Updates magnum documentation to reflect the changes made for
bp/rename-id-parameter-names.
Change-Id: I7c0d69970b9ac50649cac0e80932b5971ce09c3f
Depends-on: Idc6b9322ec7c0ecb386162ca46c7d49b910685d8
Implements: blueprint rename-id-parameter-names
The troubleshooting guide has a python script in it which has invalid
syntax. This removes the extra bracket from that script and makes it
function correctly. I also added some more specificity for credentials.
Change-Id: I9c8b8f861b4226d9eddca4ab44c811faa89160bc
We must add the user-domain to ensure than the user is from
the magnum domain.
Change-Id: I8defb35f745f7df2e20deea759a43ef1bb47d248
Closes-bug: #1635202
There is an environment that the google public DNS is not
able to use for the --dns-servername.
Before specifying the dns-servername, it's better to confirm
that the DNS can resolve server names properly at your
environment, otherwise, you will fail cluster creation.
Also this commit fixes some old links at the network
troubleshooting section of ops guide.
Change-Id: I2f765a8ff94ad54f3fb1fb029a36fdea328ea514
This is a continuation of patch 353726 and includes all of the
doc changes for replacing the term bay with cluster and
BayModel with ClusterTemplate.
Change-Id: Ia7efaed157971ad7631ddffb9c1400f3516720f0
Implements: blueprint rename-bay-to-cluster
In flanneld version 0.5.4(on fedora-atomic-23 iamge), the default flannel
etcd key changed from /coreos.com/network to /atomic.io/network.
Use $FLANNEL_ETCD_KEY in /etc/sysconfig/flanneld for debugging.
Change-Id: I13e26061b7a6b33878f90834eacde9c4bc6e72d7
Change the devstack image to the default Atomic 23 image.
Update templates not to rely on interface naming because the names
changed between Atomic 21 and Atomic 23 images.
Add a functional test to validate minion node registration.
Update docs to refer to a generic Atomic image.
Co-Authored-By: Hongbin Lu <hongbin.lu@huawei.com>
Implements: blueprint atomic23
Change-Id: Id67244e3f452b3c7c3c46fdabcca3a5c141a76c0
Add section on etcd to troubleshooting guide.
etcd provides key/value pair storage and management for
many services in the COE, therefore if it fails, many
other services will be failing also. In some cases, the
secondary failures can be rather mysterious, for instance
kube-apiserver will simply fail to start with no specific
error message.
This section covers the basic failure scenario and gives some
pointers to verify the correct configuration, operation.
Partially implements: blueprint magnum-troubleshooting-guide
Change-Id: I602a9c3b8e54796c72848cf945107a319e45b973
Add section on Flannel to troubleshooting guide.
Flannel provides the overlay network and is the default
network driver for Kubernetes COE. It is also an optional
network driver for Swarm COE. Failure in Flannel would
mean that the containers cannot communicate with each other.
This section covers common failures, how to verify configuration,
how to check for correct operation, logs, known limitations.
If you have encountered other scenarios or have additional tips,
please comment or add your own patch.
Partially implements: blueprint magnum-troubleshooting-guide
Change-Id: Icfc61d532a3381f3b70d36d8e765f0b6548e0c3f
Add guide for debugging inter-pods communication
The networking between pods is different and separate from the
neutron network set up for the cluster. Flannel is the default
network for Kubernetes cluster and Docker is configured to use
Flannel subnet. If there is a problem, the pods will not be
able to talk to each other.
This section describes techniques for debugging Kubernetes
networking based on Flannel.
Partially implements: blueprint magnum-troubleshooting-guide
Change-Id: Ifaa71b6c8a236041b033da8ea0a8e403ff884594
Cover the scenario when nodes cannot connect to internet
The nodes need to connect to the internet for discovery and
image download. No network connection is a basic failure
and can happen at any time. This patch describes some
techniques for debugging.
If you have encountered a scenario not described here, please
mention in the review.
Partially implements: blueprint magnum-troubleshooting-guide
Change-Id: I69ca9d624b960329109fbbed8634646e63aa5118
After repeated trouble digging into gate failures,
this documentation should help a new or seasoned
Magnum developer simulate gate test failures for
troubleshooting purposes.
Partially implements: blueprint magnum-troubleshooting-guide
Change-Id: Ib015b404027af30d82f6aa2538f78cba820cbce4
Initial empty framework to fill in
Following up on the discussion at the Tokyo Summit, we are
putting together a Troubleshooting Guide for Magnum.
This is organized as a list of failure symptoms, with pointers
to techniques for troubleshooting. The initial list of
failure symptoms is from brainstorming at the Summit, but
is expected to grow as more are identified.
This initial framework is not complete by any mean, but is set up
so that multiple contributors can work in parallel.
New details can be added at any time, but current details should
be kept accurate.
Partially implements: blueprint magnum-troubleshooting-guide
Change-Id: I83b9b9ba9c76e608b4adafa525caf92aaaaaf880