import logging
+import json
+import os
+
+from cwltool.errors import WorkflowException
+from cwltool.process import get_feature, UnsupportedRequirement, shortname
+from cwltool.pathmapper import adjustFiles
+
import arvados.collection
-from cwltool.process import get_feature, adjustFiles
+
from .arvdocker import arv_docker_get_image
from . import done
-from cwltool.errors import WorkflowException
-from cwltool.process import UnsupportedRequirement
+from .runner import Runner
logger = logging.getLogger('arvados.cwl-runner')
class ArvadosContainer(object):
- """Submit and manage a Crunch job for executing a CWL CommandLineTool."""
+ """Submit and manage a Crunch container request for executing a CWL CommandLineTool."""
def __init__(self, runner):
self.arvrunner = runner
self.running = False
+ self.uuid = None
def update_pipeline_component(self, r):
pass
"command": self.command_line,
"owner_uuid": self.arvrunner.project_uuid,
"name": self.name,
- "output_path": "/var/spool/cwl",
- "cwd": "/var/spool/cwl",
+ "output_path": self.outdir,
+ "cwd": self.outdir,
"priority": 1,
"state": "Committed"
}
runtime_constraints = {}
mounts = {
- "/var/spool/cwl": {
+ self.outdir: {
"kind": "tmp"
}
}
+ dirs = set()
for f in self.pathmapper.files():
- _, p = self.pathmapper.mapper(f)
- mounts[p] = {
- "kind": "collection",
- "portable_data_hash": p[6:]
- }
+ _, p, tp = self.pathmapper.mapper(f)
+ if tp == "Directory" and '/' not in p[6:]:
+ mounts[p] = {
+ "kind": "collection",
+ "portable_data_hash": p[6:]
+ }
+ dirs.add(p[6:])
+ for f in self.pathmapper.files():
+ _, p, tp = self.pathmapper.mapper(f)
+ if p[6:].split("/")[0] not in dirs:
+ mounts[p] = {
+ "kind": "collection",
+ "portable_data_hash": p[6:]
+ }
- if self.generatefiles:
- raise UnsupportedRequirement("Stdin redirection currently not suppported")
+ if self.generatefiles["listing"]:
+ raise UnsupportedRequirement("Generate files not supported")
- vwd = arvados.collection.Collection()
- container_request["task.vwd"] = {}
- for t in self.generatefiles:
- if isinstance(self.generatefiles[t], dict):
- src, rest = self.arvrunner.fs_access.get_collection(self.generatefiles[t]["path"].replace("$(task.keep)/", "keep:"))
- vwd.copy(rest, t, source_collection=src)
- else:
- with vwd.open(t, "w") as f:
- f.write(self.generatefiles[t])
- vwd.save_new()
- # TODO
- # for t in self.generatefiles:
- # container_request["task.vwd"][t] = "$(task.keep)/%s/%s" % (vwd.portable_data_hash(), t)
-
- container_request["environment"] = {"TMPDIR": "/tmp"}
+ container_request["environment"] = {"TMPDIR": self.tmpdir, "HOME": self.outdir}
if self.environment:
container_request["environment"].update(self.environment)
if self.stdin:
raise UnsupportedRequirement("Stdin redirection currently not suppported")
+ if self.stderr:
+ raise UnsupportedRequirement("Stderr redirection currently not suppported")
+
if self.stdout:
mounts["stdout"] = {"kind": "file",
- "path": "/var/spool/cwl/%s" % (self.stdout)}
+ "path": "%s/%s" % (self.outdir, self.stdout)}
(docker_req, docker_is_req) = get_feature(self, "DockerRequirement")
if not docker_req:
if resources is not None:
runtime_constraints["vcpus"] = resources.get("cores", 1)
runtime_constraints["ram"] = resources.get("ram") * 2**20
- #runtime_constraints["min_scratch_mb_per_node"] = resources.get("tmpdirSize", 0) + resources.get("outdirSize", 0)
container_request["mounts"] = mounts
container_request["runtime_constraints"] = runtime_constraints
body=container_request
).execute(num_retries=self.arvrunner.num_retries)
- self.arvrunner.jobs[response["container_uuid"]] = self
+ self.arvrunner.processes[response["container_uuid"]] = self
- logger.info("Container %s (%s) request state is %s", self.name, response["container_uuid"], response["state"])
+ logger.info("Container %s (%s) request state is %s", self.name, response["uuid"], response["state"])
if response["state"] == "Final":
self.done(response)
def done(self, record):
try:
if record["state"] == "Complete":
- processStatus = "success"
+ rcode = record["exit_code"]
+ if self.successCodes and rcode in self.successCodes:
+ processStatus = "success"
+ elif self.temporaryFailCodes and rcode in self.temporaryFailCodes:
+ processStatus = "temporaryFail"
+ elif self.permanentFailCodes and rcode in self.permanentFailCodes:
+ processStatus = "permanentFail"
+ elif rcode == 0:
+ processStatus = "success"
+ else:
+ processStatus = "permanentFail"
else:
processStatus = "permanentFail"
try:
outputs = {}
if record["output"]:
- outputs = done.done(self, record, "/tmp", "/var/spool/cwl", "/keep")
+ outputs = done.done(self, record, "/tmp", self.outdir, "/keep")
except WorkflowException as e:
logger.error("Error while collecting container outputs:\n%s", e, exc_info=(e if self.arvrunner.debug else False))
processStatus = "permanentFail"
self.output_callback(outputs, processStatus)
finally:
- del self.arvrunner.jobs[record["uuid"]]
+ del self.arvrunner.processes[record["uuid"]]
+
+
+class RunnerContainer(Runner):
+ """Submit and manage a container that runs arvados-cwl-runner."""
+
+ def arvados_job_spec(self, dry_run=False, pull_image=True, **kwargs):
+ """Create an Arvados container request for this workflow.
+
+ The returned dict can be used to create a container passed as
+ the +body+ argument to container_requests().create().
+ """
+
+ workflowmapper = super(RunnerContainer, self).arvados_job_spec(dry_run=dry_run, pull_image=pull_image, **kwargs)
+
+ with arvados.collection.Collection(api_client=self.arvrunner.api) as jobobj:
+ with jobobj.open("cwl.input.json", "w") as f:
+ json.dump(self.job_order, f, sort_keys=True, indent=4)
+ jobobj.save_new(owner_uuid=self.arvrunner.project_uuid)
+
+ workflowname = os.path.basename(self.tool.tool["id"])
+ workflowpath = "/var/lib/cwl/workflow/%s" % workflowname
+ workflowcollection = workflowmapper.mapper(self.tool.tool["id"])[1]
+ workflowcollection = workflowcollection[5:workflowcollection.index('/')]
+ jobpath = "/var/lib/cwl/job/cwl.input.json"
+
+ container_image = arv_docker_get_image(self.arvrunner.api,
+ {"dockerImageId": "arvados/jobs"},
+ pull_image,
+ self.arvrunner.project_uuid)
+
+ return {
+ "command": ["arvados-cwl-runner", "--local", "--api=containers", workflowpath, jobpath],
+ "owner_uuid": self.arvrunner.project_uuid,
+ "name": self.name,
+ "output_path": "/var/spool/cwl",
+ "cwd": "/var/spool/cwl",
+ "priority": 1,
+ "state": "Committed",
+ "container_image": container_image,
+ "mounts": {
+ "/var/lib/cwl/workflow": {
+ "kind": "collection",
+ "portable_data_hash": "%s" % workflowcollection
+ },
+ jobpath: {
+ "kind": "collection",
+ "portable_data_hash": "%s/cwl.input.json" % jobobj.portable_data_hash()
+ },
+ "stdout": {
+ "kind": "file",
+ "path": "/var/spool/cwl/cwl.output.json"
+ },
+ "/var/spool/cwl": {
+ "kind": "collection",
+ "writable": True
+ }
+ },
+ "runtime_constraints": {
+ "vcpus": 1,
+ "ram": 1024*1024*256,
+ "API": True
+ }
+ }
+
+ def run(self, *args, **kwargs):
+ kwargs["keepprefix"] = "keep:"
+ job_spec = self.arvados_job_spec(*args, **kwargs)
+ job_spec.setdefault("owner_uuid", self.arvrunner.project_uuid)
+
+ response = self.arvrunner.api.container_requests().create(
+ body=job_spec
+ ).execute(num_retries=self.arvrunner.num_retries)
+
+ self.uuid = response["uuid"]
+ self.arvrunner.processes[response["container_uuid"]] = self
+
+ logger.info("Submitted container %s", response["uuid"])
+
+ if response["state"] in ("Complete", "Failed", "Cancelled"):
+ self.done(response)