Automatically provision and manage TLS certificates in Kubernetes
flant/k8s-image-availability-exporter 87
Alert if an image used in Kubernetes cannot be pulled from container registry
A Container Storage Interface (CSI) Driver for Yandex.Cloud Compute Disks
flant/kube-resource-unit-converter 2
Convert Kubernetes Resource Quantity representation to bytes
Autoscaling components for Kubernetes
CSI driver for Amazon EBS https://aws.amazon.com/ebs/
pull request commentyandex-cloud/terraform-provider-yandex
Allow empty "static_route" Set
@GennadySpb Done.
comment created time in 21 hours
Pull request review commentyandex-cloud/terraform-provider-yandex
Allow empty "static_route" Set
func routeDescriptionToStaticRoute(v interface{}) (*vpc.StaticRoute, error) { sr.Destination = &vpc.StaticRoute_DestinationPrefix{ DestinationPrefix: v, }+ } else {+ return nil, errors.New("\"static_route\" should have a \"destination_prefix\" field") } if v, ok := res["next_hop_address"].(string); ok { sr.NextHop = &vpc.StaticRoute_NextHopAddress{ NextHopAddress: v, }+ } else {+ return nil, errors.New("\"static_route\" should have a \"next_hop_address\" field")
Done.
comment created time in 21 hours
push eventflant/terraform-provider-yandex
commit sha 6625dc6b5ab89fbc021542e154e6d8b289d85b32
Allow empty "static_route" Set Signed-off-by: Andrey Klimentyev <[email protected]>
push time in 21 hours
push eventflant/terraform-provider-yandex
commit sha b46be90d12ef3b875808611477b89cf56e2a8cb8
object storage sse support
commit sha 5cdd71164b9a7202e474cf57cdc1082c71614052
object storage sse documentation
commit sha b33040969ad8f3f4028ce01af7c4e89ec24da79d
Merge pull request #176 in CLOUD/terraform-provider-yandex-mirror from Security_Group_Sweep to master Squashed commit of the following: commit 46376374bb857e145e226b4e49b2518d0c5107df Author: Timur Girevoy <[email protected]> Date: Wed Jul 22 17:39:39 2020 +0300 Add Security Group Sweeper
commit sha ea7c88e01fa5e3f8e07ee8cebb439c891487e201
fixed storage SSE docs
commit sha c406e430b76792e460581b019258a9b985e59ae9
fixed docs for Storage SSE
commit sha 92370af5d124de6a7b82918f11bb7a8e37f92ea9
Merge pull request #180 in CLOUD/terraform-provider-yandex-mirror from CLOUD-53362 to master Squashed commit of the following: commit b8c8bd200ee12b8f8ad6a0021c80139061ae0e9b Author: Alexander Lavrukov <[email protected]> Date: Fri Sep 11 21:14:11 2020 +0300 CLOUD-53362: new security group proto spec
commit sha ff39240cf760fc1f92f132c6ba4658e75e9e9e8f
Merge pull request #177 in CLOUD/terraform-provider-yandex-mirror from fix/CLOUD-43368-lb-listener-address-as-set to master Squashed commit of the following: commit 1f5c49061072981d8d57aa3598a8e91e197de81d Author: Alexander Ovechkin <[email protected]> Date: Tue Sep 15 15:47:13 2020 +0300 CLOUD-43368: added checkWithState commit 69fa8890262c0f82be769a0c8205e56394006089 Author: Alexander Ovechkin <[email protected]> Date: Mon Sep 14 03:04:26 2020 +0300 CLOUD-43368: test fixes commit 4c7701e63a8f2304197d4f9ba677218e8811c188 Author: Alexander Ovechkin <[email protected]> Date: Mon Sep 14 02:08:11 2020 +0300 CLOUD-43368: removed tmp debug logging commit 3ddcfa26d9acb6fbe8634ab761fec9cc1e0092f6 Author: Alexander Ovechkin <[email protected]> Date: Fri Sep 11 17:13:38 2020 +0300 CLOUD-43368: using set for nlb listener representation in lb datasource commit db181da50604ab1993deca76b33968434fd6a702 Author: Alexander Ovechkin <[email protected]> Date: Fri Sep 11 02:02:29 2020 +0300 CLOUD-43368: * fixed listener hash function * representing external/internal addresses as set instead of list
commit sha f7473943072a34a1480c76111383f24f8806c693
Merge branch 'master' into master
commit sha e946199b9ef934b819c7670098bb7340328deb7b
Merge pull request #106 from tna0y/master Storage: server side encryption
commit sha 84c11d9046fd23612a691da37c1b27fa114ec568
Doc fix
commit sha 8026d9179717355bfffa40f6d23ed9573796936b
Merge pull request #182 in CLOUD/terraform-provider-yandex-mirror from new-sg-rule-fields to master Squashed commit of the following: commit e78d008f40d82b9e29df68d4ad5503d168024dd6 Author: Alexander Lavrukov <[email protected]> Date: Fri Sep 11 21:48:49 2020 +0300 security_group_id and predefined_target fields were added to the security group rule spec
commit sha 48385be5f9bc29f156a41b41882956cd8d1bb7f8
fix urls && update go sdk
commit sha 8dc6b00ebc3d7292e94fb3b811ae8d225bccbb07
Merge pull request #107 from yandex-cloud/feature/no-ticket/update-go-sdk fix urls && update go sdk
commit sha 7763b245f30281973b0e220bb30c79d31183cb98
fix linter errors (#108)
commit sha c82cba1dd0f77ae4a9d6ddf22370a409efa47a59
redis: fix example. config version is required (#109)
commit sha e0fe4a6b06d00d68b036aa16028a0650bbee4a6f
Allow empty "static_route" set
push time in 21 hours
pull request commentgardener/machine-controller-manager
@hardikdr That's what I get for trusting Goland IDE too much. Fixed!
comment created time in 21 hours
push eventflant/machine-controller-manager
commit sha 6cc726ebda7b9caa30e3784a44dde4a722941d5f
[azure] Verify tags Signed-off-by: Andrey Klimentyev <[email protected]>
push time in 21 hours
push eventflant/machine-controller-manager-provider-yandex
commit sha 4e9f31264decb60e46ba8acc6850386814b45746
Declare license and copyright information using REUSE We now use https://reuse.software/ to declare license and copyright information in a machine readable format. The copyright statement for SAP open source projects was updated to give credit to project contributors not employed by SAP.
commit sha ce9b866df05d9764df21106027f0b50d4186f3c3
Merge pull request #1 from msohn/reuse Declare license and copyright information using REUSE
push time in a day
PR merged flant/machine-controller-manager-provider-yandex
We now use https://reuse.software/ to declare license and copyright information in a machine readable format.
The copyright statement for SAP open source projects was updated to give credit to project contributors not employed by SAP.
What this PR does / why we need it: Adopt new rules set by the SAP Open Source program office.
Which issue(s) this PR fixes: Fixes #
Special notes for your reviewer:
Release note: <!-- Write your release note:
- Enter your release note in the below block.
- If no release note is required, just write "NONE" within the block.
Format of block header: <category> <target_group> Possible values:
- category: improvement|noteworthy
- target_group: user|operator|developer -->
License and copyright information is now specified in REUSE format.
pr closed time in a day
issue openedgardener/machine-controller-manager
Provide a way to initialize out-of-tree providers
What would you like to be added:
I'd like to have a special gRPC method that is called exactly once on the start-up of MCM and subsequent initialization of out-of-tree providers.
Here's how the Cloud Provider Interface approaches this: https://github.com/kubernetes/cloud-provider/blob/master/cloud.go#L44-L47
Why is this needed:
For example, while using an in-house in-tree vSphere plugin, we've hit a limit on simultaneous login attempts. An Active Directory server was under a lot of duress before I've implemented a persistent storage session storage.
Here's how I've hacked around this problem to enable persistent sessions:
- Implementation: https://github.com/flant/machine-controller-manager-provider-yandex/blob/50b1fdd35f55a09d1259f3a88b08630acdc7a02b/pkg/spi/spi.go
- Usage: https://github.com/flant/machine-controller-manager-provider-yandex/blob/50b1fdd35f55a09d1259f3a88b08630acdc7a02b/pkg/provider/core.go#L94
Basically, we add or get a session from a map. Ugly, but works.
created time in a day
Pull request review commentgardener/machine-controller-manager
func (clients *azureDriverClients) getRelevantVMs(ctx context.Context, machineID if len(machines) > 0 { for _, server := range machines {+ if !verifyAzureTags(server.Tags, searchClusterName, searchNodeRole) {+ continue
Done.
comment created time in 2 days
push eventflant/machine-controller-manager
commit sha 41a5b9063fb9e20947516cb8fce8b682afa58218
[azure] Verify tags Signed-off-by: Andrey Klimentyev <[email protected]>
push time in 2 days
create barnchflant/machine-controller-manager-provider-yandex
created branch time in 2 days
Pull request review commentgardener/machine-controller-manager
func (clients *azureDriverClients) getRelevantVMs(ctx context.Context, machineID if len(machines) > 0 { for _, server := range machines {+ if !verifyAzureTags(server.Tags, searchClusterName, searchNodeRole) {+ continue
V(4)?
comment created time in 2 days
push eventflant/yandex-cloud-controller-manager
commit sha a4e6338df13892028c0c14cc0beed3c66f08a9ed
Create dependabot.yml
push time in 3 days
delete branch flant/yandex-cloud-controller-manager
delete branch : major-rewrite
delete time in 3 days
push eventflant/yandex-cloud-controller-manager
commit sha f12ad92772010d67aa00a634ecfe8129b620419b
Allow easier testing by refactoring interactions with Yandex.Cloud go-sdk #15
push time in 3 days
PR merged flant/yandex-cloud-controller-manager
pr closed time in 3 days
push eventflant/yandex-cloud-controller-manager
commit sha 8609c0bf48588f6aed1746ae19c40e36cc161bef
Major rewrite Signed-off-by: Andrey Klimentyev <[email protected]>
push time in 3 days
PR opened flant/yandex-cloud-controller-manager
pr created time in 3 days
create barnchflant/yandex-cloud-controller-manager
created branch time in 3 days
issue closedflant/yandex-cloud-controller-manager
Hello :wave:
I was wondering what the status of the CCM is : is this sanctionned at all by Yandex Cloud ? Are there any plans, or have there been any attempts to work with Yandex to collaborate on this ?
Since Yandex has a managed kubernetes product now I imagine that they have a very similar project somewhere.
closed time in 3 days
fcantournet
delete branch flant/yandex-cloud-controller-manager
delete branch : zuzzas-patch-1
delete time in 6 days
push eventflant/yandex-cloud-controller-manager
commit sha d9d2aca2ae48c17ecbd615164e19a8edc41dc377
Mention RouteController implementation in README.md
commit sha e8bd8fd39b98363d693023a1205a41ce81d694c8
Merge pull request #14 from flant/zuzzas-patch-1 Mention RouteController implementation in README.md
push time in 6 days
PR merged flant/yandex-cloud-controller-manager
pr closed time in 6 days
PR opened flant/yandex-cloud-controller-manager
pr created time in 6 days
create barnchflant/yandex-cloud-controller-manager
created branch time in 6 days
issue openedflant/yandex-cloud-controller-manager
Remove TargetGroups when last LoadBalancer Service is deleted.
created time in 8 days
issue commentflant/yandex-cloud-controller-manager
Hi!
It's not sanctioned by the Yandex.Cloud, but was once acknowledged by them in the official Yandex.Cloud's Telegram channel. There was no news from the Yandex.Cloud team about open-sourcing their CPI and CSI implementations. That's why we have to work from the outside.
comment created time in 8 days
pull request commentgardener/machine-controller-manager
Back off failed Machine operations
@hardikdr Some preliminary work on throttling Machine operations. Would like to hear opinions regarding the proposed approach.
comment created time in 9 days
PR opened gardener/machine-controller-manager
What this PR does / why we need it:
This PR retries failed Machine operations with exponential back off.
Which issue(s) this PR fixes: Fixes #483
Special notes for your reviewer:
I've keyed back off to MachineSet since a group of identical Machines should have the same error cause.
Some things still require ironing out before moving from a Draft state.
- [ ] Unit tests for backoff_manager
- [ ] Consider whether we'd like to back off on every error or just on errors returned by cloud APIs ones. If only from cloud APIs, wrap a custom error from Driver. I can see some problems with frozen MachineSets already.
Release note: <!-- Write your release note:
- Enter your release note in the below block.
- If no release note is required, just write "NONE" within the block.
Format of block header: <category> <target_group> Possible values:
- category: improvement|noteworthy
- target_group: user|operator|developer -->
Failed Machine operations will exponentially backoff before being executed again
pr created time in 9 days
push eventflant/machine-controller-manager
commit sha 365ed671ad1abbe975bc4a79961a9e2cb364b3b7
Backoff failed Machine operations Signed-off-by: Andrey Klimentyev <[email protected]>
push time in 9 days
push eventflant/machine-controller-manager
commit sha 92c2be42cdf34621730e6e8365b91edbe2e23621
Backoff failed Machine operations Signed-off-by: Andrey Klimentyev <[email protected]>
push time in 9 days
issue commentgardener/machine-controller-manager
Adopt Cluster / Machine API Spec
Here's a convo from the last year. @vlerenc sheds some light on the issue.
https://kubernetes.slack.com/archives/CB57N0BFG/p1573835150176400
comment created time in 9 days
pull request commentgardener/machine-controller-manager
@prashanth26 We've tested the proposed filtering on existing Azure clusters. Perhaps, it can have some adverse effects on large installations, but we haven't found any bugs yet.
comment created time in 9 days
Pull request review commentgardener/machine-controller-manager
An option to throttle deletion of Failed Machines
func (c *controller) manageReplicas(allMachines []*v1alpha1.Machine, machineSet return nil } - var activeMachines, staleMachines []*v1alpha1.Machine+ var activeMachines, runningMachines, staleMachines []*v1alpha1.Machine for _, machine := range allMachines { if IsMachineActive(machine) {- //klog.Info("Active machine: ", machine.Name) activeMachines = append(activeMachines, machine)++ if IsMachineRunning(machine) {+ runningMachines = append(runningMachines, machine)+ } } else if IsMachineFailed(machine) { staleMachines = append(staleMachines, machine) } } + machineDeletionWindow := c.calculateStaleMachineDeletionWindow(activeMachines, runningMachines, staleMachines, machineSet)+ if len(staleMachines) >= 1 { klog.V(2).Infof("Deleting stale machines") }- c.terminateMachines(staleMachines, machineSet)+ deletedStaleMachines, _ := c.terminateStaleMachines(staleMachines, machineDeletionWindow, machineSet)+ staleMachinesLeft := len(staleMachines) - deletedStaleMachines diff := len(activeMachines) - int(machineSet.Spec.Replicas)+ if staleMachinesLeft > 0 {+ // Count the leftover stale machines against .spec.replicas to prevent sudden surge in machines which can happen if large number of stale machines are left over+ diff -= staleMachinesLeft
Taken care of.
comment created time in 12 days
Pull request review commentgardener/machine-controller-manager
An option to throttle deletion of Failed Machines
func (c *controller) manageReplicas(allMachines []*v1alpha1.Machine, machineSet return nil } - var activeMachines, staleMachines []*v1alpha1.Machine+ var activeMachines, runningMachines, staleMachines []*v1alpha1.Machine for _, machine := range allMachines { if IsMachineActive(machine) {- //klog.Info("Active machine: ", machine.Name) activeMachines = append(activeMachines, machine)++ if IsMachineRunning(machine) {+ runningMachines = append(runningMachines, machine)+ } } else if IsMachineFailed(machine) { staleMachines = append(staleMachines, machine) } } + machineDeletionWindow := c.calculateStaleMachineDeletionWindow(activeMachines, runningMachines, staleMachines, machineSet)+ if len(staleMachines) >= 1 { klog.V(2).Infof("Deleting stale machines") }- c.terminateMachines(staleMachines, machineSet)+ deletedStaleMachines, _ := c.terminateStaleMachines(staleMachines, machineDeletionWindow, machineSet)+ staleMachinesLeft := len(staleMachines) - deletedStaleMachines diff := len(activeMachines) - int(machineSet.Spec.Replicas)+ if staleMachinesLeft > 0 {+ // Count the leftover stale machines against .spec.replicas to prevent sudden surge in machines which can happen if large number of stale machines are left over+ diff -= staleMachinesLeft
But that won't work, because it'll panic in getMachinesToDelete() since diff will be bigger than activeMachines count.
It's my fault for moving this calculation outside of the if branch. This condition should suffice.
comment created time in 12 days
push eventflant/machine-controller-manager
commit sha d7e72b39ab4de12901d4dbf86a3886dcef6702f7
More changes Signed-off-by: Andrey Klimentyev <[email protected]>
push time in 12 days
push eventflant/machine-controller-manager
commit sha f7ea2174c419b2d76dcb0b1745f0a35c2f6dd149
Apply suggestions from code review Co-authored-by: Gaurav Gupta <[email protected]>
push time in 12 days
pull request commentcloudflare/ebpf_exporter
add kubernetes context decoder
@huaizong Would you mind re-opening the PR and restoring the branch?
The most glaring thing is support for Docker daemon and not other CRI implementations, but that can be easily rectified later. There is also the issue of using undocumented Docker container labels that are used internally by the kubelet, but this interface seems stable enough. Almost no changes in years: https://github.com/kubernetes/kubernetes/blob/4fec22b3e4da6034948fd684f60637171cb55561/pkg/kubelet/types/labels.go
comment created time in 14 days
Pull request review commentyandex-cloud/terraform-provider-yandex
Allow empty "static_route" Set
func flattenStaticRoutes(routeTable *vpc.RouteTable) *schema.Set { return staticRoutes } -func expandStaticRoutes(d *schema.ResourceData) ([]*vpc.StaticRoute, error) {+func expandStaticRoutes(v interface{}) ([]*vpc.StaticRoute, error) { staticRoutes := []*vpc.StaticRoute{} - if v, ok := d.GetOk("static_route"); ok {- routeList := v.(*schema.Set).List()- for _, v := range routeList {- sr, err := routeDescriptionToStaticRoute(v)- if err != nil {- return nil, fmt.Errorf("fail convert static route: %s", err)- }- staticRoutes = append(staticRoutes, sr)+ if v == nil {
I've found a proper way to do this. https://github.com/yandex-cloud/terraform-provider-yandex/pull/105/commits/850383b92ed92e15cbb15770eab85de4bc3e5775
comment created time in 15 days
push eventflant/terraform-provider-yandex
commit sha 850383b92ed92e15cbb15770eab85de4bc3e5775
Detect empty set element
push time in 15 days
Pull request review commentyandex-cloud/terraform-provider-yandex
Allow empty "static_route" Set
func flattenStaticRoutes(routeTable *vpc.RouteTable) *schema.Set { return staticRoutes } -func expandStaticRoutes(d *schema.ResourceData) ([]*vpc.StaticRoute, error) {+func expandStaticRoutes(v interface{}) ([]*vpc.StaticRoute, error) { staticRoutes := []*vpc.StaticRoute{} - if v, ok := d.GetOk("static_route"); ok {- routeList := v.(*schema.Set).List()- for _, v := range routeList {- sr, err := routeDescriptionToStaticRoute(v)- if err != nil {- return nil, fmt.Errorf("fail convert static route: %s", err)- }- staticRoutes = append(staticRoutes, sr)+ if v == nil {
It's not (easily) possible to verify the set's emptiness, so we check for nil. It won't catch the following set with an empty element.
static_route {}
comment created time in 15 days
PR opened yandex-cloud/terraform-provider-yandex
The "static_route" block is marked as optional. Unfortunately, a function that converts a set into a gRPC struct is not aware of that fact.
pr created time in 15 days
create barnchflant/terraform-provider-yandex
branch : fix-route-table-optionality
created branch time in 15 days
startedcilium/cilium
started time in 15 days
PR opened gardener/machine-controller-manager
What this PR does / why we need it:
Tag verification has been disabled back in May of 2019 by @prashanth26.
We can deduce that it was done due to Azure's unique ability to group resources in Resource Groups. Gardener project, presumably, packs all VMs (and relevant objects) into a single Resource Group (hence no need for tag verification).
That does not work for us, since we deploy master VMs near VMs that are managed by MCM and backed by Machine objects. It also seems weird that there are tags present in AzureMachineClass CRD, but are not used to verify actual objects.
Special notes for your reviewer:
Release note: <!-- Write your release note:
- Enter your release note in the below block.
- If no release note is required, just write "NONE" within the block.
Format of block header: <category> <target_group> Possible values:
- category: improvement|noteworthy
- target_group: user|operator|developer -->
Restored tag verification in the Azure driver
pr created time in 16 days
Pull request review commentgardener/machine-controller-manager
An option to throttle deletion of Failed Machines
func (c *controller) manageReplicas(allMachines []*v1alpha1.Machine, machineSet } diff *= -1+ if diff > BurstReplicas { diff = BurstReplicas }++ // throttle machine creation until all stale machines are deleted+ if len(staleMachines) > 0 && diff > machineDeletionWindow {+ diff = machineDeletionWindow+ }
It is done. https://github.com/gardener/machine-controller-manager/pull/482/files#diff-16f7b87f40829e0c81cb6c486f3997ccR345-R359
comment created time in 16 days
push eventflant/machine-controller-manager
commit sha d1a9293ef92bbb8a01ddfcb6ff3c2974dff51488
Return error if the secret referred by machine class is absent
commit sha 56e6831e0e89621c42391aba392b8c9fe3644de6
Merge pull request #495 from AxiomSamarth/samarth/MI-430-backoff-missing-secret Retry when secret is referred by machineClass is missing
commit sha 2034e0d601c67a858a657bd8660faa88ebdd6c51
Add termionation condition to terminating nodes
commit sha 3def9452ec9b1796e5f2171616abc31a0ce7f84e
Merge pull request #492 from guydaichs/feature/annotate-on-termination Add termination condition to deleting nodes
commit sha 82a9b49a4b317c2e18998a467ffd81c918242ad8
Limiting the random hash in the machine set to 5 characters
commit sha e16e726ce2e9ab23734ea8c8dadd8b98396dabe5
Merge pull request #500 from AxiomSamarth/samarth/MI-461-limit-random-hash-machineSet-name Limit the Random Hash in the MachineSet Name
commit sha b9cadf53f8fcc07313294b3267e3c0b455a32450
Added MigrateMachineClass interface to OOT - Any changes/updates to class.Kind propagates top-down without rolling updates - Ignore reconciling machines with unknown ClassKind at MCM - Updated docs to reflect these changes - Updated old machine class finalizers - Introduced a new flag to delete old machine classes on migration
commit sha 0a568917461a5c6385974322784b36d62762e60a
Updated finailizer deletion logic for MachineClasses - The machineClass controllers add finalizer only when reference is present, and removes it otherwise - There is also logic for setting the deletion timestamp if migration annotation and delete-migrated-machine-class flag is set
commit sha 90f8b67cc5e643feabf99ba6b8193aff47e04e79
Merge pull request #484 from prashanth26/migration Add migration logic for moving from in-tree to out-of-tree
commit sha db3e5649bbbfc0803649fdf781d47b8a2bc7f43a
Add cluster autoscaler scale-down disabled annotation on the nodes while rolling-update
commit sha 1f45f2ae34955f846363c6e024e1de89e76b8833
Add autoscaler scale-down annotation during Recreate update
commit sha f5a0478b4ce9ebbdf4d800e8ea47ee616605938a
Make annotating nodes during update configurable and add unit-tests
commit sha 91c52d568862a8c3d7425eddbed23543117dc239
Merge pull request #496 from hardikdr/feature/autoscaler-annotation Inhibit scale-down by autoscaler during roll-outs.
commit sha b178c45c06ff28487208819dbfb73d8720a651a6
Smart deletion of Failed Machines Signed-off-by: Andrey Klimentyev <[email protected]>
commit sha 5356550fb830e00504997e8d08f7d943c9a6e148
1. Calculate ratio relative to MachineSet's "spec.replicas" 2. Take MachineSet's scale-up into account Signed-off-by: Andrey Klimentyev <[email protected]>
commit sha acc8d94c347747008a7acdeff7fc07a6a73224c4
Improvements based on GitHub reviews Signed-off-by: Andrey Klimentyev <[email protected]>
commit sha 4f02715dd6d1dad36f9b39a08411de744864d8df
Clamp stale Machine deletion to machineDeletionWindow Signed-off-by: Andrey Klimentyev <[email protected]>
commit sha 8b185adbdd7850c3a157e3be8bd9c40d4e596729
Clamp stale Machine creation AND deletion to machineDeletionWindow Signed-off-by: Andrey Klimentyev <[email protected]>
commit sha e24e0a033587548a9cd56b4dfb023ccf63e2f1dc
Latest fixes Signed-off-by: Andrey Klimentyev <[email protected]>
push time in 16 days
startedhexdigest/gowrap
started time in 24 days
issue commentcloudflare/ebpf_exporter
add kubernetes namespace, container name decoder
@bobrik Would you be interested in accepting something like this into the project?
There was an attempt by @huaizong, but it went unnoticed. Perhaps, you'd like to discuss some other format of a decoder?
comment created time in a month
issue commentgardener/machine-controller-manager
Make src/dst checks configurable on awsmachineclass
@zanetworker That is correct.
comment created time in a month
issue commentgardener/machine-controller-manager
Make src/dst checks configurable on awsmachineclass
If you use AWS cloud-controller-manager (separately or as a legacy part of kube-controller-manager), it'll set this option on every Node. That's why we aren't interested in this feature being part of machine-controller-manager.
comment created time in a month
push eventflant/examples
commit sha cd32fbc2c8c4b55559f12c721a035e298d1a87fa
KubeCon fixes (#11)
push time in a month
push eventflant/examples
commit sha 5b7510e18e54ce3885e0ff83c6ab8bcc2b381668
Kubecon 2020 (#10)
commit sha eda829f73564ba0c7e96d238b10d24f61eb950df
KubeCon fixes
push time in a month
push eventzuzzas/shell-operator
commit sha 317e68fffaf41b9b0a9187a2e69723e21f07144a
+
push time in a month
push eventflant/examples
commit sha 5b7510e18e54ce3885e0ff83c6ab8bcc2b381668
Kubecon 2020 (#10)
push time in a month
push eventflant/examples
commit sha 601f4523c4cf9c561ecd2d988326ffb4a04f30a7
Update Dockerfile
push time in a month
push eventzuzzas/shell-operator
commit sha 97112332b831e1fbef3e00b918b14aaeff2ddbcf
Add a sourceable shell_lib.sh
push time in a month
PR opened flant/shell-operator
Perhaps, there could be a better place for this, alas, its location is already a part of the KubeCon 2020 video.
pr created time in a month
push eventzuzzas/shell-operator
commit sha 2a639883a0dd23586cc56de0a89b6d0c6a3fc952
+
push time in a month
issue openedflant/yandex-csi-driver
ControllerUnpublishVolume logic error
Should stick closer to the spec and query attachment information directly from the Disk object.
created time in a month
Pull request review commentgardener/machine-controller-manager
An option to throttle deletion of Failed Machines
func (c *controller) manageReplicas(allMachines []*v1alpha1.Machine, machineSet } diff *= -1+ if diff > BurstReplicas { diff = BurstReplicas }++ // throttle machine creation until all stale machines are deleted+ if len(staleMachines) > 0 && diff > machineDeletionWindow {+ diff = machineDeletionWindow+ }
Thanks. This and other suggestions should be encoded into unit tests. I'll get right to it.
comment created time in a month
issue commentflant/shell-operator
Add building instructions to docs to guide outside cluster debugging
If you've got a concrete feature request then open an issue. If you've got a general question, you can reach us on Slack or Telegram. They are
comment created time in a month
issue commentflant/shell-operator
Add building instructions to docs to guide outside cluster debugging
@rspurgeon We'll soon have a talk about this project. We'll certainly do some user-facing improvements before this date. Thanks for reporting a usability issue!
comment created time in a month
Pull request review commentgardener/machine-controller-manager
An option to throttle deletion of Failed Machines
func (c *controller) terminateMachines(inactiveMachines []*v1alpha1.Machine, mac return nil } +func (c *controller) terminateStaleMachines(inactiveMachines []*v1alpha1.Machine, deletionWindow int, machineSet *v1alpha1.MachineSet) error {+ var (+ wg sync.WaitGroup+ errCh = make(chan error, deletionWindow)+ )+ defer close(errCh)++ wg.Add(deletionWindow)+ for i := 0; i < deletionWindow; i++ {+ go c.prepareMachineForDeletion(inactiveMachines[i], machineSet, &wg, errCh)+ }+ wg.Wait()++ select {+ case err := <-errCh:+ // all errors have been reported before and they're likely to be the same, so we'll only return the first one we hit.+ if err != nil {+ return err+ }+ default:+ }++ return nil+}++// calculateStaleMachineDeletionWindow computes an amount of Failed Machines that can be deleted based on FailedMachineDeletionRatio+func (c *controller) calculateStaleMachineDeletionWindow(activeMachines, runningMachines, inactiveMachines []*v1alpha1.Machine, machineSet *v1alpha1.MachineSet) int {+ var (+ deletionWindow int+ numOfActiveMachines = len(activeMachines)+ numOfRunningMachines = len(runningMachines)+ numOfInactiveMachines = len(inactiveMachines)+ desiredMachineSetReplica = int(machineSet.Spec.Replicas)+ )++ // nothing to do+ if numOfInactiveMachines == 0 {+ return 0+ }++ // Will not delete any stale machines. It is expected that the user deletes these machines manually.+ if c.safetyOptions.FailedMachineDeletionRatio == 0 {+ return 0+ }++ if c.safetyOptions.FailedMachineDeletionRatio == 1 {+ return numOfInactiveMachines+ }++ if desiredMachineSetReplica == 0 {+ deletionWindow = numOfInactiveMachines+ } else {+ deletionWindow = int(math.Round(float64(desiredMachineSetReplica) * c.safetyOptions.FailedMachineDeletionRatio))
After some deliberation, I fully agree. There is no reason to burst Machine creation if there is no need. Pushed the change.
comment created time in a month
push eventflant/machine-controller-manager
commit sha cacd37647e8ac1e8cdc5b95136d5121e4c16b7a8
Clamp stale Machine creation AND deletion to machineDeletionWindow Signed-off-by: Andrey Klimentyev <[email protected]>
push time in a month
created tagflant/cloud-provider-vsphere
Kubernetes Cloud Provider for vSphere https://cloud-provider-vsphere.sigs.k8s.io
created time in a month
push eventflant/cloud-provider-vsphere
commit sha 07a5a7313ae4a3c8a043674b89bb798385d7dd46
Fixed a bug in InstanceExistsByProviderID() Signed-off-by: Andrey Klimentyev <[email protected]>
push time in a month
created tagflant/cloud-provider-vsphere
Kubernetes Cloud Provider for vSphere https://cloud-provider-vsphere.sigs.k8s.io
created time in a month
created tagflant/cloud-provider-vsphere
Kubernetes Cloud Provider for vSphere https://cloud-provider-vsphere.sigs.k8s.io
created time in a month