Video in TIB AV-Portal: Kubernetes+GlusterFS

Formal Metadata

Lightning Ver.
Title of Series
CC Attribution 2.0 Belgium:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor.
Release Date
Production Year

Content Metadata

Subject Area
You can now deploy GlusterFS on Kubernetes, and have your Kubernetesapplications access hyper-converged persistent storage. Ashiq and Jose willexplain how. AMENDMENT: Replacement for "The Easiest Way to Deploy OpenShift", which had tobe cancelled due to speaker travel issues.
Surface Electronic mailing list Data storage device Control flow Volume (thermodynamics) Representational state transfer Mereology Revision control Arithmetic mean Data management Software Different (Kate Ryan album) Term (mathematics) File system MiniDisc Energy level Communications protocol
Mobile app Software Multiplication sign Surface Network topology Data storage device Data center Set (mathematics) MiniDisc Cartesian coordinate system Flow separation
Point (geometry) Group action Demo (music) System administrator Data storage device Volume (thermodynamics) Hecke operator Representational state transfer Cartesian coordinate system Personal digital assistant Videoconferencing Right angle Social class
Slide rule Presentation of a group Link (knot theory) Videoconferencing
go ahead all right how do people this is
kubernetes Singh cluster FS finally plain nice together lightning version I'm Jose this is a
chic let's see what we can do in five minutes hi so bless profess so mostly everyone knows glass surface I think so glycerin versus a distributed storage software different storage file system so it it a grits stories from different nodes and gives out us a single volume to the customer or the user so how its minimum level is breaks and from my device and LV is created and that small part of LV is a break we take Elvis from each node mean bricks from each node in terms of glycerin face and create a volume out of it and mount it anywhere we have protocol supporting fuse and NFS and samba so that's what glistrup versus in shorter version and hecka theha katisa restful interface which can manage glossary FS volumes or create lists refers volumes for you you don't have to go behind the screen and create the lv SVG's and manage your disks so Hecate will take care of it and jose is going to continue on the cool part cool part i say alright so these
days closed surface and hikari are fully hyper-converged with kubernetes that just means that they're running as containers in a kubernetes cluster serving storage for other for other kubernetes applications obviously with hyperconvergence this means that you guys can save money in your data centers by not having to buy separate storage appliances just buy a couple SSDs or something these the applications have applications in kubernetes have native access to storage from kubernetes which means that it's you know a little less network latency and it's less prone to and it's less prone to outages from the network because if your if your apps can't access your storage that means your that means your kubernetes cluster is down and you have other problems and the glossary first node can don't need to run on every node all the time they only need to run the nodes actually have storage on them so you can actually have nodes that aren't there aren't come that are not completely alike and the gloss the glossary FS Damon set will take care of making sure they're running in the right place and we kept the ability to easily scale out that is given that is that is brought
with glossary FS here's a little picture of how things are supposed to look you'll see that Hetty is running it's a single pod on one of the nodes it doesn't matter which note it's on and the three nodes I have storage attached to them have Gluster FS pods running in them that are all logically linked together into one cluster FS cluster and as I'm as you can see there they don't have to have the same topology of disks on every node we just need to know what the topology is so that we can serve it
to the group of kubernetes applications and we also we also allow for dynamic provisioning if you're familiar with that in Kerber Nettie's the administrator on the right there provides a starch class which defines that applications can use some underlying storage and the admin manages that by saying an endpoint that tells it where to get access to the glossary FS pods and a rest URL that points to the heck Hettie rest interface and after that the user on the left there only needs to know the name of the storage class in this case cluster and when you create your persistent volume claim over on the left there it gets attached to the starch class and when and then creates a persistent volume in some storage somewhere in this case it create that Nemec li creates a cluster volume on your on your Gluster storage and that then gets attached to the PVC and can be used by an applicator and that that is now a persistent volume that can get attached to a PVC and used by a kubernetes application from the user and
unfortunately at this point there would be a demo but the demo video itself is about four minutes long which is about
as long as its presentation so thank you very much he's a sheik i'm jose here's a couple of links these slides will be online after this presentation and hopefully the video - we'll see thank you [Applause]


  300 ms - page object


AV-Portal 3.20.2 (36f6df173ce4850b467c9cb7af359cf1cdaed247)