我有包含python文件的docker映像,应该从scihub网站下载卫星映像。泊坞窗映像工作正常。现在,当我想创建部署方式kubectl
以便将其作为服务公开时,它的容器将继续崩溃。这就是通过kubectl describe pod
进行查看时广告连播描述的内容。
这就是我尝试部署sudo kubectl run back --image=back:latest --port=8080 --image-pull-policy Never
的方式。我也尝试更改端口,但没有成功。这是docker映像中的文件。
Docker文件
FROM python:3.7-stretch
COPY . /code
WORKDIR /code
RUN pip install -r requirements.txt
ENTRYPOINT ["python", "ingestion.py"]
**摄入**
import os
import shutil
import logging
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(name)s - %(message)s')
logger = logging.getLogger("ingestion")
import requests
import datahub
scihub_username = os.environ["scihub_username"]
scihub_password = os.environ["scihub_password"]
result_url = "http://" + os.environ["CDINRW_BASE_URL"] + "/jobs/" + os.environ["CDINRW_JOB_ID"] + "/results"
logger.info("Searching the Copernicus Open Access Hub")
scenes = datahub.search(username=scihub_username,
password=scihub_password,
producttype=os.getenv("producttype"),
platformname=os.getenv("platformname"),
days_back=os.getenv("days_back", 2),
footprint=os.getenv("footprint"),
max_cloud_cover_percentage=os.getenv("max_cloud_cover_percentage"),
start_date = os.getenv("start_date"),
end_date = os.getenv("end_date"))
logger.info("Found {} relevant scenes".format(len(scenes)))
job_results = []
for scene in scenes:
# do not donwload a scene that has already been ingested
if os.path.exists(os.path.join("/out_data", scene["title"]+".SAFE")):
logger.info("The scene {} already exists in /out_data and will not be downloaded again.".format(scene["title"]))
filename = scene["title"]+".SAFE"
else:
logger.info("Starting the download of scene {}".format(scene["title"]))
filename = datahub.download(scene, "/tmp", scihub_username, scihub_password, unpack=True)
logger.info("The download was successful.")
shutil.move(filename, "/out_data")
result_message = {"description": "test",
"type": "Raster",
"format": "SAFE",
"filename": os.path.basename(filename)}
job_results.append(result_message)
res = requests.put(result_url, json=job_results, timeout=60)
res.raise_for_status()
* datahub
import logging
import os
import urllib.parse
import zipfile
import requests
# constructing URLs for querying the data hub
_BASE_URL = "https://scihub.copernicus.eu/dhus/"
SITE = {}
SITE["SEARCH"] = _BASE_URL + "search?format=xml&sortedby=beginposition&order=desc&rows=100&start={offset}&q="
_PRODUCT_URL = _BASE_URL + "odata/v1/Products('{uuid}')/"
SITE["CHECKSUM"] = _PRODUCT_URL + "Checksum/Value/$value"
SITE["SAFEZIP"] = _PRODUCT_URL + "$value"
logger = logging.getLogger(__name__)
def _build_search_url(producttype=None, platformname=None, days_back=2, footprint=None, max_cloud_cover_percentage=None, start_date=None, end_date=None):
search_terms = []
if producttype:
search_terms.append("producttype:{}".format(producttype))
if platformname:
search_terms.append("platformname:{}".format(platformname))
if start_date and end_date:
search_terms.append(
"beginPosition:[{}+TO+{}]".format(start_date, end_date))
elif days_back:
search_terms.append(
"beginPosition:[NOW-{}DAYS+TO+NOW]".format(days_back))
if footprint:
search_terms.append("footprint:%22Intersects({})%22".format(
footprint.replace(" ", "+")))
if max_cloud_cover_percentage:
search_terms.append("cloudcoverpercentage:[0+TO+{}]".format(max_cloud_cover_percentage))
url = SITE["SEARCH"] + "+AND+".join(search_terms)
return url
def _unpack(zip_file, directory, remove_after=False):
with zipfile.ZipFile(zip_file) as zf:
# This assumes that the zipfile only contains the .SAFE directory at root level
safe_path = zf.namelist()[0]
zf.extractall(path=directory)
if remove_after:
os.remove(zip_file)
return os.path.normpath(os.path.join(directory, safe_path))
def search(username, password, producttype=None, platformname=None ,days_back=2, footprint=None, max_cloud_cover_percentage=None, start_date=None, end_date=None):
""" Search the Copernicus SciHub
Parameters
----------
username : str
user name for the Copernicus SciHub
password : str
password for the Copernicus SciHub
producttype : str, optional
product type to filter for in the query (see https://scihub.copernicus.eu/userguide/FullTextSearch#Search_Keywords for allowed values)
platformname : str, optional
plattform name to filter for in the query (see https://scihub.copernicus.eu/userguide/FullTextSearch#Search_Keywords for allowed values)
days_back : int, optional
number of days before today that will be searched. Default are the last 2 days. If start and end date are set the days_back parameter is ignored
footprint : str, optional
well-known-text representation of the footprint
max_cloud_cover_percentage: str, optional
percentage of cloud cover per scene. Can only be used in combination with Sentinel-2 imagery.
(see https://scihub.copernicus.eu/userguide/FullTextSearch#Search_Keywords for allowed values)
start_date: str, optional
start point of the search extent has to be used in combination with end_date
end_date: str, optional
end_point of the search extent has to be used in combination with start_date
Returns
-------
list
a list of scenes that match the search parameters
"""
import xml.etree.cElementTree as ET
scenes = []
search_url = _build_search_url(producttype, platformname, days_back, footprint, max_cloud_cover_percentage, start_date, end_date)
logger.info("Search URL: {}".format(search_url))
offset = 0
rowsBreak = 5000
name_space = {"atom": "http://www.w3.org/2005/Atom",
"opensearch": "http://a9.com/-/spec/opensearch/1.1/"}
while offset < rowsBreak: # Next pagination page:
response = requests.get(search_url.format(offset=offset), auth=(username, password))
root = ET.fromstring(response.content)
if offset == 0:
rowsBreak = int(
root.find("opensearch:totalResults", name_space).text)
for e in root.iterfind("atom:entry", name_space):
uuid = e.find("atom:id", name_space).text
title = e.find("atom:title", name_space).text
begin_position = e.find(
"atom:date[@name='beginposition']", name_space).text
end_position = e.find(
"atom:date[@name='endposition']", name_space).text
footprint = e.find("atom:str[@name='footprint']", name_space).text
scenes.append({
"id": uuid,
"title": title,
"begin_position": begin_position,
"end_position": end_position,
"footprint": footprint})
# Ultimate DHuS pagination page size limit (rows per page).
offset += 100
return scenes
def download(scene, directory, username, password, unpack=True):
""" Download a Sentinel scene based on its uuid
Parameters
----------
scene : dict
the scene to be downloaded
path : str
the path where the file will be downloaded to
username : str
username for the Copernicus SciHub
password : str
password for the Copernicus SciHub
unpack: boolean, optional
flag that defines whether the downloaded product should be unpacked after download. defaults to true
Raises
------
ValueError
if the size of the downloaded file does not match the Content-Length header
ValueError
if the checksum of the downloaded file does not match the checksum provided by the Copernicus SciHub
Returns
-------
str
path to the downloaded file
"""
import hashlib
md5hash = hashlib.md5()
md5sum = requests.get(SITE["CHECKSUM"].format(
uuid=scene["id"]), auth=(username, password)).text
download_path = os.path.join(directory, scene["title"] + ".zip")
# overwrite if path already exists
if os.path.exists(download_path):
os.remove(download_path)
url = SITE["SAFEZIP"].format(uuid=scene["id"])
rsp = requests.get(url, auth=(username, password), stream=True)
cl = rsp.headers.get("Content-Length")
size = int(cl) if cl else -1
# Actually fetch now:
with open(download_path, "wb") as f: # Do not read as a whole into memory:
written = 0
for block in rsp.iter_content(8192):
f.write(block)
written += len(block)
md5hash.update(block)
written = os.path.getsize(download_path)
if size > -1 and written != size:
raise ValueError("{}: size mismatch, {} bytes written but expected {} bytes to write!".format(
download_path, written, size))
elif md5sum:
calculated = md5hash.hexdigest()
expected = md5sum.lower()
POD事件
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning BackOff 2m39s (x18636 over 2d19h) kubelet, minikube Back-off restarting failed container
要使用此服务的系统已经在8081上运行了另一个主要的前端服务(仅运行应用程序),所以也许我需要在同一端口上公开该服务。如何使部署运行?