5b6aa3bc95
* k8s: add tolerations for 'nodeType=crawling:NoSchedule' to allow scheduling crawling on designated nodes for crawler and profiles jobs and statefulsets * add affinity for 'nodeType=crawling' on crawling and profile browser statefulsets * refactor crawljob: combine crawl_updater logic into base crawl_job * increment new 'crawlAttemptCount' counter crawlconfig when crawl is started, not necessarily finished, to avoid deleting configs that had attempted but not finished crawls. * better external mongodb support: use MONGO_DB_URL to set custom url directly, otherwise build from username, password and mongo host
77 lines
1.8 KiB
YAML
77 lines
1.8 KiB
YAML
apiVersion: batch/v1
|
|
kind: Job
|
|
metadata:
|
|
name: job-{{ id }}
|
|
annotations:
|
|
btrix.run.manual: "{{ manual }}"
|
|
|
|
labels:
|
|
btrix.user: {{ userid }}
|
|
btrix.archive: {{ aid }}
|
|
btrix.crawlconfig: {{ cid }}
|
|
|
|
spec:
|
|
backoffLimit: 1000
|
|
ttlSecondsAfterFinished: 20
|
|
template:
|
|
metadata:
|
|
labels:
|
|
btrix.user: {{ userid }}
|
|
btrix.archive: {{ aid }}
|
|
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: Always
|
|
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
|
|
|
|
- configMapRef:
|
|
name: crawl-config-{{ cid }}
|
|
|
|
env:
|
|
- name: JOB_ID
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.labels['job-name']
|
|
|
|
- name: RUN_MANUAL
|
|
value: "{{ manual }}"
|
|
|
|
volumes:
|
|
- name: config-volume
|
|
configMap:
|
|
name: shared-job-config
|
|
items:
|
|
- key: config.yaml
|
|
path: config.yaml
|
|
|
|
|
|
|