key and block size from object store constructor
* move encrypter/decrypter initialization
* remove unnecessary cast
* Fix padding issue
* Fix linter
* add todos
* use random encryption key for data encryption. Store an encrypted copy of this key in segment metadata
* use different encryption key for each segment
* encrypt data in one step if it is small enough
* refactor and move encryption stuff
* fix errors related to nil slices passed to copy
* fix encrypter vs. decrypter bug
* put encryption stuff in eestream
* get captplanet test to pass
* fix linting errors
* add types for encryption keys/nonces and clean up
* fix tests
* more review changes
* add Cipher type for encryption stuff
* fix rs_test
* Simplify type casting of key and nonce
* Init starting nonce to the segment index
* don't copy derived key
* remove default encryption key; force user to explicitly set it
* move getSegmentPath to streams package
* dont require user to specify encryption key for captplanet
* rename GenericKey and GenericNonce to Key and Nonce
* review changes
* fix linting error
* Download uses the encryption type from metadata
* Store enc block size in metadata and use it for download
* storage node quick check and startup validation
* rearranged the startup validation and quick check logic
* travis lint warning fixes
* travis lint warning fixes
* travis lint warning fixes
* code changes per review comments
* code clean dev debug info
* travis lint wranings
* code changes per code review comments
* code changes per code review comments
* code update per review
* sqlite SUM is having issue when getting the SUM of an empty column; filepath was checking a directory that doesn't exist when starting server; Example updated to print allocated and used space
* storage node quick check and startup validation
* rearranged the startup validation and quick check logic
* travis lint warning fixes
* travis lint warning fixes
* travis lint warning fixes
* code changes per review comments
* code clean dev debug info
* travis lint wranings
* code changes per code review comments
* code changes per code review comments
* code update per review
* no file or directory error
* Updated mock PSClient
1. Added KeyValueStore.Iterate for implementing the different List, ListV2 etc. implementations. This allows for more efficient use of memory depending on the situation.
2. Implemented an inmemory teststore for running tests. This should allow to replace MockKeyValueStore in most places.
3. Rewrote tests
4. Pulled out logger from bolt implementation so it can be used for all other storage implementations.
5. Fixed multiple things in bolt and redis implementations.
* add CopyObject method
* use utils.LogClose
* extract common code from GetObject to getObject helper
* remove rr.Range from getObject helper, create helper putObject
* return rr, err in getObject helper
* extract code from PutObject into putObject helper
* remove commented out text
* remove other commented out code
* WIP trying to get storj cp command to work with copyObject
* fix typo in rb and now it works
* use rr.Size() instead of srcInfo.Size
* Revert "WIP trying to get storj cp command to work with copyObject"
This reverts commit e256b9f9a0fda728d41eb5b9d7a98b5446825842.
* add CopyObject test
* rebase and fix merge conflicts
* check error in gateway-storj test
* fix typo
* Don't use url.Parse for bolt paths: filepaths may not be valid URL-s.
* go.mod: update dependencies
* README.md: add Windows instructions
* pkg/overlay: check for the correct path and text in error
* pkg/overlay: fix tests for windows
* pkg/piecestore: make windows tests pass
* pkg/telemetry: skip test, as it doesn't shutdown nicely
* storage/redis: ensure that redis is clean before running tests
