Proxmox Virtual Environment Complete Course Part 15 - Clustering

[@MichaelCook-oo8lj]

There are precious few tech KZbinrs making content as truly engaging and and educational as Learn Linux TV. Applause.

[@DaddyElfMan]

took me about a week to get thru all 16 vids.. very well done. very informative. been in IT for about 40 yrs now, novice/intermediate to linux and proxmox, and last time I used or worked with vm's was back in 2012-13. building a home lab with some old decomissioned dell power edge servers (r710 & 720), for learning and for fun, (setting up minecraft server (java) and media server (plex more that likely)... but HUGE THANK YOU for these vids.... seen you had some about linux as well, will be watching those this week coming...

[@theSwift715]

I can always count on your tutorials. Thank you for not making half of the guide a Zima board advertisement

[@samzick6301]

At the end of your course, you should do a mini series of random things to do with proxmox, like device passthrough to LXC, or passing audio from VM to host with scream.

[@SwiatLinuksa]

+1

[@Jordan-hz1wr]

Device passthrough would be very helpful. I spent well over an hour a few weeks ago trying to get a legacy software license USB key working on a migrated VM.

[@houserouterhouserouter5808]

I agree, would be much more fun and interesting!

[@Vaillant44]

Jay, thank you for the video. I just want to share some tips that I faced as a newbie. You cannot add to a cluster a Proxmox which already has containers or VM and clustering is easy to create but a bit more challenging to alter. Also, clustering two different install formats, for exemple ext4 and zfs raid 1 is creating issues. I'm sure there are ways to solve that last one but it was too advanced for me.

[@TheClembo]

Just loving this comprehensive tutorial on Proxmox, especially having watched the previous episode on shared storage, been waiting for that. Goes without saying it but I do anyway, thanks so much!

[@microcisco]

Hi Jay - really enjoying your videos. Just clarification on the cluster - You actually have to apply an IP address for each network you create for the bridge. You created networks and you didnt set up the cluster in a different network but rather created the cluster in the mgmnt network and not cluster network. I have 4 networks (mgmnt, cluster, VM's and Ceph) and I followed your example but found I couldn't add in the cluster network as it needed an IP address and not network address. If you see your vmbr1 bridge under networks it has an IP address and not a network address. As soon as I changed mine, i could add in the cluster network and all worked fine - Just my 2c :)

[@alain-pierrep.312]

Wonderful series, very informative. Side note : I highly recommend "prettyping" over ping. It's a simple bash script to "prettyfy" ping, showing latencies and lost packets graphically.

[@davidg4512]

The reason the output stops is because it joins the cluster and invalidates your current session as your user is no longer valid. There is some delay when this occurs, hence why sometimes you see some output and other times you see half of the expected output. It really depends. You can see this because the GUI now has a connection error.

[@ManishSahay]

Hey, Congrats Jay on 200K :)

[@gilbert1975nf]

We get lucky nowadays. Back in 2015 create a cluster and HA into Proxmox was a pain in the ass.

[@s.z.6118]

What is so lucky about it ? It still does not work. Do not believe the video. It is NOT like "click, click , next, next... finish". If you follow the video then you get locked out of the nodes' Web interfaces.

[@jenniferw8963]

1:34 start

[@MatthewCrofut]

Would like to see a video on quicksync passthrough. Some more info on privileged vs non-privileged containers would be great as well.

[@felicemassaro9564]

I noted that after clustering 3 servers, VMs do not start after rebooting server one if the other two servers of the cluster are still switched off. Moreover in such conditions even the backup on server one does not run until I switch on the other two servers. Is this normal? Would be nice to have a video on "how to undo a cluster".

[@jasonli2287]

Love it. Thanks so much for the help.

[@ericsong9433]

+1

[@zyghom]

I am just creating my clusters so thank you for this video. However, "it requires reboot" on Linux is so much... WINDOWS ;-)

[@5thpixel]

love these! thanks!

[@applemodus]

Thanks for the video

[@JaminFernandez]

Amazing, thanks

[@madmuke]

thank you so much for doing this

[@heckyes]

Incredible. How is promox free and open! How the hell can it do this while the VM is active. At what point in time does the virtual kernel leave one CPU and start on the other?

[@rars007]

thank you, very cool and informative!!!

[@SwiatLinuksa]

Big thanks! Jay ! And i have question. It's possible that your book about mastering Ubuntu will be translated to polish language? I think it's very interesting but i and many other in Poland doesn't know English so well too understand it, specially a technic language :( Maybe this is stupid question but better ask than not. Thx

[@udayarpandey3937]

Thank you Jay.. you are great :)

[@plazotronik]

what about ceph?

[@MrWadezz]

yes please !

[@jelliott3604]

If you have one server running proxmox then surely you have access to 3 virtual proxmox machines for a cluster? (to expt and follow along on)

[@nyanates]

Nice vid. I'm trying to set up a similar PM cluster but one of them can't join. QQ: Do all the nodes in a Proxmox cluster need to be at the same version? Not the case with these 2x.

[@BrianThomas]

Hello. Great video. Question: What benefits would you have on putting Proxmox on different devices? Aside from high availability.

[@help2day]

nice channel, love it

[@Techtips200]

Superb content

[@463mick]

I noticed in the linux bridge you put a network id in and not a static address. Will this allow for the dhcp to work on that interface.

[@Jordan-hz1wr]

Jay, is it possible to migrate a VM that is using shared storage, onto another nodes local storage?

[@Gheotic]

I'm thinking the same since it does not let me press migrate without. Haven't tried with shared storage though.

[@vancedajcg8724]

By shared storage do you mean an NFS mount?

[@crimsionCoder42]

So does a proxmox cluster increase availability and stability or reduce the time for task completion by separating out the tasks, like, let's say, machine learning and data science calculations? Or is it both tasks are separated across the node, and high availability is there by design?

[@fuseteam]

Hmmm so we don't need to set up corosync and passwordless login between the two uhh three?

[@tokoiaoben3842]

Thank you very much. You created second network (brides) on the pve2 and pve3 to match the network settings on the pve1, is that requirement b4 you that live migrate ? I have several VMs on one server and each VM assinged to different vlan. Creating other vlan networks on other servers (or pves) is not possible since I have only few intherfaces on the node.

[@stevennicholas5472]

How does clustering handle shared VM /Container ID? If is have two servers, each with webserver (100), how would clustering be affected? Would a backup and restore auto,matically reassign that ID?

[@somchaisichon7264]

How to access promox cluster in case of the Primary server failed ?

[@majorgear1021]

Good question. I ran into this yesterday while changing hardware on one node in my 2 node cluster. On boot, cluster members vote on who's in charge, and the logic is you need a 51% or higher vote to win, or reach "Quorum". That's why the TrueNAS forums will tell you to have at least 3 nodes in a cluster. That way you can get a 2/3'd or 66% majority vote. If you don't mind editing conf files, you can change the vote requirements to "1". That allows a single server to vote itself the "boss" of the cluster. However, that setting probably gets overwritten when you upgrade Proxmox. So keep that in mind!. I read that you can run proxmox on a raspberry pi and join it to a 2 node cluster to make a 3 node cluster. I have a few spare rpi's so I'm going to go that route.

[@ierosgr]

At last you reached the part of the series I was expecting. Maybe you could help me with what I am dealing with,. In the first node (2 ssds in mirror for prox, 4 sas hdds 10k in raid 10 for VMs and 4 sata hdds in raid 10 for data )I have a VM which resides in a raid 10 sas drives. This VM has an extra storage to it from another raid 10 configuration (tje sata drives) . Now how could i migrate the VM to another node since it has attached to it a second storage from the first node? Any thoughts it would be very helpful since I am struggling with this for months.

[@fbifido2]

When you migrated VM 101 & 102, could you have migrate the storage from local-lvm to the NFS storage (pve-shared) ??

[@Gilgwathir]

What a cluster.... win?

[@lechaldon]

I plan on working my way through these videos and creating ansible playbooks/roles for these tasks in proxmox. How difficult do you think that will be to achieve clustering+ha in ansible and has anyone already done it? Is there a role to automate this deployment?

[@mohammedsaif3922]

Do i need to keep identical hardware for the migration support.

[@majorgear1021]

From my experience , you need to have volumes with the same names on each server, because Web based migration only goes from same name on source to same name volume on the destination . I think you can override this one the command line. So, not the same hardware, but the same names ( and enough room!) on storage volumes.

[@BrianThomas]

So I follow the steps exactly as they were presented. However I'm constantly being asked for username and password on the other machines that have joined the cluster. I enter in the route username and the root password, and it still comes back invalid. Does anyone have any suggestions as to what I might be doing wrong?

[@psionicp.7962]

Why are you using a bridge for cluster network?

[@davidclark9173]

can you do a video on removing a server from a cluter say like when one dies?

[@arielgonzalez6282]

Is there a way to migrate VMs that have snapshots and not using ZFS or shared storage?

[@brianhansen9578]

Not a single word about quorum in a cluster video ?

[@RodolfoPalau]

I need to solve this. In 2 pve cluster, i can't start VMs if i loose 1 pve. Quórum problem

[@brianhansen9578]

@@RodolfoPalau you can run a 3rd very small instance , just for that purpose... 2 instances than runs the VM's and one small instance just to add quorum

[@MrWadezz]

Quorum isn't a cluster spec, it's a HA one, it's in the next episode :)

[@haiderkhan4167]

Great sir, I have one confusion, I changes IP for one of them server, then cluster will break or not, can you explain litle bit more?

[@vancedajcg8724]

Changing any node IP will definitely break the connectivity. You should exclude the node from cluster before doing something like that and then only adding the node back to cluster once all is good again. Proxmox clustering is not easy. I'd recommend staying away from it if you have anything as small as a 3 node home setup or less.

[@muhammadejaz9841]

Thanks for sharing such a great experience. very nice explanation. Sir Please include class on Replication (live and scheduled replication). Related to Clustering please tell me can we share resources of servers in clusters (i.e CPU, RAM or HDDs) in case one server utilizing full resources and we need to increase it CPUs or RAM can we take these from other server in the cluster??? or this can be done by some other method? I really appreciate your Hard work. Thanks

[@meerhassan8066]

according this tutorial when i am trying to join a cluster i am getting this error message below, (detected the following error(s): * this host already contains virtual guests TASK ERROR: Check if node may join a cluster failed!) How Can i solved this problem?? need some advice.... Thanks in Advance...

[@vancedajcg8724]

You'll have to stop all vms, back them up , delete them, add the node to cluster, then restore the vms.

[@smash_shane]

I understand that you can't join a node to a cluster if it has an existing VM or CT. I have 3 nodes with VMs on them that want to join together. Can I backup up the VMs and containers.... Delete them from the node.. Join them to the cluster and then restore the backups? Will that work?

[@brianhansen9578]

That should work just fine, a shared storage makes this very easy. Also remember that all 3 nodes also need to have same type of storage, you can't mix LVM on one and ZFS on another

[@AmrFayezAmer]

when migrate vm proxmox don't show target storage, how show target store for select store another proxmox for migrate vm?

[@voiceoftreason1760]

I think the novelty of the matrix in the terminal scripts really wore off by now

[@GHETTOHAMBURG]

didnt work at all bro, im quiting my IT job now

[@floriantthebault521]

I would definitely support you as a patron, but.... with 5 to 6 ad slots inside a video of 20 minutes... I guess that's far enough (because it's soooooo annoying as f******) - content is top notch, though. That's too bad it's polluted by so many ads it makes the video unwatchable.