CrowdSec•2mo ago

New issue: $LAPI_HOST resolves to the wrong service name

 acheong@fishy ~/P/k/ingress (master) [1]> kubectl exec -it -n crowdsec pods/crowdsec-agent-k5nxj --container wait-for-lapi-and-register -- sh
/ # ps aux
PID   USER     TIME  COMMAND
    1 root      0:00 sh -c until nc "$LAPI_HOST" "$LAPI_PORT" -z; do echo waiting for lapi to start; sleep 5; done; ln -s /staging/etc/crowdsec /etc/
    8 root      0:00 nc crowdsec-service.crowdsec 8080 -z
    9 root      0:00 sh
   15 root      0:00 ps aux
/ #

 acheong@fishy ~/P/k/ingress (master) [1]> kubectl exec -it -n crowdsec pods/crowdsec-agent-k5nxj --container wait-for-lapi-and-register -- sh
/ # ps aux
PID   USER     TIME  COMMAND
    1 root      0:00 sh -c until nc "$LAPI_HOST" "$LAPI_PORT" -z; do echo waiting for lapi to start; sleep 5; done; ln -s /staging/etc/crowdsec /etc/
    8 root      0:00 nc crowdsec-service.crowdsec 8080 -z
    9 root      0:00 sh
   15 root      0:00 ps aux
/ #

It tries to connect to crowdsec-service.crowdsec which fails when the correct address is crowdsec-service.crowdsec.svc.cluster.local only seems to be happening on one node when all other nodes are running fine. It's been running fine for months and this just randomly popped up

6 Replies

CrowdSec•2mo ago

Important Information

This post has been marked as resolved. If this is a mistake please press the red button below or type /unresolve

Fish alt#3000OP•2mo ago

crowdsec-agent-9q64f             1/1     Running    0          23m     10.42.4.42    kube-worker-3             <none>           <none>
crowdsec-agent-brx52             1/1     Running    0          25m     10.42.7.96    kube-worker-2.novalocal   <none>           <none>
crowdsec-agent-k5nxj             0/1     Init:0/1   0          4m54s   10.42.0.47    kube.duti.dev             <none>           <none>
crowdsec-agent-qjvt6             1/1     Running    0          24m     10.42.8.250   rock-3c                   <none>           <none>
crowdsec-agent-t5m4j             1/1     Running    0          23m     10.42.2.104   kube-worker-4             <none>           <none>
crowdsec-agent-vn2zh             1/1     Running    0          24m     10.42.3.47    kube-worker-1.novalocal   <none>           <none>
crowdsec-lapi-769559c7ff-2cbvz   1/1     Running    0          9m59s   10.42.7.97    kube-worker-2.novalocal   <none>           <none>

crowdsec-agent-9q64f             1/1     Running    0          23m     10.42.4.42    kube-worker-3             <none>           <none>
crowdsec-agent-brx52             1/1     Running    0          25m     10.42.7.96    kube-worker-2.novalocal   <none>           <none>
crowdsec-agent-k5nxj             0/1     Init:0/1   0          4m54s   10.42.0.47    kube.duti.dev             <none>           <none>
crowdsec-agent-qjvt6             1/1     Running    0          24m     10.42.8.250   rock-3c                   <none>           <none>
crowdsec-agent-t5m4j             1/1     Running    0          23m     10.42.2.104   kube-worker-4             <none>           <none>
crowdsec-agent-vn2zh             1/1     Running    0          24m     10.42.3.47    kube-worker-1.novalocal   <none>           <none>
crowdsec-lapi-769559c7ff-2cbvz   1/1     Running    0          9m59s   10.42.7.97    kube-worker-2.novalocal   <none>           <none>

_KaszpiR_•2mo ago

it's rather that node has dns issues or has some additional things related to the DNS config added to the spawned pods

CrowdSec•2mo ago

Resolving New issue: $LAPI_HOST resolves to the wrong service name This has now been resolved. If you think this is a mistake please run /unresolve

Fish alt#3000OP•2mo ago

yup dns seemed to be the issue. not sure what the problem was but I just respawned the node

_KaszpiR_•2mo ago

usually systemd going mental (sticking to the other dns server and not switching back to the one it had )

Gaming

Programming

New issue: $LAPI_HOST resolves to the wrong service name

Did you find this page helpful?