Commit History for krib¶
This is the commit history for the krib component.
v4.15.0, v4.15.0-alpha01 (2025/06/12)¶
- Fix up .gitignore and .gitlab-ci
- Initial post-split commit for branches v4
v4.13.32 (2025/06/12)¶
- Fix up .gitignore and .gitlab-ci
- Initial post-split commit for branches v4.13.0-branch
v4.14.20 (2024/10/29)¶
- docs: replace old readthedocs links with rackn.io links
- fix(token): removing all usages of GenerateProfileTokens and replacing with profiles.. extra claims
- fix(params): use secure param where appropriate
- doc(fix): update links and fix render references
- doc: update provision ref and clean links
v4.12.17, v4.13.31 (2023/04/13)¶
- Add yamllint to gitlab build pipeline
- docs: rst to md conversion, parts missed by previous passes
- docs(krib): rst to md conversion
v4.10.9, v4.11.30 (2022/07/19)¶
- doc: Fix embed youtube link
- doc: Fix youtube secure references
- doc: update cluster references
v4.9.10 (2022/02/28)¶
- feat: convert meta data to a single file
v4.8.12 (2021/09/09)¶
- feat(gitlab): Migrate module to gitlab
v4.7.0 (2021/05/24)¶
- Avoid a client ca for etcd
v4.6.12 (2021/03/20)¶
- fix(doc): Various content doc fixes
v4.4.2, v4.5.5, v4.6.0-beta01 (2020/06/18)¶
- feat(agent): Add content for installing the agent.
- krib: Fix for metallb configs
- krib: move to use download() and apply proxy when necessary
- krib: docker/apply-http-proxy param to install proxy env vars
- krib: fix metallb upstream branch name
v4.3.1 (2020/04/22)¶
- docs(krib): fix duplicates and refs
- docs(krib): fix formatting and indenting
- Fix missing 'type' on param schema
- take advantage of new UX metadata field
- include feature check for ExtraClaims
- remove unsafe hack and use ExtraClaims instead. Will require 4x DRP
v4.3.0-beta2 (2020/02/21)¶
- workaround: remove helm until updated to v3
- placeholder for slowly pulling work from edgelab into release
- edge helm install runs
- rename helm/ to krib/helm-
v4.1.2, v4.2.25, v4.3.0-alpha2 (2019/10/21)¶
- for contexts, we want a finish that does not change the bootenv
v4.1.1 (2019/09/24)¶
- downgrade to 1.15.4 for better chart compatibility
- move download functions to task-library
- remove not needed serviceaccount re-applying
- lint dashboard admin cat yaml
- calico 3.9 update
- cni 0.8.2 update
- update various versions
- use download & with_backoff function for docker script
- Restart containerd if it was already running
- Override ExecStart properly
- Be idempotent
- Fix wrong param
- Damn those races
- Add option to customize loglevel
- Bump vault version
- Update consul version to 1.6.1
- Change consul agent to DEBUG mode
v4.0.15, v4.1.0-beta (2019/09/02)¶
- Delete krib.yaml (added in error)
- Delete krib.yaml (added in error)
- Ensure versioned nginx containers are used
- Fix metallb container versions to chosen version
- Obtain metallb version via param
v4.0.12 (2019/08/23)¶
- Add missing params skipped by aggressive .gitignore
- Added explicit IPs for Nginx ingress(es)
v4.0.9 (2019/08/20)¶
- Add ability to customize tcp/udp services for nginx
- Remove conntrack.max
- Update the krib to use the parameters from dev-library
- Safer reset
v4.0.8 (2019/08/15)¶
- Increase backoff to avoid races
- Allow fixing rook-ceph version
- Enforce v.1.0.0 for now
- crictl version 1.15 matching k8s 1.15
- re-add cleanup without heapster
- k8s 1.15.2
- kubernetes dashboard v2
v4.0.7 (2019/08/13)¶
- Protect mounted filesystems
- Improve ceph resetting
- Permit reset with custom user/password
- As you weeeeesh
- Standardize Ceph manifests to 1.0
v4.0.6 (2019/08/10)¶
- Fix doc links
v4.0.4 (2019/08/08)¶
- Set sane log level
v4.0.2 (2019/08/06)¶
- Start containerd on boot
- don't hardcode the profile brah
- as per @galthaus, use krib/k3s param. this is basic level for now but good pattern
- as per request by @galthaus
- cert plugin does not do secure yet
- move to secure data - most are not passwords. REQUIRES TIP DRP
- missed if
- minor docs
- remove krib/k3s param
- undo the secure/password mess - it's not good but needs a real test pass to fix
- more delay, zzzz
- k3s multi-node working
- containerd install fixes
- add containerd install
- WIP k3s config process
- respect the k3s param
- improve log messages
- cert plugin does not do secure yet
- include feature flag to catch Profile Secure Get/Update requirement
- move to secure data - most are not passwords. REQUIRES TIP DRP
- add workflow, stage includes special k3s param
- WIP krib-get-masters failing
- Another place to handle ha
- Adjust the vip to also watch for master-count > 1
- Fix parameter references to correctly match protections.
- Remove some dead spaces
- Add ceph/ceph to private registry
- Update for storing calico images in trusted registry
- Improve regex
- cert plugin does not do secure yet
- include feature flag to catch Profile Secure Get/Update requirement
- move to secure data - most are not passwords. REQUIRES TIP DRP
- add workflow, stage includes special k3s param
- Un-dumbassed the sed (only myself to blame)
- Make ceph default storageclass
- Added further add-on containers
- Progress pinning containers
- WIP krib-get-masters failing
- Add option to specify metallb container image
- !#/bin/bash (my head in)
- Avoid races on vault
- Prevent races
- More backoffs on nginx
- Tired of re-trying nginx stage
- Implemented (aws only for now)
- Update nginx deployment for mutually exclusive options
- Tidy up templates
- Whitespace fixed
- Standardize param names
- Parametize target disk
- Initial addition of option
- Add default issuer to ceph dashboard cert if configured
- Fix errant quote
- Improve calico behavior in multi-NIC deployments
- Latest progress
- Streamlined installation
- Merged with logging branch
- Further genericized
- Add params to genericize graylogging
- Remove local rook files on reset
- Added fluent-bit to DRP
- Zap worker disks (hardcoded to /dev/sda currently)
- Progress thus far - no drives working yet
- fail gracefully if namespace already exsits
- By default, avoid buggy IPs
- Front proxy certs must match on all masters
- Slightly less sleepy
- add ability to ignore preflight errors, streamline docs
- Fix nginx to v0.24.1
- Updated for signed kubelets
- Add signed kubelet config
- Restore validating webhook
- Make cleanup.. cleaner
- Correct name of AWS secret key param
- Working ingress, without admission controller for now
- Tidy up reset to avoid races
- Ingress working, certs unsure yet
- Working with two ingresses, no certs yet
- Store consul client certs to facilitate backup
- Generate clients certs to backup etcd externally
- Add delay before testing vault status to allow for startup
- Generate a client cert for controller
- Reduce likelihood of races to init vault
- Tidy up certs
- Got automated install with vault working
- Tidied up some issues discovered on reinstall
- Support vault token as param in DRP
- Add support for auto-unsealing with awskms
- Add (untested) KMS auto-unseal
- Add TLS to vault
- Tweak client params
- Works co-locating agent and server on the same host
- Initial version of vault
- Make consul deployment more user-friendly
- Add gossip encryption
- Split consul into install and configure steps
- Make consul work with DRP-generated certs
- Self-assemble TLS-secured consul cluster (WIP)
- Improve consul workflow
- Prepare params
- Progressed up to a point - left notes to followup
- Got initial consul cluster built (the dirty way)
- Installed /usr/local/bin/consul
- First cut at consul setup
- Improve presentation and docs of krib-container-runtime.yaml
- Add option to install files from local repo
- Fix test for local repo
- Add details on switching to containerd
- Implement containerd, fix bugs
- Make install-containerd agnostic to a local http repo
- Update for generic runtime stage
- Applies, waiting validation
- First cut (is the deepest!)
- Correctly use HA cluster endpoint
- Increase random sleep from 5 to 25 sec
- Decomposed further
- Safely introduce optional params for selective mastership
- Only use HA etcd master if necessary params are defined
- Prevent failure if krib/cluster-master-vip is undefined
- Add brutal machine-wiping script
- Add etcd/cluster-client-vip-port as optional param
- add prereqs meta to help users know certs is required
- Update kubeadm.conf
- Implement load-balanced etcd
- update kubernetes to v1.14.3
- fix cni plugin link
- update kubernetes to v1.14.2, cni to 0.8.0
- update docker, etcd, cni, crictl, kubernetes
- Make sure docker config dir exists before trying to create daemon.json
- update calico to version 3.6
- use -o since -O is used to set remote name as file name
- Update example-krib-ha.yaml
- Add prerequisites to community content bundles
- Update mount-disks.sh.tmpl
- Adding comment to add unsafe/rs-password
- Adding comment after installing helm
- Change Option -o to -O on Line 200
- Rename "Name: example-krib-ha" to "Name: krib-ha"
- Update mount-disks.sh.tmpl
- Adding -f option on mkfs.xfs command
- adding function to umount when the /dev/sda1
- updated metallb to be re-runnable. Refs digitalrebar/provision-content/issues/243
- add krib apiserver extra SANs param
- updated helm install to support master recovery
- added basic support for self healing etcd cluster
- Get profile data from profile and not aggregate through machine.
- fix CertSANs -> certSANS; add static pod path for all
- fix certSANs indent
- turn off debug messages
- remove unneeded extra volumes and static pod path
- add missing cluster profile data for krib-lib
- source krib-lib
- move with_backoff and download function to krib-lib and use for etcd
- tweaks for catalog generation
- remove extra volumes mount of hyperkube
- load kernel module br_netfilter if already not loaded
- check if /etc/fstab exists before remove swap entry from it
- unify wording
- no need for full git repo; remove heapster temp dir
- remove debug output
- update config for v1.13
- add backoff&download to handle bad connections; increased verbosity
- update defaults:
- Add full content package info
- fix etcd-config call to non-existent xiterr()
- fix etcd-config call to non-existent xiterr()
- added a check in ingress to not create the ingress for the dashboard if it's disabled
- added weave networking provider
- fixed unworkable flannel config option
- set the icon in all cases
- bad syntax
- added param to allow disabling k8s dashboard install
- adding line breaks inside ranges so lines don't get run together
- turn off swap permanently
- --sort-backends no longer being accepted on current ingress-nginx release
- added param to configure additional kubeadm apiServerExtraArgs
- Allow etcd and krib IP's to be overridden
- change pocket to docker icon
- Fix provider/flannel-config URL
- use known icon
- adding more flexibility for templatebefore and templateafer
- Updates to repos support for helm charts
- Changed to apply
- Added ceph dashboard login info to docs
- Added documentation and examples
- Updated rook to add ceph dashboard service + ingress and update default path
- use more reliable kubectl apply instead of create
- use master label for taint
- add IFS to while loop
- disable debug for krib keepalived
- ensure taints for masters
- use rook version 0.9.1
- add rook/ceph example
- allow multiple mon for small setups
- update rook (0.9) and ceph (13.2.2-20181023)
- added krib rancher block storage
- Accidentally removed one too many lines.
- typo for tls secret name
- added ingress name to default $CLUSTERIP.xip.io hostname so multiple ingresses can be defined
- removed NodePort addition to kubernetes-dashboard service since it didn't work for non-proxy access anyhow
- moved dashboard ingress template inline in the shell script to prevent inadvertent reuse
- Added krib nginx-ingress
- Fixed heapster unable to talk to API
- formatting cleanup
- tabs to spaces
- added param and ability to remove taints from master nodes
- Updated docs
- krib/cluster-master-vip shouldn't be cleaned up since it's manually set
- Removed bad doc
- Added krib-metallb
- update etcd version to 3.3.10
- fix calico config url
- remove calicoctl
- update doc for calicoctl
- kubernetes version 1.12.4 didn't work for me
-
- use k8s datastore instead of calico etcd
- add krib/cluster-pod-subnet to calico
- Expanded on accessing the kubernetes dashboard
- fix HOSTNAME for worker nodes
-
- add support for custom calico etcd clusterIP
- fix typo
- update kubernetes version to v1.12.4
- update cni to v0.7.1
- add custom cluster ip for calico etcd
- allow label overwrite
- update to calico 3.4; own etcd config
- update dashboard link
- Fix centos7 update breaking krib.
- fix multi-arch
- cleanup and add multiarch back
- as per @faniand, remove unneeded packages. host check is optional
- working first pass for KubeVirt.io stage
- Make sure socat is installed.
- Add docker specific version.
- Deal with docker issues
- Update krib and friends to work with CoreOS
- Fix krib in tip
- version in community changed...
- skip daemon if no param defined
- remove insecure default
- undo typo in config
- use daemon config instead of contrail reg hack
- change icons
- move docker daemon change to docker install
- tweaks from debug
- update for new release
- cleanups from testing
- remove helm, put contrail install in right place. NOT TESTED
- WIP contrail bones for basic install
- Changes to support curl with ipv6 addresses.
- rook install needs helm AND templates - this is a powerful addition
- WIP Rook Ceph - add repo to helm
- Add 'ntpdate' binary check, install if missing
- update docs - see matching provision patch
- add some logging
- include wait option for charts
- helm advanced is ONLY way, move to helm/charts
- good working point for helm install updates
- move name and namespace to top level
- fix yaml parsing error
- add targz suppport
- WIP advanced helm install builder working
- change approach to simpler charts-advanced array instead
- initial helm directives object
- initial helm directives object
- WIP
- fix extra newline in etcd endpoints list iteration
- Desensitize check for 'Ready' status
- move label code into place where it works for all nodes
- pull node labels from krib/labels param
- add require features flags to content
- adds 'delete' and 'cordon' Operations to KRIB
- timing tweaks, default to no wait
- initial sonobuoy support - not automatically included
- label nodes - use inventory if available
- add 'krib/cluster-masters' to example profile
- add override for kubeadm.cfg
- fixes,enhancements, testing
- move set-networking-provider to networking-provider
- Fix spaces at end-of-lines
- Fix bad template syntax
- updates to support v1alpha2 for etcd, etc
- updates, still broken - etcd cluster conflict
- WIP: updates to KRIB
- add some docs
- add security policy and repos list
- fix bug in reset
- helm stage
- add password as param
- use rocketskates auth to reset cert in dev reset
- move doc from provision to krib component
- testing terraform, small tweaks
- update for new meta UX field name
- add wait stage to reset
- add compute nodes wait for controller loop
- WIP initial files checkin - begin testing
- Installing wget since it's not included in the base CentOS install
- Even better solution that is permanent.
- Fix #106 by forcing 1 into /proc/sys/net/bridge/bridge-nf-call-iptables
- remove set -x, add doc line
- detect and halt if CA root is present
- add needed backoff for reset, include render
- missed exit
- remove plugin:
- move cluster reset into its own stage
- troubleshooting and docs tweaks
- Updated etcd version to match kubernetes 1.11
- kubeadm installs coredns instead of kube-dns
- fix etcd version
- adding more echos for troubleshooting
- add documentation to params
- WIP for building KRIB clusters
- Fix case for wait_for_count returning something.
- Fix templates again
- Fix CLUSTER PROFILE here too
- Same thing with PROFILE_CLUSTER var
- Make sure profile token is set before loading library.
- Fix spaces at EOL
- Add KRIB to community content.
- Initial commit