Commit Graph

14 Commits

Author SHA1 Message Date
Dalton Hubble 991a5c6cee Add new tutorial docs and links 2018-08-27 23:30:32 -07:00
Dalton Hubble e58b424882 Fix firewall to allow etcd client traffic between controllers
* Broaden internal-etcd firewall rule to allow etcd client
traffic (2379) from other controller nodes
* Previously, kube-apiservers were only able to connect to their
node's local etcd peer. While master node outages were tolerated,
reaching a healthy peer took longer than neccessary in some cases
* Reduce time needed to bootstrap a cluster
2018-08-21 23:51:40 -07:00
Dalton Hubble 0764bd30b5 Fix typo in AWS MTU tip for using jumbo packets 2018-06-11 18:11:50 -07:00
Dalton Hubble 4ac4d7cbaf Add docs fixes and Flatcar Linux announcement 2018-05-22 21:22:50 -07:00
Dalton Hubble cd913986df Write documentation for Fedora Atomic 2018-04-24 01:10:27 -07:00
Dalton Hubble ad2e4311d1 Switch GCP network lb to global TCP proxy lb
* Allow multi-controller clusters on Google Cloud
* GCP regional network load balancers have a long open
bug in which requests originating from a backend instance
are routed to the instance itself, regardless of whether
the health check passes or not. As a result, only the 0th
controller node registers. We've recommended just using
single master GCP clusters for a while
* https://issuetracker.google.com/issues/67366622
* Workaround issue by switching to a GCP TCP Proxy load
balancer. TCP proxy lb routes traffic to a backend service
(global) of instance group backends. In our case, spread
controllers across 3 zones (all regions have 3+ zones) and
organize them in 3 zonal unmanaged instance groups that
serve as backends. Allows multi-controller cluster creation
* GCP network load balancers only allowed legacy HTTP health
checks so kubelet 10255 was checked as an approximation of
controller health. Replace with TCP apiserver health checks
to detect unhealth or unresponsive apiservers.
* Drawbacks: GCP provision time increases, tailed logs now
timeout (similar tradeoff in AWS), controllers only span 3
zones instead of the exact number in the region
* Workaround in Typhoon has been known and posted for 5 months,
but there still appears to be no better alternative. Its
probably time to support multi-master and accept the downsides
2018-04-18 00:09:06 -07:00
Khris Richardson e623439eec Fix typos in docs and CONTRIBUTING.md 2017-12-09 19:58:09 -08:00
Dalton Hubble 58cf82da56 Promote AWS platform from alpha to beta 2017-11-06 21:38:24 -08:00
Dalton Hubble 805dd772a8 Run etcd cluster on-host, across controllers on AWS
* Change controllers ASG to heterogeneous EC2 instances
* Create DNS records for each controller's private IP for etcd
* Change etcd to run on-host, across controllers (etcd-member.service)
* Reduce time to bootstrap a cluster
* Deprecate self-hosted-etcd on the AWS platform
2017-11-06 01:03:53 -08:00
Dalton Hubble 7b38271212 Run etcd cluster on-host, across controllers on Google Cloud
* Change controllers from a managed group to individual instances
* Create discrete DNS records to each controller's private IP for etcd
* Change etcd to run on-host, across controllers (etcd-member.service)
* Reduce time to bootstrap a cluster
* Deprecate self-hosted-etcd on the Google Cloud platform
2017-11-05 11:03:35 -08:00
Dalton Hubble 67e3d2b86e docs: GCE network bandwidth is excellent, even btw zones
* Remove performance note that the GCE vs AWS network performance
is not an equal comparison. On both platforms, workers now span the
(availability) zones of a region.
* Testing host-to-host and pod-to-pod network bandwidth between nodes
(now located in different zones) showed no reduction in bandwidth
2017-11-04 14:08:20 -07:00
Dalton Hubble 734bc1d32a Add performance benchmark for flannel with bonded NICs 2017-10-16 01:12:13 -07:00
Dalton Hubble 308c7dfb6e digital-ocean: Run etcd cluster on-host, across controllers
* Run etcd peers with TLS across controller nodes
* Deprecate self-hosted-etcd on the Digital Ocean platform
* Distribute etcd TLS certificates as part of initial provisioning
* Check the status of etcd by running `systemctl status etcd-member`
2017-10-09 22:43:23 -07:00
Dalton Hubble 62d7ccfff3 Add docs on provision time and network performance 2017-10-04 00:05:43 -07:00