DRBD in PVE 2.0

[Andre Köhler]

Good Morning,
its Monday morning and now... Proxmox time...

So i have a cluster with two nodes (Node 1: 172.16.5.122 and Node 2: 172.16.5.123)

I saw that HA is available with the update, so i've made the update like wiki.

My wish is, that the 2 nodes share 1 HDD... so "normal" DRBD. (So that the sda4 from node2 is exactly the same as sda4 from node1)

Can i activate DRBD over the GUI oder only in CLI?

Can you tell me how i configure a normal DRBD without FailOver etc... just a primary/primary or primary(node .122) / secondary(node .123)
is that the same as PVE 1.9?? so in CLI?? or can i activate it in the GUI?!

Thanks for help.

Andre


EDIT:

So i would like to test is with the wiki from Proxmox http://pve.proxmox.com/wiki/DRBD

But the installation get an error...


Unpacking drbd8-utils (from .../drbd8-utils_2%3a8.3.7-2.1_amd64.deb) ...
dpkg: error processing /var/cache/apt/archives/drbd8-utils_2%3a8.3.7-2.1_amd64.deb (--unpack):
trying to overwrite '/usr/share/cluster/drbd.sh', which is also in package resource-agents-pve 3.9.2-2
configured to not write apport reports
dpkg-deb: subprocess paste killed by signal (Broken pipe)
Processing triggers for man-db ...
Errors were encountered while processing:
/var/cache/apt/archives/drbd8-utils_2%3a8.3.7-2.1_amd64.deb
E: Sub-process /usr/bin/dpkg returned an error code (1)



So ... i just want a NORMAL DRBD with 2 Nodes... iam a noob... that must be so easy or not?!

 

[xelaneb]

Bump, ditto

 

[jva1601]

It would be nice to have a web interface to manage DRBD implementation but since the developers of proxmox are very busy I did a search online if there was one available this came up http://lcmc.sourceforge.net/. I haven't tried it yet maybe someone who has could share their experience. After posting I also found this one also http://oss.linbit.com/drbd-mc/

 

[e100]

DRBD is not necessarily "hard" but I would not call it "easy" either.
It is not something you want to use in production until you are very familiar with it.

Currently DRBD can only be setup using the CLI.

I am not sure why you had issues with installing the drbd8-utils, on the wiki page there is a link to directions on building the proper drbd8-utils package:
http://forum.proxmox.com/threads/70....3.10-to-match-kernel-in-1.9?p=40033#poststop

Maybe that will fix your problem?

I would also highly recommend reading this section of the wiki and using two DRBD volumes as suggested:
http://pve.proxmox.com/wiki/DRBD#Recovery_from_communication_failure

 

[xelaneb]

jva 1601 Thanks for sharing that link. http://oss.linbit.com/drbd-mc/ looks great, it's a no go with proxmox 2.0 because of conflicting packages in pve specific implementations of corosync etc. I love the power and simplicity of proxmox and love the improvements in 2.0. Wish the documentation was as complete as it is with 1.x versions (I know it's still in beta).

e100, I re-ran the scripts and successfully built the utils etc.

using this as a guide for drbd with proxmox: http://pve.proxmox.com/wiki/DRBD#Introduction

Bit murky on how exactly to accomplish 2 volumes but this seems to be working so far (both are in the process of sync)

In drbd.conf moved "include "drbd.d/*.res";" to the end of the global section

Created r0.res and r1.res as seperate files under /etc/drbd.d/

r0.res contents:
"
resource r0 {
on proxmox0 {


device /dev/drbd0;
disk /dev/sdb1;
meta-disk internal;


address 192.168.200.200:7788;
}
on proxmox1 {
device /dev/drbd0;
disk /dev/sdb1;
meta-disk internal;


address 192.168.200.201:7788;
}
}
"

copied r0.res to r1.res and changed sdb1 to sdc1 and drbd0 to drbd1 as well as :7789 for port

proceed as normal in http://pve.proxmox.com/wiki/DRBD

Hope this is useful to others