[kubernetes-users] Re: EFS and InitContainer Error

2017-09-26 Thread Rémon Sinnema
Just a random thought: Does the directory /neo4jEfs/conf/ exist? If not, you have to create it before you can write a file into it. On Sunday, September 24, 2017 at 3:15:46 PM UTC+2, shivani...@gmail.com wrote: > > I am trying to setup Neo4jDB with EFS on a specific conf directory. > Using

Re: [kubernetes-users] Need some guidance/help: howto diagnose an oomkill

2017-09-26 Thread John VanRyn
helps some... we made the kube pods have almost twice as much memory as we are allocating the jvm.. and it seems to get us out of the woods but it totally means we need to look into a jdk upgrade from 8. Thanks On Tue, Sep 26, 2017 at 7:50 AM, Davanum Srinivas wrote: >

Re: [kubernetes-users] Resolv.conf override for a pod

2017-09-26 Thread Rodrigo Campos
What I tried to say is using this: http://blog.kubernetes.io/2017/04/configuring-private-dns-zones-upstream-nameservers-kubernetes.html?m=1 in kube-dns configuration. Not sure how your consul name is and, with all you said in the previous mail, a service type external will help. As not even able

[kubernetes-users] edit kube-proxy configmap programmatically

2017-09-26 Thread testmirji
hello. we are automating bootstrapping of k8's on on-prem setup. One of the step includes editing of kube-proxy configmap. we are using command "kubectl edit -n kube-system configmap/kube-proxy". to edit kube-proxy. When I run this command, it opens up the editor. I can modify it and save it.

[kubernetes-users] Servicing Leaders in StatefulSets

2017-09-26 Thread me
Hi all, Given a statefulset that has a single read/write leader and hot standbys, is there an appropriate way to direct the traffic in a Service directly to the leader? The leader is randomly elected from the set of nodes, and while currently the other nodes forward the request internally to

Re: [kubernetes-users] Resolv.conf override for a pod

2017-09-26 Thread Simone D'Andreta
It creates records such as myservice.service.domain.io, so your application must be able to contact a dns forwarder which has the zone for that domain.io. What I am using at the moment are stubdomains to include the domain.io but it's at cluster level. What I need to do is to solve different

Re: [kubernetes-users] Resolv.conf override for a pod

2017-09-26 Thread Rodrigo Campos
Sorry, never used consul and I don't follow what you said. Does it create records like .k8s-service that won't work with just a k8s external type service? Then it might be possible to say to kube dns to use some upstream to some domains, probably? I've not played with it, as I don't need it. But

Re: [kubernetes-users] Need some guidance/help: howto diagnose an oomkill

2017-09-26 Thread Davanum Srinivas
John, Does this help? https://developers.redhat.com/blog/2017/03/14/java-inside-docker/ There are some details here as well: https://github.com/moby/moby/issues/15020 Thanks, Dims On Tue, Sep 26, 2017 at 7:37 AM, John VanRyn wrote: > I have a kube cluster running on

[kubernetes-users] Need some guidance/help: howto diagnose an oomkill

2017-09-26 Thread John VanRyn
I have a kube cluster running on n1-highmem-16 (16 vCPUs, 104 GB memory), using the unmodified cos-stable-60-9592-84-0 image. I have a java app running under wildfly apiVersion: extensions/v1beta1 kind: Deployment metadata: name: cas-unicas-ws labels: name: cas-unicas-ws model:

Re: [kubernetes-users] Resolv.conf override for a pod

2017-09-26 Thread Simone D'Andreta
I am afraid it won't work. I will be able to solve consul.service.domain but I won't be able to solve external services registered within consul, such as mydatabase.service.domain because I need a NS record.. unless I can use that CNAME as a NS record, but I don't think it's the right

Re: [kubernetes-users] Resolv.conf override for a pod

2017-09-26 Thread Simone D'Andreta
Hi Rodrigo, ideally we would need this per pod, but I can give it a try with creating a service per namespace. Thanks for the hint, I will let you know how it goes. Simone Il giorno lunedì 25 settembre 2017 18:34:07 UTC+2, Rodrigo Campos ha scritto: > > Sorry, I must be missing something. But