86ca9c4bac
* backend: fix for total crawl timelimit: - time limit is computed for total job run time - when limit is exceeded, job starts to stop crawls gracefully, equivalent to 'stop crawl' operation - fix for #664 * rename crawl-timeout -> crawl_expire_time * fix lint
116 lines
2.8 KiB
YAML
116 lines
2.8 KiB
YAML
apiVersion: batch/v1
|
|
kind: Job
|
|
metadata:
|
|
name: job-{{ id }}
|
|
annotations:
|
|
btrix.run.manual: "{{ manual }}"
|
|
|
|
labels:
|
|
btrix.user: {{ userid }}
|
|
btrix.org: {{ oid }}
|
|
btrix.crawlconfig: {{ cid }}
|
|
|
|
spec:
|
|
backoffLimit: 1000
|
|
ttlSecondsAfterFinished: 20
|
|
template:
|
|
metadata:
|
|
labels:
|
|
btrix.user: {{ userid }}
|
|
btrix.org: {{ oid }}
|
|
btrix.crawlconfig: {{ cid }}
|
|
spec:
|
|
restartPolicy: OnFailure
|
|
affinity:
|
|
nodeAffinity:
|
|
preferredDuringSchedulingIgnoredDuringExecution:
|
|
- weight: 1
|
|
preference:
|
|
matchExpressions:
|
|
- key: nodeType
|
|
operator: In
|
|
values:
|
|
- "{{ crawler_node_type }}"
|
|
|
|
tolerations:
|
|
- key: "nodeType"
|
|
operator: "Equal"
|
|
value: "crawling"
|
|
effect: "NoSchedule"
|
|
|
|
containers:
|
|
- name: crawl-job
|
|
image: {{ job_image }}
|
|
imagePullPolicy: {{ job_pull_policy }}
|
|
command: ["uvicorn", "btrixcloud.k8s.crawl_job:app", "--host", "0.0.0.0", "--access-log", "--log-level", "info"]
|
|
|
|
volumeMounts:
|
|
- name: config-volume
|
|
mountPath: /config
|
|
|
|
envFrom:
|
|
- secretRef:
|
|
name: mongo-auth
|
|
|
|
env:
|
|
- name: JOB_ID
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.labels['job-name']
|
|
|
|
- name: RUN_MANUAL
|
|
value: "{{ manual }}"
|
|
|
|
- name: USER_ID
|
|
value: "{{ userid }}"
|
|
|
|
- name: ORG_ID
|
|
value: "{{ oid }}"
|
|
|
|
- name: CRAWL_CONFIG_ID
|
|
value: "{{ cid }}"
|
|
|
|
- name: REV
|
|
value: "{{ rev }}"
|
|
|
|
- name: TAGS
|
|
value: "{{ tags }}"
|
|
|
|
- name: CRAWL_EXPIRE_TIME
|
|
value: "{{ crawl_expire_time }}"
|
|
|
|
- name: STORE_PATH
|
|
valueFrom:
|
|
configMapKeyRef:
|
|
name: crawl-config-{{ cid }}
|
|
key: STORE_PATH
|
|
|
|
- name: STORE_FILENAME
|
|
valueFrom:
|
|
configMapKeyRef:
|
|
name: crawl-config-{{ cid }}
|
|
key: STORE_FILENAME
|
|
|
|
- name: STORAGE_NAME
|
|
valueFrom:
|
|
configMapKeyRef:
|
|
name: crawl-config-{{ cid }}
|
|
key: STORAGE_NAME
|
|
|
|
- name: PROFILE_FILENAME
|
|
valueFrom:
|
|
configMapKeyRef:
|
|
name: crawl-config-{{ cid }}
|
|
key: PROFILE_FILENAME
|
|
|
|
volumes:
|
|
- name: config-volume
|
|
configMap:
|
|
name: shared-job-config
|
|
items:
|
|
- key: config.yaml
|
|
path: config.yaml
|
|
|
|
|
|
|