/// <summary>
/// Process a single dataset and fetch the info about it.
/// </summary>
protected override void ProcessRecord()
{
var trimmedDSName = DatasetName.Trim();
if (_resultsCache.Value[trimmedDSName] is PSGRIDDatasetInfo cHit)
{
WriteObject(cHit);
}
else
{
// Setup for verbosity if we need it.
var listener = new PSListener(this);
Trace.Listeners.Add(listener);
try
{
// Where are we going to be doing query on - we need a machine.
var sm = JobParser.GetSubmissionMachine();
// Get the remove environment configured if it needs to be
if (_connection == null)
{
_connection = new SSHConnection(sm.MachineName, sm.UserName);
_connection
.Apply(() => DisplayStatus("Setting up ATLAS"))
.setupATLAS()
.Apply(() => DisplayStatus("Setting up Rucio"))
.setupRucio(_gridCredentials.Username)
.Apply(() => DisplayStatus("Acquiring GRID credentials"))
.VomsProxyInit("atlas", failNow: () => Stopping);
}
// Great - get the info on this dataset.
var fileInfo = _connection
.Apply(() => DisplayStatus($"Checking for info on {trimmedDSName}."))
.FileInfoFromGRID(trimmedDSName, failNow: () => Stopping);
// Next, build the resulting thingy.
var r = new PSGRIDDatasetInfo()
{
DatasetName = trimmedDSName,
nFiles = fileInfo.Count,
TotalSizeMB = (int)fileInfo.Sum(fi => fi.size),
FileInfo = fileInfo.ToArray()
};
_resultsCache.Value[trimmedDSName] = r;
using (var pp = listener.PauseListening())
{
WriteObject(r);
}
}
finally
{
Trace.Listeners.Remove(listener);
}
}
}
/// <summary>
/// Fetch a particular dataset.
/// </summary>
protected override void ProcessRecord()
{
var listener = new PSListener(this);
Trace.Listeners.Add(listener);
try
{
// Get the dataset name that we are to process. Depending on parameters, it can come
// from one of two places.
var dataset = ParameterSetName == "job"
? GetJobDatasetName()
: DatasetName.Trim();
// Find all the members of this dataset.
var allFilesToCopy = DataSetManager.ListOfFilesInDataSetAsync(dataset, m => DisplayStatus($"Listing Files in {dataset}", m), failNow: () => Stopping).Result;
if (nFiles != 0)
{
allFilesToCopy = allFilesToCopy
.OrderBy(u => u.AbsolutePath)
.Take(nFiles)
.ToArray();
}
// If we have a location, we want to do a copy. If we don't have a location, then we just
// want to get the files somewhere local.
var loc = Location.AsIPlace().Result;
//WriteError(new ErrorRecord(err, "NoSuchLocation", ErrorCategory.InvalidArgument, null));
var resultUris = loc == null
? DataSetManager.MakeFilesLocalAsync(allFilesToCopy, m => DisplayStatus($"Downloading {dataset}", m), failNow : () => Stopping, timeout : Timeout).Result
: DataSetManager.CopyFilesToAsync(loc, allFilesToCopy, m => DisplayStatus($"Downloading {dataset}", m), failNow: () => Stopping, timeout: Timeout).Result;
// Dump all the returned files out to whatever is next in the pipeline.
if (PassThru.IsPresent)
{
using (var pl = listener.PauseListening())
{
foreach (var ds in resultUris)
{
WriteObject(ds);
}
}
}
}
finally
{
ClearStatus();
Trace.Listeners.Remove(listener);
}
}
/// <summary>
/// Get the list and write it out as objects.
/// </summary>
protected override void ProcessRecord()
{
var listener = new PSListener(this);
Trace.Listeners.Add(listener);
try
{
var list = DataSetManager.ValidLocations;
foreach (var l in list)
{
using (var pl = listener.PauseListening())
{
WriteObject(l);
}
}
} finally
{
Trace.Listeners.Remove(listener);
}
}
/// <summary>
/// Fetch a particular dataset.
/// </summary>
protected override void ProcessRecord()
{
var listener = new PSListener(this);
Trace.Listeners.Add(listener);
try
{
// Get the actual dataset name.
var dataset = DatasetName;
if (ParameterSetName == "job")
{
// Get the job, see if it is finished, and then get the output dataset.
var job = JobParser.FindJob(JobName, JobVersion);
if (job == null)
{
throw new ArgumentException($"Job {JobName} v{JobVersion} is not known to the system!");
}
// Get the resulting job name for this guy.
var pandaJobName = job.ResultingDataSetName(JobSourceDatasetName) + "/";
// Now, to get the output dataset, we need to fetch the job.
var pandaTask = pandaJobName.FindPandaJobWithTaskName(true);
if (pandaTask == null)
{
throw new ArgumentException($"No panda task found with name '{pandaJobName}' for job '{JobName}' v{JobVersion}.");
}
var containers = pandaTask.DataSetNamesOUT();
if (containers.Length > 1)
{
throw new ArgumentException($"There are more than one output container for the panda task {pandaTask.jeditaskid} - can't decide. Need code upgrade!! Thanks for the fish!");
}
dataset = containers.First();
}
// Find all the members of this dataset.
var allFilesToCopy = DataSetManager.ListOfFilesInDataSetAsync(dataset, m => DisplayStatus($"Listing Files in {dataset}", m), failNow: () => Stopping)
.Result;
if (nFiles != 0)
{
allFilesToCopy = allFilesToCopy
.OrderBy(u => u.AbsolutePath)
.Take(nFiles)
.ToArray();
}
// Next, see what the test is. It is only exists right now.
if (!Exists)
{
throw new ArgumentException("Test-GRIDDataset just use the -Exists flag.");
}
// Query the location to see if we have a decent copy there.
var loc = Location.AsIPlace().Result;
var hasFiles = Task.WhenAll(allFilesToCopy
.Select(f => loc.HasFileAsync(f, m => DisplayStatus($"Checking Files {dataset}", m), failNow: () => Stopping)))
.Result
.All(f => f);
// Dump all the returned files out to whatever is next in the pipeline.
using (var pl = listener.PauseListening())
{
WriteObject(hasFiles);
}
}
finally
{
Trace.Listeners.Remove(listener);
}
}
/// <summary>
/// Fetch a particular dataset.
/// </summary>
protected override void ProcessRecord()
{
var listener = new PSListener(this);
Trace.Listeners.Add(listener);
try
{
// Get the actual dataset name.
var dataset = DatasetName.Trim();
if (ParameterSetName == "job")
{
// Get the job, see if it is finished, and then get the output dataset.
var job = JobParser.FindJob(JobName, JobVersion);
if (job == null)
{
throw new ArgumentException($"Job {JobName} v{JobVersion} is not known to the system!");
}
// Get the resulting job name for this guy.
var pandaJobName = job.ResultingDataSetName(JobSourceDatasetName, JobIteration) + "/";
// Now, to get the output dataset, we need to fetch the job.
var pandaTask = pandaJobName.FindPandaJobWithTaskName(true);
if (pandaTask == null)
{
throw new ArgumentException($"No panda task found with name '{pandaJobName}' for job '{JobName}' v{JobVersion}.");
}
var containers = pandaTask.DataSetNamesOUT();
if (containers.Length > 1)
{
throw new ArgumentException($"There are more than one output container for the panda task {pandaTask.jeditaskid} - can't decide. Need code upgrade!! Thanks for the fish!");
}
dataset = containers.First();
}
// Find all the members of this dataset.
var allFilesToCopy = DataSetManager.ListOfFilesInDataSetAsync(dataset, m => DisplayStatus($"Listing Files in {dataset}", m), failNow: () => Stopping)
.Result;
if (nFiles != 0)
{
allFilesToCopy = allFilesToCopy
.OrderBy(u => u.AbsolutePath)
.Take(nFiles)
.ToArray();
}
// Turn the source and destination locations into actual locations.
var locSource = SourceLocation.AsIPlace().Result;
var locDest = DestinationLocation.AsIPlace().Result;
// Do the actual copy. This will fail if one of the files can't be found at the source.
var resultUris = DataSetManager
.CopyFilesAsync(locSource, locDest, allFilesToCopy, mbox => DisplayStatus($"Downloading {dataset}", mbox), failNow: () => Stopping, timeout: Timeout)
.Result;
// Dump all the returned files out to whatever is next in the pipeline.
if (PassThru.IsPresent)
{
using (var pl = listener.PauseListening())
{
foreach (var ds in resultUris)
{
WriteObject(ds);
}
}
}
}
finally
{
Trace.Listeners.Remove(listener);
}
}
/// <summary>
/// Load up the job requested. Fail, obviously, if we can't.
/// </summary>
protected override void ProcessRecord()
{
// Setup for verbosity if we need it.
var listener = new PSListener(this);
Trace.Listeners.Add(listener);
try
{
// Get the job
var job = JobParser.FindJob(JobName, JobVersion);
// Get the expected resulting dataset name. Since this will be a personal
// dataset, we need to get the GRID info.
var originalDatasetName = DatasetName.Trim();
string resultDatasetName = job.ResultingDataSetName(originalDatasetName, _gridCredentials, JobIteration);
// See if there is already a job defined that will produce this
var pandaJob = (resultDatasetName + "/").FindPandaJobWithTaskName(useCacheIfPossible: !DoNotUsePandaTaskCache.IsPresent);
if (pandaJob == null)
{
// Where are we going to be doing the submission on?
var sm = JobParser.GetSubmissionMachine();
// Get the remove environment configured if it needs to be
var firstJob = false;
if (_connection == null)
{
firstJob = true;
_connection = new SSHConnection(sm.MachineName, sm.UserName);
_connection
.Apply(() => DisplayStatus("Setting up ATLAS"))
.setupATLAS(dumpOnly: WhatIf.IsPresent)
.Apply(() => DisplayStatus("Setting up Rucio"))
.setupRucio(_gridCredentials.Username, dumpOnly: WhatIf.IsPresent)
.Apply(() => DisplayStatus("Acquiring GRID credentials"))
.VomsProxyInit("atlas", failNow: () => Stopping, dumpOnly: WhatIf.IsPresent);
}
// Check to see if the original dataset exists. We will use the location known as Local for doing the
// setup, I suppose.
var files = _connection
.Apply(() => DisplayStatus("Checking dataset exists on the GRID"))
.FilelistFromGRID(originalDatasetName, failNow: () => Stopping, dumpOnly: WhatIf.IsPresent);
if (files.Length == 0 && !WhatIf.IsPresent)
{
throw new ArgumentException($"Dataset '{originalDatasetName}' has zero files - won't submit a job against it!");
}
// Submit the job
_connection
.SubmitJobAsync(job, originalDatasetName, resultDatasetName, DisplayStatus, failNow: () => Stopping, sameJobAsLastTime: !firstJob, dumpOnly: WhatIf.IsPresent)
.WaitAndUnwrapException();
// Try to find the job again if requested. The submission can take a very long time to show up in
// big panda, so skip unless requested.
if (WaitForPandaRegistration)
{
var pandJobResult = Policy
.Handle <InvalidOperationException>()
.WaitAndRetryForever(nthRetry => TimeSpan.FromMinutes(1),
(e, ts) => { WriteWarning($"Failed to find the submitted panda job on bigpanda: {e.Message}. Will wait one minute and try again."); })
.ExecuteAndCapture(() => FindPandaJobForDS(resultDatasetName));
if (pandJobResult.Outcome != OutcomeType.Successful)
{
throw pandJobResult.FinalException;
}
pandaJob = pandJobResult.Result;
}
}
// Return a helper obj that contains the info about this job that can be used by other commands.
if (pandaJob != null)
{
var r = new AtlasPandaTaskID()
{
ID = pandaJob.jeditaskid, Name = pandaJob.taskname
};
using (var pp = listener.PauseListening())
{
WriteObject(r);
}
}
} finally
{
Trace.Listeners.Remove(listener);
}
}