- Kubernetes in the Cloud (Part 1)
- Kubernetes in the Cloud (Part 2)
- Kubernetes in the Cloud (Part 3) – (Almost) Everything You Need to Know about Stateful Workloads
Introducing the Storage Networking Security Webcast Series
It’s a Wrap for SNIA and the Solid State Storage Initiative at Flash Memory Summit 2019
A Best of Show award, over 12 hours of content, three days of demos, and a new program drawing attention to persistent memory programming completed – Flash Memory Summit 2019 is officially a success!
SNIA volunteers were again recognized for their hard work developing standards for datacenters and storage professionals with a “Most Innovative Flash Memory Technology” FMS Best of Show award. This year, it was SNIA’s Object Drive Technical Work Group who received kudos for the SNIA Technical Position Key Value Storage API Specification. Jay Kramer, head of the FMS awards program, presented the award to Bill Martin, Chair of the Object Drive TWG, commenting “Key value store technology can enable NVM storage devices to map and store data more efficiently and with enhanced performance, which is of paramount significance to facilitate computational storage. Flash Memory Summit is proud to recognize the SNIA Object Drive Technical Work Group (TWG) for creating the SNIA Technical Position Key Value Storage API Specification Version 1.0 defining an application programming interface (API) for key value storage devices and making this available to the public for download.
SNIA Sessions at FMS Now Available for Viewing and Download
SNIA Executive Director Michael Oros again took the mainstage to describe “Standards that Can Change Your Job and Your Life” encapsulating SNIA work in three core areas: persistent memory, computational storage, and storage management.
Also at Flash Memory Summit, SNIA work and volunteers were on display in eight sessions on persistent memory, highlighting advances in persistent memory, PM software and applications, remote persistent memory, and current research in PM, sponsored by SNIA, JEDEC, and the OpenFabrics Alliance. A new 2019 SNIA-sponsored track on computational storage featured four sessions on controllers and technology, deploying solutions, implementation methods, and applications.
SNIA’s SFF Technology Affiliate highlighted their work on the Enterprise and Datacenter 1U Short SSD Form Factor (E1.S) specification SFF-TA 1006, while the Object Drive TWG expanded on their work in standardization for a key value storage interface underway at SNIA and NVM Express. SNIA also presented a preconference seminar tutorial on persistent memory and NVDIMM, and a session on Storage Management with Swordfish APIs for Open Channel SSDs.
A new session on programming to persistent memory featured a tutorial (video available soon) and a 2 ½ day Persistent Memory Programming Hackathon where attendees programmed to persistent memory systems and discussed their applications. Next up for the Hackathon series – a 2-day event at SNIA Storage Developer Conference.
Find PDFs of these sessions by clicking on Flash Memory Summit 2019 under Associated Event in the SNIA Educational Library.
We continued our discussions on the exhibit floor featuring JEDEC-compliant NVDIMM-Ns from SNIA Persistent Memory and NVDIMM SIG members AgigA Tech, Micron, SMART Modular Technologies, and Viking in a Supermicro box running an open source performance demonstration. If you missed it, the SIG will showcase a similar demonstration at the upcoming SNIA Storage Developer Conference September 23-26, 2019, and at the 2020 SNIA Persistent Memory Summit January 23, 2020, both at the Hyatt Regency Santa Clara. Click on the conference names to register for both events.
It’s a Wrap for SNIA and the Solid State Storage Initiative at Flash Memory Summit 2019
Kubernetes Links & Resources to Keep You in the Know
Our recent SNIA CSTI webcast, “Kubernetes in the Cloud (Part 3): (Almost) Everything You Need to Know about Stateful Workloads” offered a wealth of insight on how to address the challenges of running stateful workloads in Kubernetes. This webcast was the third installment of our Kubernetes in the Cloud webcast series and it is now available on-demand as are “Kubernetes in the Cloud (Part 1)” and “Kubernetes in the Cloud (Part 2).”
Our expert presenters, Paul Burt and Ingo Fuchs, have provided additional resources to help keep you in the know on Kubernetes. Here they all are:
- Anyone running Kubernetes needs to put a bit of thought into how they approach stateful workloads. Containers by their nature are ephemeral. So, what options are there for the average Kubernetes user? This talk aims to provide an overview of the most obvious options, and provide resources for learning more about them. https://twitter.com/kubernetesio/status/840257886202683392 and https://www.youtube.com/watch?v=4x1r3Osu1Kg
- Dockerizing MySQL at Uber Engineering https://eng.uber.com/dockerizing-mysql/
- Kelsey Hightower Tweet https://twitter.com/kelseyhightower/status/963413508300812295?lang=en
- Kubernetes secret objects https://kubernetes.io/docs/concepts/configuration/secret/
- Kubernetes external secrets https://github.com/godaddy/kubernetes-external-secrets
- Secrets management and data protection with HashiCorp Vault https://learn.hashicorp.com/vault
- Katakoda: Learn Kubernetes Using Interactive Browser-Based Scenarios https://www.katacoda.com/courses/kubernetes
- Operating Kubernetes Clusters and Applications Safely https://kubernetes-security.info/ and https://info.aquasec.com/kubernetes-security
- Kubernetes Best Practices: Mapping External Services https://cloud.google.com/blog/products/gcp/kubernetes-best-practices-mapping-external-services
- Example of mapping an external service by URI https://cloud.google.com/blog/products/gcp/kubernetes-best-practices-mapping-external-services
- Map ports and IP address directly https://cloud.google.com/blog/products/gcp/kubernetes-best-practices-mapping-external-services
- Kubernetes introductory tutorials https://kubernetes.io/docs/tutorials/stateful-application/mysql-wordpress-persistent-volume/
- Kubernetes introductory tasks https://kubernetes.io/docs/tasks/run-application/run-single-instance-stateful-application/
- To run or not run a database on Kubernetes: What to Consider https://cloud.google.com/blog/products/databases/to-run-or-not-to-run-a-database-on-kubernetes-what-to-consider
- Introducing Operators: Putting Operational Knowledge into Software https://coreos.com/blog/introducing-operators.html
- Kubernetes Concepts: Custom Resources https://kubernetes.io/docs/concepts/extend-kubernetes/api-extension/custom-resources/
- Rules and criteria for Operators https://coreos.com/blog/introducing-operators.html
- GitHub etcd-io/etcd https://github.com/etcd-io/etcd/blob/master/Procfile.v2
- Couchbase cbbackupmgr Tool https://docs.couchbase.com/server/6.0/backup-restore/enterprise-backup-restore.html
- List of Operators https://operatorhub.io/
- Operator Frameworks https://github.com/operator-framework
- Should I use a configMap or a custom resource https://kubernetes.io/docs/concepts/extend-kubernetes/api-extension/custom-resources/#should-i-use-a-configmap-or-a-custom-resource
- AWS Service Operator for Kubernetes https://aws.amazon.com/blogs/opensource/aws-service-operator-kubernetes-available/
- Examples how simple accessing an external service can be https://github.com/awslabs/aws-service-operator
- Kelsey Hightower Tweet https://twitter.com/kelseyhightower/status/822488055709712384?lang=en
If you know of some Kubernetes resources to share, please comment on this blog and we’ll add them to our list.
Kubernetes Links & Resources to Keep You in the Know
Kubernetes in the Cloud Q&A
Kubernetes is a hot topic these days, generating lots of interest and questions. The goal of our SNIA Cloud Storage Technologies Initiative Kubernetes in the Cloud webcast series is to cut through the hype and provide a vendor neutral look at what Kubernetes is and how it is being used. Our most recent webcast, Kubernetes in the Cloud (Part 2), generated some interesting questions. Here are answers from our expert presenters.
Q. If I’m running my Kubernetes infrastructure at a cloud service provider, do I need CSI support by the cloud provider? If this is not available, I will need a virtual storage array that provides CSI leveraging the underlying cloud storage. Do you know whether there are solutions on the market that I can deploy as a virtual machine at my cloud provider?
A. Current solutions using the CSI interface for public cloud storage are not available at this point. It will be up to the cloud provider to decide whether to support those interfaces to their storage layers.
Q. Does each pod run on one CPU core? I am trying to understand how to size the server configuration?
A. Containers use Linux cgroups to limit the amount of CPU and memory a container can consume and this is exposed in Kubernetes as limits that you can set.
Q. In today’s environment for Kubernetes Flex storage, what is the suggested process to “backup” these stateful PVs or is that not necessary anymore?
A. Backups are still as important with containers as they are today with traditional applications. There are many different approaches available to backup containers: storage snapshots via native storage interfaces, deployment of backup clients in containers, application level backups, etc.
Q. You mentioned Kubernetes a lot – but what is the status of native Docker CSI support? Can I use CSI for usage within Docker without deploying Kubernetes? And if yes: Can I get rid of the need for docker volume drivers then?
A. Docker universal control point (UCP) support of CSI is currently in beta. Once UCP is generally available, we’ll be able to answer your question in more detail.
Interested in more Kubernetes in the Cloud information? Watch our first installment Kubernetes in the Cloud (Part 1) on-demand at your convenience and sign up for our next webcast, Kubernetes in the Cloud (Part 3): Stateful Workloads which will be live on August 20, 2019 and available on demand after that.
Kubernetes in the Cloud Q&A
Stateful Workloads on Kubernetes: (Almost) Everything You Need to Know
Kubernetes is great for running stateless workloads, like web servers. It’ll run health checks, restart containers when they crash, and do all sorts of other wonderful things. So, what about stateful workloads? Large implementers like Uber say to avoid it if you can [1], and gurus like Kelsey Hightower echo that sentiment [2].
It’s the topic we’ll address on August 20th at our live SNIA Cloud Storage Technologies Initiative webcast “Kubernetes in the Cloud (Part 3): Stateful Workloads.” In this session, we’ll explore when it’s appropriate to run a stateful workload in cluster, or out. We’ll discuss the best options for running a workload like a database on the cloud, or in the cluster, and what’s needed to set that up.
We’ll cover:
- Secrets management
- Running a database on a VM and connecting it to Kubernetes as a service
- Running a database in Kubernetes using a `stateful set`
- Running a database in Kubernetes using an Operator
- Running a database on a cloud managed service
Register today to save your place on August 20th. This is the 3rd installment of our Kubernetes in the Cloud webcast series. Kubernetes in the Cloud (Part 1) and Kubernetes in the Cloud (Part 2) are available on demand. I encourage you to check them out for great information and demonstrations on Kubernetes.
[1] https://eng.uber.com/dockerizing-mysql/
[2] https://twitter.com/kelseyhightower/status/963413508300812295