Re: [kubernetes-users] pod migration if the node goes down
Feel free to search the documentation for the default values. But in my experience, it was never an issue (we don't have one replica of anything, though. So a single node going down is pretty tolerable) On Monday, August 6, 2018, Niranjan Kolly wrote: > Hi MR, > > What about any application which are running(like nginx) , what if the > node goes down and how quick master would spin those pods in the available > nodes(for load distribution) > How to tweak the timeout in the contol manager. > > > Thanks, > Niranjan > > On Mon, Aug 6, 2018 at 1:12 PM, 'Matthias Rampke' via Kubernetes user > discussion and Q wrote: > >> It takes a few minutes to declare a node lost, this is configurable via >> kube-controller-manager flags. >> >> There are a few things you can do on a pod that prevents them being put >> on the same node, such as declaring a hostPort. >> >> How will Cassandra react when a pod disappears and another one appears >> (the is no "pod migration", only replacement)? I think this is a fairly >> complex problem; research how others have solved this. A common pattern is >> to name a thing that manages some technology in Kubernetes an Operator, >> googling "cassandra operator kubernetes" I see several candidates but I >> haven't used any so I can't say which one works best for you. >> >> /MR >> >> On Mon, Aug 6, 2018, 09:06 Niranjan Kolly >> wrote: >> >>> Hi, >>> >>> I have a K8 cluster with 3 master and 3 slave on centOS VM. >>> >>> we have installed KONG and cassandra with 3 replica each. As a part of >>> the resiliency we brought down one node , but still the "kubectl get pods" >>> command shows the pods running in that node showing up. >>> >>> As a K8 feature the master should spin the pods in the available node. >>> How to do this configuration, but the master is not doing until we delete >>> the pod manually. >>> >>> Please help me how to setup this. >>> >>> >>> -- >>> Regards >>> NiranjanKolly >>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "Kubernetes user discussion and Q" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to kubernetes-users+unsubscr...@googlegroups.com. >>> To post to this group, send email to kubernetes-users@googlegroups.com. >>> Visit this group at https://groups.google.com/group/kubernetes-users. >>> For more options, visit https://groups.google.com/d/optout. >>> >> -- >> You received this message because you are subscribed to the Google Groups >> "Kubernetes user discussion and Q" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to kubernetes-users+unsubscr...@googlegroups.com. >> To post to this group, send email to kubernetes-users@googlegroups.com. >> Visit this group at https://groups.google.com/group/kubernetes-users. >> For more options, visit https://groups.google.com/d/optout. >> > > > > -- > Regards > NiranjanKolly > > -- > You received this message because you are subscribed to the Google Groups > "Kubernetes user discussion and Q" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to kubernetes-users+unsubscr...@googlegroups.com. > To post to this group, send email to kubernetes-users@googlegroups.com. > Visit this group at https://groups.google.com/group/kubernetes-users. > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "Kubernetes user discussion and Q" group. To unsubscribe from this group and stop receiving emails from it, send an email to kubernetes-users+unsubscr...@googlegroups.com. To post to this group, send email to kubernetes-users@googlegroups.com. Visit this group at https://groups.google.com/group/kubernetes-users. For more options, visit https://groups.google.com/d/optout.
Re: [kubernetes-users] pod migration if the node goes down
Hi MR, What about any application which are running(like nginx) , what if the node goes down and how quick master would spin those pods in the available nodes(for load distribution) How to tweak the timeout in the contol manager. Thanks, Niranjan On Mon, Aug 6, 2018 at 1:12 PM, 'Matthias Rampke' via Kubernetes user discussion and Q wrote: > It takes a few minutes to declare a node lost, this is configurable via > kube-controller-manager flags. > > There are a few things you can do on a pod that prevents them being put on > the same node, such as declaring a hostPort. > > How will Cassandra react when a pod disappears and another one appears > (the is no "pod migration", only replacement)? I think this is a fairly > complex problem; research how others have solved this. A common pattern is > to name a thing that manages some technology in Kubernetes an Operator, > googling "cassandra operator kubernetes" I see several candidates but I > haven't used any so I can't say which one works best for you. > > /MR > > On Mon, Aug 6, 2018, 09:06 Niranjan Kolly wrote: > >> Hi, >> >> I have a K8 cluster with 3 master and 3 slave on centOS VM. >> >> we have installed KONG and cassandra with 3 replica each. As a part of >> the resiliency we brought down one node , but still the "kubectl get pods" >> command shows the pods running in that node showing up. >> >> As a K8 feature the master should spin the pods in the available node. >> How to do this configuration, but the master is not doing until we delete >> the pod manually. >> >> Please help me how to setup this. >> >> >> -- >> Regards >> NiranjanKolly >> >> -- >> You received this message because you are subscribed to the Google Groups >> "Kubernetes user discussion and Q" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to kubernetes-users+unsubscr...@googlegroups.com. >> To post to this group, send email to kubernetes-users@googlegroups.com. >> Visit this group at https://groups.google.com/group/kubernetes-users. >> For more options, visit https://groups.google.com/d/optout. >> > -- > You received this message because you are subscribed to the Google Groups > "Kubernetes user discussion and Q" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to kubernetes-users+unsubscr...@googlegroups.com. > To post to this group, send email to kubernetes-users@googlegroups.com. > Visit this group at https://groups.google.com/group/kubernetes-users. > For more options, visit https://groups.google.com/d/optout. > -- Regards NiranjanKolly -- You received this message because you are subscribed to the Google Groups "Kubernetes user discussion and Q" group. To unsubscribe from this group and stop receiving emails from it, send an email to kubernetes-users+unsubscr...@googlegroups.com. To post to this group, send email to kubernetes-users@googlegroups.com. Visit this group at https://groups.google.com/group/kubernetes-users. For more options, visit https://groups.google.com/d/optout.
[kubernetes-users] pod migration if the node goes down
Hi, I have a K8 cluster with 3 master and 3 slave on centOS VM. we have installed KONG and cassandra with 3 replica each. As a part of the resiliency we brought down one node , but still the "kubectl get pods" command shows the pods running in that node showing up. As a K8 feature the master should spin the pods in the available node. How to do this configuration, but the master is not doing until we delete the pod manually. Please help me how to setup this. -- Regards NiranjanKolly -- You received this message because you are subscribed to the Google Groups "Kubernetes user discussion and Q" group. To unsubscribe from this group and stop receiving emails from it, send an email to kubernetes-users+unsubscr...@googlegroups.com. To post to this group, send email to kubernetes-users@googlegroups.com. Visit this group at https://groups.google.com/group/kubernetes-users. For more options, visit https://groups.google.com/d/optout.
