my journey to proxmox

V

virttom

Member
Proxmox Subscriber
Mar 20, 2020
47
14
8
50
France
Hi all,
I'm an old ESXi/M$ guy who's trying to convert to opensource and to convert my client since I'm consultant/architect in virtualisation / cloud domain. This is not a testimony just to give you some context about my "noobitude" on gnu/linux in general and proxmox in particular.

So I'm actively trying to convince a client to go on Proxomox and doing a POC on a random machine before architecting the true product. Working on 6.2 with a lonely server installed on a lonely disk in my companies R&D environnement.

This is day 2 and I've to confess that it's kind of weird experience ! yesterday my host was up but I couldn't connect to my node by GUI. After some little check, i decided to reboot since restarting daemons seems too long (maybe 5 minute after pvedaemon restart didn't finished). Everything was ok then. Today i can connect on graphical interface but the cluster (yes 1 node cluster) is some sort of grayed and cannot acces to anything. Had a error 595 and after restarting
pve-cluster
pvedaemon
pveproxy

I now have communication failure (0)

any idea on this ?
 
Thanks for your answer ! but definitly not the case. I have a lonely HDD for the purpose of this POC. I'll have another but it's definitively local ones.

I rebooted since I need to configure some stuff on it. here is my "pvesm status"

root@pve:~# pvesm status
Name Type Status Total Used Available %
local dir active 34829920 2663100 30367860 7.65%
local-lvm lvmthin active 79933440 0 79933440 0.00%
Click to expand...
 
What is the output of the following?
Code: 
pvecm status
 
Code: 
Cluster information
-------------------
Name:             CS
Config Version:   1
Transport:        knet
Secure auth:      on

Quorum information
------------------
Date:             Tue Oct 13 10:38:16 2020
Quorum provider:  corosync_votequorum
Nodes:            1
Node ID:          0x00000001
Ring ID:          1.3c
Quorate:          Yes

Votequorum information
----------------------
Expected votes:   1
Highest expected: 1
Total votes:      1
Quorum:           1
Flags:            Quorate

Membership information
----------------------
    Nodeid      Votes Name
0x00000001          1 172.x.x.x (local)
 
Last edited:
ok, now that I have another hard drive and no more intempestive proxmox graphical lost, it's working great ! seems that my current client is convinced on the first project relative to potential as a stand alone product.

next step will be :
  1. ovf import,
  2. FS consideration (ZFS or LVM),
  3. QM practices
  4. Backup/Restore
  5. FW configuration,
  6. Monitoring
in a random approach, sounds good ! really nice product.

Next stair will be HCI cluster approach, probably with CEPH, need to investigate on gluster too.
 
hello all !

due to dicontinuation of project currently on proxmox, i was not on it but hey i'm here and back again ;)

So yesterday I did work on migration of W2K16 on my standalone node. I had some confusion regardin to admin guide that is not aligned w/ wiki so for now on, i'll be doublechecking on both to be sure. VM was perfectly migrated even if my storage is too light ( 76Go free before migration). So today, after a windows update, well you know what happened ^^ GUI was down (but SSH and hypervisor consol still responding) and after some investigation :

Code: 
root@pve:~# qm status 150
status: io-error

#### not good !!!

root@pve:~# lvs


#### no response !!! definitly not good...

and after a host reboot (yes I know that i need to work heavily on my linux/debian knowledge... this on top of my todo list).

Code: 
root@pve:~# lvs
  LV            VG  Attr       LSize  Pool Origin Data%  Meta%  Move Log Cpy%Sync Convert
  data          pve twi-aotz-- 76.23g             100.00 5.35                   
  root          pve -wi-ao---- 34.00g                                           
  swap          pve -wi-ao----  8.00g                                           
  vm-150-disk-0 pve Vwi-a-tz-- 80.00g data        95.29

now I need to investigate on how to hot add a hard drive on proxmox ^^
 
I had some confusion regardin to admin guide that is not aligned w/ wiki so for
Click to expand...
In doubt the administration guide should be correct.

now I need to investigate on how to hot add a hard drive on proxmox ^^
Click to expand...
You plug it in and mount (fstab) it like in every other distribution ;) Potentially adding a storage in the GUI afterwards.
 
I need to use pvcreate in order to initialize le PV to be used by LVM no ?
Click to expand...

If the disk is empty then you have to plug it in and do the rest in the GUI. You'll see the disk in the GUI if you click on your node->Disks. It should say "Usage no" for the disk.
Then you can click on LVM-Thin (or anything else) and click Create Thinpool and leave the "Add Storage" checkbox ticked. This should be the whole procedure.

If you cannot select it in the Create Thinpool dialog (=> not empty) and do not need the data on it anymore, then you have to make really really sure, that you know the name (for example /dev/sdc ...) of your disk and run (assuming your disk is /dev/sdc)
Code: 
wipefs -a /dev/sdc
dd if=/dev/zero of=/dev/sdc bs=1M count=200
Then do the procedure from above.
 
Hi all hope you're fine.

So as I said I'm working on a new architecture on my second project which can be described as a k8s/docker/rancher + S3 like storage and PostgresSQL for metadata concern app. Since the begining I was working on standard w/ 3 worker / master + 1 rancher node. So It was somewhat matching HCI target. So I started my studies on it (and proxmox + CEPH or proxmox + gluster was one of the competitor). But it was yesterday. Now it seems that my client is thinking on evolving its product so they asked me :

"hey bob ! (let say my name is bob) what do you think about puting all this stuff in a single k3s cluster and get rid of those pesky HCI cluster (since their application is not that demanding on the server side and let client side using its CPU). can you think about it ?"

So as my brain is bleeding, I decided to work on this new fresh (even desturbing) choice. I proposed them to still work w/ an hypervisor + a SDS on a standalone mode and make Async replication on a second hypervisor just to have some sort of resiliency (they have a 4 hour restart after crash requierment). Do you think it's a good idea ? do you have some reco / pros and cons on HCI vs 2 bigger active/standby node ? w/ the use of PROXMOX VE ?+ CEPH on both case just to compare same things.

My thinking on it is that it can be a good idea to simplify architecture on the network side and not have to maintain a ditributed storage if it's not a requierement to kkep application working whatever the cost.
 
ok thanks. Is there some rule of thumb to calculate what will be necessary for my hypervisor to maintain backup / monitoring services to calculate the overall needs of my infrastructure ? It is a for an evaluation to give some comparisaon element to my management. It is not a firm evaluation of hardware cost at this stage. This will be precisely calculated under real conditions during R&D POC. A thing like what I have on PVE admin Guide for evaluating my needs for CEPH storage ?
 
I'm not 100% sure what you mean.

If you want to do ZFS storage replication, then an important consideration might be getting enough (ECC) RAM. This actually also holds for Ceph. Fast & durable disks are relevant for all types of storages, too. That is probably more expensive aspect.

Resource requirements for unknown/external monitoring services that you want to run on your PVE nodes depend highly on which service you use.

We did some benchmarks for Ceph. They also contain speed tests for a couple of hard disks and SSDs and the 2018 benchmark also contains a comparison of networks with different bandwidth. Maybe you can use this for your evaluation.
 
well, I've read your stuff relative to benchmarks. IMO, it's a bit overkill to work on this for my present concern. I'll rely heavily on those when i'll be testing / contructing my POC, and it will be precious at this time. Or juste before to give my cookin' recipe to my hardware reseler. For now on I just want some sort of "relative" indications like : reserve 10% of your compute for your HCI or as a rule of thumb count 1Go de RAM by To of storage. Just to have an vague knowledge of what i will invest to have my stuff working and compare to other hardware/HCI plateform.
 
Thx for the ref jsterr, i've used this to evaluate my needs on storage. My concern is now to evaluate the same thing on backup & monitoring so a a kind of abacus like the one you pointed me would have been priceless ;) I've made some approximation and it will be ok for this time. Next step will be POC !
 
You must log in or register to reply here.

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Quick Navigation

Home Get Subscription Wiki Downloads Proxmox Customer Portal About

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!
Community platform by XenForo® © 2010-2024 XenForo Ltd.
Top Bottom