pachyderm: github.com/pachyderm/pachyderm/src/client Index | Files | Directories

package client

import "github.com/pachyderm/pachyderm/src/client"

Index

Package Files

admin.go client.go debug.go health.go pfs.go pps.go version.go

Constants

const (
    // MaxListItemsLog specifies the maximum number of items we log in response to a List* API
    MaxListItemsLog = 10
    // StorageSecretName is the name of the Kubernetes secret in which
    // storage credentials are stored.
    StorageSecretName = "pachyderm-storage-secret"

    // DefaultPachdNodePort is the pachd kubernetes service's default
    // NodePort.Port setting.
    DefaultPachdNodePort = "30650"

    // DefaultPachdPort is the pachd kubernetes service's default
    // Port (often used with Pachyderm ELBs)
    DefaultPachdPort = "650"
)
const (
    // PPSEtcdPrefixEnv is the environment variable that specifies the etcd
    // prefix that PPS uses.
    PPSEtcdPrefixEnv = "PPS_ETCD_PREFIX"
    // PPSWorkerIPEnv is the environment variable that a worker can use to
    // see its own IP.  The IP address is made available through the
    // Kubernetes downward API.
    PPSWorkerIPEnv = "PPS_WORKER_IP"
    // PPSPodNameEnv is the environment variable that a pod can use to
    // see its own name.  The pod name is made available through the
    // Kubernetes downward API.
    PPSPodNameEnv = "PPS_POD_NAME"
    // PPSPipelineNameEnv is the env var that sets the name of the pipeline
    // that the workers are running.
    PPSPipelineNameEnv = "PPS_PIPELINE_NAME"
    // PPSNamespaceEnv is the namespace in which pachyderm is deployed
    PPSNamespaceEnv = "PPS_NAMESPACE"
    // PPSJobIDEnv is the env var that sets the ID of the job that the
    // workers are running (if the workers belong to an orphan job, rather than a
    // pipeline).
    PPSJobIDEnv = "PPS_JOB_ID"
    // PPSSpecCommitEnv is the namespace in which pachyderm is deployed
    PPSSpecCommitEnv = "PPS_SPEC_COMMIT"
    // PPSInputPrefix is the prefix of the path where datums are downloaded
    // to.  A datum of an input named `XXX` is downloaded to `/pfs/XXX/`.
    PPSInputPrefix = "/pfs"
    // PPSScratchSpace is where pps workers store data while it's waiting to be
    // processed.
    PPSScratchSpace = "/pfs/.scratch"
    // PPSWorkerPort is the port that workers use for their gRPC server
    PPSWorkerPort = 80
    // PPSWorkerVolume is the name of the volume in which workers store
    // data.
    PPSWorkerVolume = "pachyderm-worker"
    // PPSWorkerUserContainerName is the name of the container that runs
    // the user code to process data.
    PPSWorkerUserContainerName = "user"
    // PPSWorkerSidecarContainerName is the name of the sidecar container
    // that runs alongside of each worker container.
    PPSWorkerSidecarContainerName = "storage"
    // GCGenerationKey is the etcd key that stores a counter that the
    // GC utility increments when it runs, so as to invalidate all cache.
    GCGenerationKey = "gc-generation"
    // JobIDEnv is an env var that is added to the environment of user pipeline
    // code and indicates the id of the job currently being run.
    JobIDEnv = "PACH_JOB_ID"
    // OutputCommitIDEnv is an env var that is added to the environment of user
    // pipelined code and indicates the id of the output commit.
    OutputCommitIDEnv = "PACH_OUTPUT_COMMIT_ID"
)
const DefaultDialTimeout = 30 * time.Second

DefaultDialTimeout is the max amount of time APIClient.connect() will wait for a connection to be established unless overridden by WithDialTimeout()

const DefaultMaxConcurrentStreams = 100

DefaultMaxConcurrentStreams defines the max number of Putfiles or Getfiles happening simultaneously

func DatumTagPrefix Uses

func DatumTagPrefix(salt string) string

DatumTagPrefix hashes a pipeline salt to a string of a fixed size for use as the prefix for datum output trees. This prefix allows us to do garbage collection correctly.

func DefaultDialOptions Uses

func DefaultDialOptions() []grpc.DialOption

DefaultDialOptions is a helper returning a slice of grpc.Dial options such that grpc.Dial() is synchronous: the call doesn't return until the connection has been established and it's safe to send RPCs

func GetDatumTotalTime Uses

func GetDatumTotalTime(s *pps.ProcessStats) time.Duration

GetDatumTotalTime sums the timing stats from a DatumInfo

func NewAtomInput Uses

func NewAtomInput(repo string, glob string) *pps.Input

NewAtomInput returns a new atom input. It only includes required options.

func NewAtomInputOpts Uses

func NewAtomInputOpts(name string, repo string, branch string, glob string, lazy bool) *pps.Input

NewAtomInputOpts returns a new atom input. It includes all options.

func NewBlock Uses

func NewBlock(hash string) *pfs.Block

NewBlock creates a pfs.Block.

func NewBranch Uses

func NewBranch(repoName string, branchName string) *pfs.Branch

NewBranch creates a pfs.Branch

func NewCommit Uses

func NewCommit(repoName string, commitID string) *pfs.Commit

NewCommit creates a pfs.Commit.

func NewCronInput Uses

func NewCronInput(name string, spec string) *pps.Input

NewCronInput returns an input which will trigger based on a timed schedule. It uses cron syntax to specify the schedule. The input will be exposed to jobs as `/pfs/<name>/time` which will contain a timestamp.

func NewCrossInput Uses

func NewCrossInput(input ...*pps.Input) *pps.Input

NewCrossInput returns an input which is the cross product of other inputs. That means that all combination of datums will be seen by the job / pipeline.

func NewFile Uses

func NewFile(repoName string, commitID string, path string) *pfs.File

NewFile creates a pfs.File.

func NewJob Uses

func NewJob(jobID string) *pps.Job

NewJob creates a pps.Job.

func NewJobInput Uses

func NewJobInput(repoName string, commitID string, glob string) *pps.JobInput

NewJobInput creates a pps.JobInput.

func NewObject Uses

func NewObject(hash string) *pfs.Object

NewObject creates a pfs.Object.

func NewPipeline Uses

func NewPipeline(pipelineName string) *pps.Pipeline

NewPipeline creates a pps.Pipeline.

func NewPipelineInput Uses

func NewPipelineInput(repoName string, glob string) *pps.PipelineInput

NewPipelineInput creates a new pps.PipelineInput

func NewRepo Uses

func NewRepo(repoName string) *pfs.Repo

NewRepo creates a pfs.Repo.

func NewTag Uses

func NewTag(name string) *pfs.Tag

NewTag creates a pfs.Tag.

func NewUnionInput Uses

func NewUnionInput(input ...*pps.Input) *pps.Input

NewUnionInput returns an input which is the union of other inputs. That means that all datums from any of the inputs will be seen individually by the job / pipeline.

type APIClient Uses

type APIClient struct {
    PfsAPIClient
    PpsAPIClient
    ObjectAPIClient
    AuthAPIClient
    DeployAPIClient
    VersionAPIClient
    AdminAPIClient
    DebugClient
    Enterprise enterprise.APIClient // not embedded--method name conflicts with AuthAPIClient
    // contains filtered or unexported fields
}

An APIClient is a wrapper around pfs, pps and block APIClients.

func NewFromAddress Uses

func NewFromAddress(addr string, options ...Option) (*APIClient, error)

NewFromAddress constructs a new APIClient for the server at addr.

func NewInCluster Uses

func NewInCluster(options ...Option) (*APIClient, error)

NewInCluster constructs a new APIClient using env vars that Kubernetes creates. This should be used to access Pachyderm from within a Kubernetes cluster with Pachyderm running on it.

func NewOnUserMachine Uses

func NewOnUserMachine(reportMetrics bool, prefix string, options ...Option) (*APIClient, error)

NewOnUserMachine constructs a new APIClient using env vars that may be set on a user's machine (i.e. ADDRESS), as well as $HOME/.pachyderm/config if it exists. This is primarily intended to be used with the pachctl binary, but may also be useful in tests.

TODO(msteffen) this logic is fairly linux/unix specific, and makes the pachyderm client library incompatible with Windows. We may want to move this (and similar) logic into src/server and have it call a NewFromOptions() constructor.

func (*APIClient) AddMetadata Uses

func (c *APIClient) AddMetadata(ctx context.Context) context.Context

AddMetadata adds necessary metadata (including authentication credentials) to the context 'ctx', preserving any metadata that is present in either the incoming or outgoing metadata of 'ctx'.

func (APIClient) BlockCommit Uses

func (c APIClient) BlockCommit(repoName string, commitID string) (*pfs.CommitInfo, error)

BlockCommit returns info about a specific Commit, but blocks until that commit has been finished.

func (APIClient) BuildCommit Uses

func (c APIClient) BuildCommit(repoName string, branch string, parent string, treeObject string) (*pfs.Commit, error)

BuildCommit builds a commit in a single call from an existing HashTree that has already been written to the object store. Note this is a more advanced pattern for creating commits that's mostly used internally.

func (*APIClient) Close Uses

func (c *APIClient) Close() error

Close the connection to gRPC

func (APIClient) Compact Uses

func (c APIClient) Compact() error

Compact forces compaction of objects.

func (APIClient) CopyFile Uses

func (c APIClient) CopyFile(srcRepo, srcCommit, srcPath, dstRepo, dstCommit, dstPath string, overwrite bool) error

CopyFile copys a file from one pfs location to another. It can be used on directories or regular files.

func (APIClient) CreateBranch Uses

func (c APIClient) CreateBranch(repoName string, branch string, commit string, provenance []*pfs.Branch) error

CreateBranch creates a new branch

func (APIClient) CreateJob Uses

func (c APIClient) CreateJob(pipeline string, outputCommit *pfs.Commit) (*pps.Job, error)

CreateJob creates and runs a job in PPS. This function is mostly useful internally, users should generally run work by creating pipelines as well.

func (APIClient) CreatePipeline Uses

func (c APIClient) CreatePipeline(
    name string,
    image string,
    cmd []string,
    stdin []string,
    parallelismSpec *pps.ParallelismSpec,
    input *pps.Input,
    outputBranch string,
    update bool,
) error

CreatePipeline creates a new pipeline, pipelines are the main computation object in PPS they create a flow of data from a set of input Repos to an output Repo (which has the same name as the pipeline). Whenever new data is committed to one of the input repos the pipelines will create jobs to bring the output Repo up to data. image is the Docker image to run the jobs in. cmd is the command passed to the Docker run invocation. NOTE as with Docker cmd is not run inside a shell that means that things like wildcard globbing (*), pipes (|) and file redirects (> and >>) will not work. To get that behavior you should have your command be a shell of your choice and pass a shell script to stdin. stdin is a slice of lines that are sent to your command on stdin. Lines need not end in newline characters. parallelism is how many copies of your container should run in parallel. You may pass 0 for parallelism in which case PPS will set the parallelism based on available resources. input specifies a set of Repos that will be visible to the jobs during runtime. commits to these repos will cause the pipeline to create new jobs to process them. update indicates that you want to update an existing pipeline

func (APIClient) CreatePipelineService Uses

func (c APIClient) CreatePipelineService(
    name string,
    image string,
    cmd []string,
    stdin []string,
    parallelismSpec *pps.ParallelismSpec,
    input *pps.Input,
    update bool,
    internalPort int32,
    externalPort int32,
) error

CreatePipelineService creates a new pipeline service.

func (APIClient) CreateRepo Uses

func (c APIClient) CreateRepo(repoName string) error

CreateRepo creates a new Repo object in pfs with the given name. Repos are the top level data object in pfs and should be used to store data of a similar type. For example rather than having a single Repo for an entire project you might have separate Repos for logs, metrics, database dumps etc.

func (*APIClient) Ctx Uses

func (c *APIClient) Ctx() context.Context

Ctx is a convenience function that returns adds Pachyderm authn metadata to context.Background().

func (APIClient) DeleteAll Uses

func (c APIClient) DeleteAll() error

DeleteAll deletes everything in the cluster. Use with caution, there is no undo.

func (APIClient) DeleteBranch Uses

func (c APIClient) DeleteBranch(repoName string, branch string, force bool) error

DeleteBranch deletes a branch, but leaves the commits themselves intact. In other words, those commits can still be accessed via commit IDs and other branches they happen to be on.

func (APIClient) DeleteCommit Uses

func (c APIClient) DeleteCommit(repoName string, commitID string) error

DeleteCommit deletes a commit. Note it is currently not implemented.

func (APIClient) DeleteFile Uses

func (c APIClient) DeleteFile(repoName string, commitID string, path string) error

DeleteFile deletes a file from a Commit. DeleteFile leaves a tombstone in the Commit, assuming the file isn't written to later attempting to get the file from the finished commit will result in not found error. The file will of course remain intact in the Commit's parent.

func (APIClient) DeleteJob Uses

func (c APIClient) DeleteJob(jobID string) error

DeleteJob deletes a job.

func (APIClient) DeletePipeline Uses

func (c APIClient) DeletePipeline(name string, force bool) error

DeletePipeline deletes a pipeline along with its output Repo.

func (APIClient) DeleteRepo Uses

func (c APIClient) DeleteRepo(repoName string, force bool) error

DeleteRepo deletes a repo and reclaims the storage space it was using. Note that as of 1.0 we do not reclaim the blocks that the Repo was referencing, this is because they may also be referenced by other Repos and deleting them would make those Repos inaccessible. This will be resolved in later versions. If "force" is set to true, the repo will be removed regardless of errors. This argument should be used with care.

func (APIClient) DiffFile Uses

func (c APIClient) DiffFile(newRepoName, newCommitID, newPath, oldRepoName,
    oldCommitID, oldPath string, shallow bool) ([]*pfs.FileInfo, []*pfs.FileInfo, error)

DiffFile returns the difference between 2 paths, old path may be omitted in which case the parent of the new path will be used. DiffFile return 2 values (unless it returns an error) the first value is files present under new path, the second is files present under old path, files which are under both paths and have identical content are omitted.

func (APIClient) Dump Uses

func (c APIClient) Dump(w io.Writer) error

Dump returns debug information from the server.

func (APIClient) Extract Uses

func (c APIClient) Extract(objects bool, f func(op *admin.Op) error) error

Extract all cluster state, call f with each operation.

func (APIClient) ExtractAll Uses

func (c APIClient) ExtractAll(objects bool) ([]*admin.Op, error)

ExtractAll cluster state as a slice of operations.

func (APIClient) ExtractPipeline Uses

func (c APIClient) ExtractPipeline(pipelineName string) (*pps.CreatePipelineRequest, error)

ExtractPipeline extracts a single pipeline.

func (APIClient) ExtractURL Uses

func (c APIClient) ExtractURL(url string) error

ExtractURL extracts all cluster state and marshalls it to object storage.

func (APIClient) ExtractWriter Uses

func (c APIClient) ExtractWriter(objects bool, w io.Writer) error

ExtractWriter extracts all cluster state and marshals it to w.

func (APIClient) FinishCommit Uses

func (c APIClient) FinishCommit(repoName string, commitID string) error

FinishCommit ends the process of committing data to a Repo and persists the Commit. Once a Commit is finished the data becomes immutable and future attempts to write to it with PutFile will error.

func (APIClient) FlushCommit Uses

func (c APIClient) FlushCommit(commits []*pfs.Commit, toRepos []*pfs.Repo) (CommitInfoIterator, error)

FlushCommit returns an iterator that returns commits that have the specified `commits` as provenance. Note that the iterator can block if jobs have not successfully completed. This in effect waits for all of the jobs that are triggered by a set of commits to complete.

If toRepos is not nil then only the commits up to and including those repos will be considered, otherwise all repos are considered.

Note that it's never necessary to call FlushCommit to run jobs, they'll run no matter what, FlushCommit just allows you to wait for them to complete and see their output once they do.

func (APIClient) FlushCommitF Uses

func (c APIClient) FlushCommitF(commits []*pfs.Commit, toRepos []*pfs.Repo, f func(*pfs.CommitInfo) error) error

FlushCommitF calls f with commits that have the specified `commits` as provenance. Note that it can block if jobs have not successfully completed. This in effect waits for all of the jobs that are triggered by a set of commits to complete.

If toRepos is not nil then only the commits up to and including those repos will be considered, otherwise all repos are considered.

Note that it's never necessary to call FlushCommit to run jobs, they'll run no matter what, FlushCommit just allows you to wait for them to complete and see their output once they do.

func (APIClient) FlushJob Uses

func (c APIClient) FlushJob(commits []*pfs.Commit, toPipelines []string, f func(*pps.JobInfo) error) error

FlushJob calls f with all the jobs which were triggered by commits. If toPipelines is non-nil then only the jobs between commits and those pipelines in the DAG will be returned.

func (APIClient) FlushJobAll Uses

func (c APIClient) FlushJobAll(commits []*pfs.Commit, toPipelines []string) ([]*pps.JobInfo, error)

FlushJobAll returns all the jobs which were triggered by commits. If toPipelines is non-nil then only the jobs between commits and those pipelines in the DAG will be returned.

func (APIClient) GarbageCollect Uses

func (c APIClient) GarbageCollect(memoryBytes int64) error

GarbageCollect garbage collects unused data. Currently GC needs to be run while no data is being added or removed (which, among other things, implies that there shouldn't be jobs actively running). Pfs Garbage collection uses bloom filters to keep track of live objects because it can store more objects than can be indexed in memory. This means that there is a chance for unreferenced objects to not be GCed, this chance increases as the number of objects in the system increases. You can tradeoff using more memory to get a lower chance of collisions, the default value is 10 MB and collisions should be unlikely until you have 10 million objects.

func (*APIClient) GetAddress Uses

func (c *APIClient) GetAddress() string

GetAddress returns the pachd host:port with which 'c' is communicating. If 'c' was created using NewInCluster or NewOnUserMachine then this is how the address may be retrieved from the environment.

func (APIClient) GetFile Uses

func (c APIClient) GetFile(repoName string, commitID string, path string, offset int64, size int64, writer io.Writer) error

GetFile returns the contents of a file at a specific Commit. offset specifies a number of bytes that should be skipped in the beginning of the file. size limits the total amount of data returned, note you will get fewer bytes than size if you pass a value larger than the size of the file. If size is set to 0 then all of the data will be returned.

func (APIClient) GetFileReadSeeker Uses

func (c APIClient) GetFileReadSeeker(repoName string, commitID string, path string) (io.ReadSeeker, error)

GetFileReadSeeker returns a reader for the contents of a file at a specific Commit that permits Seeking to different points in the file.

func (APIClient) GetFileReader Uses

func (c APIClient) GetFileReader(repoName string, commitID string, path string, offset int64, size int64) (io.Reader, error)

GetFileReader returns a reader for the contents of a file at a specific Commit. offset specifies a number of bytes that should be skipped in the beginning of the file. size limits the total amount of data returned, note you will get fewer bytes than size if you pass a value larger than the size of the file. If size is set to 0 then all of the data will be returned.

func (APIClient) GetLogs Uses

func (c APIClient) GetLogs(
    pipelineName string,
    jobID string,
    data []string,
    datumID string,
    master bool,
    follow bool,
    tail int64,
) *LogsIter

GetLogs gets logs from a job (logs includes stdout and stderr). 'pipelineName', 'jobID', 'data', and 'datumID', are all filters. To forego any filter, simply pass an empty value, though one of 'pipelineName' and 'jobID' must be set. Responses are written to 'messages'

func (APIClient) GetObject Uses

func (c APIClient) GetObject(hash string, writer io.Writer) error

GetObject gets an object out of the object store by hash.

func (APIClient) GetObjectReader Uses

func (c APIClient) GetObjectReader(hash string) (io.ReadCloser, error)

GetObjectReader returns a reader for an object in object store by hash.

func (APIClient) GetObjects Uses

func (c APIClient) GetObjects(hashes []string, offset uint64, size uint64, totalSize uint64, writer io.Writer) error

GetObjects gets several objects out of the object store by hash.

func (APIClient) GetTag Uses

func (c APIClient) GetTag(tag string, writer io.Writer) error

GetTag gets an object out of the object store by tag.

func (APIClient) GetTagReader Uses

func (c APIClient) GetTagReader(tag string) (io.ReadCloser, error)

GetTagReader returns a reader for an object in object store by tag.

func (APIClient) GlobFile Uses

func (c APIClient) GlobFile(repoName string, commitID string, pattern string) ([]*pfs.FileInfo, error)

GlobFile returns files that match a given glob pattern in a given commit. The pattern is documented here: https://golang.org/pkg/path/filepath/#Match

func (APIClient) Health Uses

func (c APIClient) Health() error

Health health checks pachd, it returns an error if pachd isn't healthy.

func (APIClient) InspectBranch Uses

func (c APIClient) InspectBranch(repoName string, branch string) (*pfs.BranchInfo, error)

InspectBranch returns information on a specific PFS branch

func (APIClient) InspectCluster Uses

func (c APIClient) InspectCluster() (*admin.ClusterInfo, error)

InspectCluster retrieves cluster state

func (APIClient) InspectCommit Uses

func (c APIClient) InspectCommit(repoName string, commitID string) (*pfs.CommitInfo, error)

InspectCommit returns info about a specific Commit.

func (APIClient) InspectDatum Uses

func (c APIClient) InspectDatum(jobID string, datumID string) (*pps.DatumInfo, error)

InspectDatum returns info about a single datum

func (APIClient) InspectFile Uses

func (c APIClient) InspectFile(repoName string, commitID string, path string) (*pfs.FileInfo, error)

InspectFile returns info about a specific file.

func (APIClient) InspectJob Uses

func (c APIClient) InspectJob(jobID string, blockState bool) (*pps.JobInfo, error)

InspectJob returns info about a specific job. blockState will cause the call to block until the job reaches a terminal state (failure or success).

func (APIClient) InspectJobOutputCommit Uses

func (c APIClient) InspectJobOutputCommit(repoName, commitID string, blockState bool) (*pps.JobInfo, error)

InspectJobOutputCommit returns info about a job that created a commit. blockState will cause the call to block until the job reaches a terminal state (failure or success).

func (APIClient) InspectObject Uses

func (c APIClient) InspectObject(hash string) (*pfs.ObjectInfo, error)

InspectObject returns info about an Object.

func (APIClient) InspectPipeline Uses

func (c APIClient) InspectPipeline(pipelineName string) (*pps.PipelineInfo, error)

InspectPipeline returns info about a specific pipeline.

func (APIClient) InspectRepo Uses

func (c APIClient) InspectRepo(repoName string) (*pfs.RepoInfo, error)

InspectRepo returns info about a specific Repo.

func (APIClient) ListBranch Uses

func (c APIClient) ListBranch(repoName string) ([]*pfs.BranchInfo, error)

ListBranch lists the active branches on a Repo.

func (APIClient) ListCommit Uses

func (c APIClient) ListCommit(repoName string, to string, from string, number uint64) ([]*pfs.CommitInfo, error)

ListCommit lists commits. If only `repo` is given, all commits in the repo are returned. If `to` is given, only the ancestors of `to`, including `to` itself, are considered. If `from` is given, only the descendents of `from`, including `from` itself, are considered. `number` determines how many commits are returned. If `number` is 0, all commits that match the aforementioned criteria are returned.

func (APIClient) ListCommitByRepo Uses

func (c APIClient) ListCommitByRepo(repoName string) ([]*pfs.CommitInfo, error)

ListCommitByRepo lists all commits in a repo.

func (APIClient) ListCommitF Uses

func (c APIClient) ListCommitF(repoName string, to string, from string, number uint64, f func(*pfs.CommitInfo) error) error

ListCommitF lists commits, calling f with each commit. If only `repo` is given, all commits in the repo are returned. If `to` is given, only the ancestors of `to`, including `to` itself, are considered. If `from` is given, only the descendents of `from`, including `from` itself, are considered. `number` determines how many commits are returned. If `number` is 0, all commits that match the aforementioned criteria are returned.

func (APIClient) ListDatum Uses

func (c APIClient) ListDatum(jobID string, pageSize int64, page int64) (*pps.ListDatumResponse, error)

ListDatum returns info about all datums in a Job

func (APIClient) ListDatumF Uses

func (c APIClient) ListDatumF(jobID string, pageSize int64, page int64, f func(di *pps.DatumInfo) error) error

ListDatumF returns info about all datums in a Job, calling f with each datum info.

func (APIClient) ListFile Uses

func (c APIClient) ListFile(repoName string, commitID string, path string) ([]*pfs.FileInfo, error)

ListFile returns info about all files in a Commit under path.

func (APIClient) ListFileF Uses

func (c APIClient) ListFileF(repoName string, commitID string, path string, f func(fi *pfs.FileInfo) error) error

ListFileF returns info about all files in a Commit under path, calling f with each FileInfo.

func (APIClient) ListJob Uses

func (c APIClient) ListJob(pipelineName string, inputCommit []*pfs.Commit, outputCommit *pfs.Commit) ([]*pps.JobInfo, error)

ListJob returns info about all jobs. If pipelineName is non empty then only jobs that were started by the named pipeline will be returned If inputCommit is non-nil then only jobs which took the specific commits as inputs will be returned. The order of the inputCommits doesn't matter. If outputCommit is non-nil then only the job which created that commit as output will be returned.

func (APIClient) ListJobF Uses

func (c APIClient) ListJobF(pipelineName string, inputCommit []*pfs.Commit, outputCommit *pfs.Commit, f func(*pps.JobInfo) error) error

ListJobF returns info about all jobs, calling f with each JobInfo. If f returns an error iteration of jobs will stop and ListJobF will return that error, unless the error is errutil.ErrBreak in which case it will return nil. If pipelineName is non empty then only jobs that were started by the named pipeline will be returned If inputCommit is non-nil then only jobs which took the specific commits as inputs will be returned. The order of the inputCommits doesn't matter. If outputCommit is non-nil then only the job which created that commit as output will be returned.

func (APIClient) ListObject Uses

func (c APIClient) ListObject(f func(*pfs.Object) error) error

ListObject lists objects stored in pfs.

func (APIClient) ListPipeline Uses

func (c APIClient) ListPipeline() ([]*pps.PipelineInfo, error)

ListPipeline returns info about all pipelines.

func (APIClient) ListRepo Uses

func (c APIClient) ListRepo() ([]*pfs.RepoInfo, error)

ListRepo returns info about all Repos. provenance specifies a set of provenance repos, only repos which have ALL of the specified repos as provenance will be returned unless provenance is nil in which case it is ignored.

func (APIClient) ListTag Uses

func (c APIClient) ListTag(f func(*pfs.ListTagsResponse) error) error

ListTag lists tags stored in pfs.

func (APIClient) NewPutFileClient Uses

func (c APIClient) NewPutFileClient() (PutFileClient, error)

NewPutFileClient returns a new client for putting files into pfs in a single request.

func (APIClient) PutFile Uses

func (c APIClient) PutFile(repoName string, commitID string, path string, reader io.Reader) (_ int, retErr error)

PutFile writes a file to PFS from a reader.

func (APIClient) PutFileOverwrite Uses

func (c APIClient) PutFileOverwrite(repoName string, commitID string, path string, reader io.Reader, overwriteIndex int64) (_ int, retErr error)

PutFileOverwrite is like PutFile but it overwrites the file rather than appending to it. overwriteIndex allows you to specify the index of the object starting from which you'd like to overwrite. If you want to overwrite the entire file, specify an index of 0.

func (APIClient) PutFileSplit Uses

func (c APIClient) PutFileSplit(repoName string, commitID string, path string, delimiter pfs.Delimiter, targetFileDatums int64, targetFileBytes int64, headerRecords int64, overwrite bool, reader io.Reader) (_ int, retErr error)

PutFileSplit writes a file to PFS from a reader delimiter is used to tell PFS how to break the input into blocks

func (APIClient) PutFileSplitWriter Uses

func (c APIClient) PutFileSplitWriter(repoName string, commitID string, path string,
    delimiter pfs.Delimiter, targetFileDatums int64, targetFileBytes int64, headerRecords int64, overwrite bool) (io.WriteCloser, error)

PutFileSplitWriter writes a multiple files to PFS by splitting up the data that is written to it. NOTE: PutFileSplitWriter returns an io.WriteCloser you must call Close on it when you are done writing.

func (APIClient) PutFileURL Uses

func (c APIClient) PutFileURL(repoName string, commitID string, path string, url string, recursive bool, overwrite bool) (retErr error)

PutFileURL puts a file using the content found at a URL. The URL is sent to the server which performs the request. recursive allow for recursive scraping of some types URLs for example on s3:// urls.

func (APIClient) PutFileWriter Uses

func (c APIClient) PutFileWriter(repoName string, commitID string, path string) (io.WriteCloser, error)

PutFileWriter writes a file to PFS. NOTE: PutFileWriter returns an io.WriteCloser you must call Close on it when you are done writing.

func (APIClient) PutObject Uses

func (c APIClient) PutObject(_r io.Reader, tags ...string) (object *pfs.Object, _ int64, retErr error)

PutObject puts a value into the object store and tags it with 0 or more tags.

func (APIClient) PutObjectAsync Uses

func (c APIClient) PutObjectAsync(tags []*pfs.Tag) (*PutObjectWriteCloserAsync, error)

PutObjectAsync puts a value into the object store asynchronously.

func (APIClient) PutObjectSplit Uses

func (c APIClient) PutObjectSplit(_r io.Reader) (objects []*pfs.Object, _ int64, retErr error)

PutObjectSplit is the same as PutObject except that the data is splitted into several smaller objects. This is primarily useful if you'd like to be able to resume upload.

func (APIClient) ReadObject Uses

func (c APIClient) ReadObject(hash string) ([]byte, error)

ReadObject gets an object by hash and returns it directly as []byte.

func (APIClient) ReadObjects Uses

func (c APIClient) ReadObjects(hashes []string, offset uint64, size uint64) ([]byte, error)

ReadObjects gets several objects by hash and returns them directly as []byte.

func (APIClient) ReadTag Uses

func (c APIClient) ReadTag(tag string) ([]byte, error)

ReadTag gets an object by tag and returns it directly as []byte.

func (APIClient) RerunPipeline Uses

func (c APIClient) RerunPipeline(name string, include []*pfs.Commit, exclude []*pfs.Commit) error

RerunPipeline reruns a pipeline over a given set of commits. Exclude and include are filters that either include or exclude the ancestors of the given commits. A commit is considered the ancestor of itself. The behavior is the same as that of ListCommit.

func (APIClient) RestartDatum Uses

func (c APIClient) RestartDatum(jobID string, datumFilter []string) error

RestartDatum restarts a datum that's being processed as part of a job. datumFilter is a slice of strings which are matched against either the Path or Hash of the datum, the order of the strings in datumFilter is irrelevant.

func (APIClient) Restore Uses

func (c APIClient) Restore(ops []*admin.Op) (retErr error)

Restore cluster state from an extract series of operations.

func (APIClient) RestoreFrom Uses

func (c APIClient) RestoreFrom(objects bool, otherC *APIClient) (retErr error)

RestoreFrom restores state from another cluster which can be access through otherC.

func (APIClient) RestoreReader Uses

func (c APIClient) RestoreReader(r io.Reader) (retErr error)

RestoreReader restores cluster state from a reader containing marshaled ops. Such as those written by ExtractWriter.

func (APIClient) RestoreURL Uses

func (c APIClient) RestoreURL(url string) (retErr error)

RestoreURL restures cluster state from object storage.

func (*APIClient) SetAuthToken Uses

func (c *APIClient) SetAuthToken(token string)

SetAuthToken sets the authentication token that will be used for all API calls for this client.

func (APIClient) SetBranch Uses

func (c APIClient) SetBranch(repoName string, commit string, branch string) error

SetBranch sets a commit and its ancestors as a branch. SetBranch is deprecated in favor of CommitBranch.

func (APIClient) SetMaxConcurrentStreams Uses

func (c APIClient) SetMaxConcurrentStreams(n int)

SetMaxConcurrentStreams Sets the maximum number of concurrent streams the client can have. It is not safe to call this operations while operations are outstanding.

func (APIClient) StartCommit Uses

func (c APIClient) StartCommit(repoName string, branch string) (*pfs.Commit, error)

StartCommit begins the process of committing data to a Repo. Once started you can write to the Commit with PutFile and when all the data has been written you must finish the Commit with FinishCommit. NOTE, data is not persisted until FinishCommit is called. branch is a more convenient way to build linear chains of commits. When a commit is started with a non empty branch the value of branch becomes an alias for the created Commit. This enables a more intuitive access pattern. When the commit is started on a branch the previous head of the branch is used as the parent of the commit.

func (APIClient) StartCommitParent Uses

func (c APIClient) StartCommitParent(repoName string, branch string, parentCommit string) (*pfs.Commit, error)

StartCommitParent begins the process of committing data to a Repo. Once started you can write to the Commit with PutFile and when all the data has been written you must finish the Commit with FinishCommit. NOTE, data is not persisted until FinishCommit is called. branch is a more convenient way to build linear chains of commits. When a commit is started with a non empty branch the value of branch becomes an alias for the created Commit. This enables a more intuitive access pattern. When the commit is started on a branch the previous head of the branch is used as the parent of the commit. parentCommit specifies the parent Commit, upon creation the new Commit will appear identical to the parent Commit, data can safely be added to the new commit without affecting the contents of the parent Commit. You may pass "" as parentCommit in which case the new Commit will have no parent and will initially appear empty.

func (APIClient) StartPipeline Uses

func (c APIClient) StartPipeline(name string) error

StartPipeline restarts a stopped pipeline.

func (APIClient) StopJob Uses

func (c APIClient) StopJob(jobID string) error

StopJob stops a job.

func (APIClient) StopPipeline Uses

func (c APIClient) StopPipeline(name string) error

StopPipeline prevents a pipeline from processing things, it can be restarted with StartPipeline.

func (APIClient) SubscribeCommit Uses

func (c APIClient) SubscribeCommit(repo string, branch string, from string, state pfs.CommitState) (CommitInfoIterator, error)

SubscribeCommit is like ListCommit but it keeps listening for commits as they come in.

func (APIClient) SubscribeCommitF Uses

func (c APIClient) SubscribeCommitF(repo, branch, from string, state pfs.CommitState, f func(*pfs.CommitInfo) error) error

SubscribeCommitF is like ListCommit but it calls a callback function with the results rather than returning an iterator.

func (APIClient) TagObject Uses

func (c APIClient) TagObject(hash string, tags ...string) error

TagObject applies a tag to an existing object.

func (APIClient) Version Uses

func (c APIClient) Version() (string, error)

Version returns the version of pachd as a string.

func (APIClient) Walk Uses

func (c APIClient) Walk(repoName string, commitID string, path string, f WalkFn) error

Walk walks the pfs filesystem rooted at path. walkFn will be called for each file found under path, this includes both regular files and directories.

func (*APIClient) WithCtx Uses

func (c *APIClient) WithCtx(ctx context.Context) *APIClient

WithCtx returns a new APIClient that uses ctx for requests it sends. Note that the new APIClient will still use the authentication token and metrics metadata of this client, so this is only useful for propagating other context-associated metadata.

type AdminAPIClient Uses

type AdminAPIClient admin.APIClient

AdminAPIClient is an alias of admin.APIClient

type AuthAPIClient Uses

type AuthAPIClient auth.APIClient

AuthAPIClient is an alias of auth.APIClient

type CommitInfoIterator Uses

type CommitInfoIterator interface {
    Next() (*pfs.CommitInfo, error)
    Close()
}

CommitInfoIterator wraps a stream of commits and makes them easy to iterate.

type DebugClient Uses

type DebugClient debug.DebugClient

DebugClient is an alias of debug.DebugClient

type DeployAPIClient Uses

type DeployAPIClient deploy.APIClient

DeployAPIClient is an alias of auth.APIClient

type LogsIter Uses

type LogsIter struct {
    // contains filtered or unexported fields
}

LogsIter iterates through log messages returned from pps.GetLogs. Logs can be fetched with 'Next()'. The log message received can be examined with 'Message()', and any errors can be examined with 'Err()'.

func (*LogsIter) Err Uses

func (l *LogsIter) Err() error

Err retrieves any errors encountered in the course of calling 'Next()'.

func (*LogsIter) Message Uses

func (l *LogsIter) Message() *pps.LogMessage

Message returns the most recently retrieve log message (as an annotated log line, in the form of a pps.LogMessage)

func (*LogsIter) Next Uses

func (l *LogsIter) Next() bool

Next retrieves the next relevant log message from pachd

type ObjectAPIClient Uses

type ObjectAPIClient pfs.ObjectAPIClient

ObjectAPIClient is an alias for pfs.ObjectAPIClient

type Option Uses

type Option func(*clientSettings) error

Option is a client creation option that may be passed to NewOnUserMachine(), or NewInCluster()

func WithAdditionalPachdCert Uses

func WithAdditionalPachdCert() Option

WithAdditionalPachdCert instructs the New* functions to additionally trust the signed cert mounted in Pachd's cert volume. This is used by Pachd when connecting to itself (if no cert is present, the clients cert pool will not be modified, so that if no other options have been passed, pachd will connect to itself over an insecure connection)

func WithAdditionalRootCAs Uses

func WithAdditionalRootCAs(pemBytes []byte) Option

WithAdditionalRootCAs instructs the New* functions to additionally trust the given base64-encoded, signed x509 certificates as root certificates. Introduced to pass certs in the Pachyderm config

func WithDialTimeout Uses

func WithDialTimeout(t time.Duration) Option

WithDialTimeout instructs the New* functions to use 't' as the deadline to connect to pachd

func WithMaxConcurrentStreams Uses

func WithMaxConcurrentStreams(streams int) Option

WithMaxConcurrentStreams instructs the New* functions to create client that can have at most 'streams' concurrent streams open with pachd at a time

func WithRootCAs Uses

func WithRootCAs(path string) Option

WithRootCAs instructs the New* functions to create client that uses the given signed x509 certificates as the trusted root certificates (instead of the system certs). Introduced to pass certs provided via command-line flags

type PfsAPIClient Uses

type PfsAPIClient pfs.APIClient

PfsAPIClient is an alias for pfs.APIClient.

type PpsAPIClient Uses

type PpsAPIClient pps.APIClient

PpsAPIClient is an alias for pps.APIClient.

type PutFileClient Uses

type PutFileClient interface {
    // PutFileWriter writes a file to PFS.
    // NOTE: PutFileWriter returns an io.WriteCloser that you must call Close on when
    // you are done writing.
    PutFileWriter(repoName, commitID, path string) (io.WriteCloser, error)

    // PutFileSplitWriter writes multiple files to PFS by splitting up the data
    // that is written to it.
    // NOTE: PutFileSplitWriter returns an io.WriteCloser that you must call Close on when
    // you are done writing.
    PutFileSplitWriter(repoName string, commitID string, path string, delimiter pfs.Delimiter, targetFileDatums int64, targetFileBytes int64, headerRecords int64, overwrite bool) (io.WriteCloser, error)

    // PutFile writes a file to PFS from a reader.
    PutFile(repoName string, commitID string, path string, reader io.Reader) (_ int, retErr error)

    // PutFileOverwrite is like PutFile but it overwrites the file rather than
    // appending to it. overwriteIndex allows you to specify the index of the
    // object starting from which you'd like to overwrite. If you want to
    // overwrite the entire file, specify an index of 0.
    PutFileOverwrite(repoName string, commitID string, path string, reader io.Reader, overwriteIndex int64) (_ int, retErr error)

    // PutFileSplit writes a file to PFS from a reader.
    // delimiter is used to tell PFS how to break the input into blocks.
    PutFileSplit(repoName string, commitID string, path string, delimiter pfs.Delimiter, targetFileDatums int64, targetFileBytes int64, headerRecords int64, overwrite bool, reader io.Reader) (_ int, retErr error)

    // PutFileURL puts a file using the content found at a URL.
    // The URL is sent to the server which performs the request.
    // recursive allows for recursive scraping of some types URLs. For example on s3:// urls.
    PutFileURL(repoName string, commitID string, path string, url string, recursive bool, overwrite bool) error

    // Close must be called after you're done using a PutFileClient.
    // Further requests will throw errors.
    Close() error
}

PutFileClient is a client interface for putting files. There are 2 implementations, 1 that does each file as a seperate request and one that does them all together in the same request.

type PutObjectWriteCloserAsync Uses

type PutObjectWriteCloserAsync struct {
    // contains filtered or unexported fields
}

PutObjectWriteCloserAsync wraps a put object call in an asynchronous buffered writer.

func (*PutObjectWriteCloserAsync) Close Uses

func (w *PutObjectWriteCloserAsync) Close() error

Close closes the writer.

func (*PutObjectWriteCloserAsync) Object Uses

func (w *PutObjectWriteCloserAsync) Object() (*pfs.Object, error)

Object gets the pfs object for this writer. This can only be called when the writer is closed (the put object call is complete)

func (*PutObjectWriteCloserAsync) Write Uses

func (w *PutObjectWriteCloserAsync) Write(p []byte) (int, error)

Write performs a write.

type VersionAPIClient Uses

type VersionAPIClient versionpb.APIClient

VersionAPIClient is an alias of versionpb.APIClient

type WalkFn Uses

type WalkFn func(*pfs.FileInfo) error

WalkFn is the type of the function called for each file in Walk. Returning a non-nil error from WalkFn will result in Walk aborting and returning said error.

Directories

PathSynopsis
adminPackage admin is a generated protocol buffer package.
authPackage auth is a generated protocol buffer package.
debugPackage debug is a generated protocol buffer package.
deployPackage deploy is a generated protocol buffer package.
enterprisePackage enterprise is a generated protocol buffer package.
healthPackage health is a generated protocol buffer package.
limitPackage limit provides primitives to limit concurrency.
pfsPackage pfs is a generated protocol buffer package.
pkg/configPackage config is a generated protocol buffer package.
pkg/discovery
pkg/grpcutil
pkg/pbutil
pkg/require
pkg/shardPackage shard is a generated protocol buffer package.
ppsPackage pps is a generated protocol buffer package.
version
version/versionpbPackage versionpb is a generated protocol buffer package.

Package client imports 37 packages (graph) and is imported by 109 packages. Updated 2018-12-07. Refresh now. Tools for package owners.