Does the Proxmox VE Cluster require DRBD?

  • Thread starter Thread starter joec22
  • Start date Start date
DRBD is just an option. (but a quite useful)
 
Ok I should be more clear in my question....ProxmoxVE Cluster does not require drbd for live migration to work?
 
DRBD has quite good performance... In comparison to what?? To the speed of an IBM XT floppy drive? It takes me 6 hours to install Mandriva 2010 to DRBD - instead of 30 minutes this process normally taken!!!
 
>ProxmoxVE Cluster does not require drbd for live migration to work?
You may use NFS share for example. Fortunately, Proxmox doesnot depends on DRBD at all, you may use any shared storage solution as you wish.
 
If proxmoxve cluster does not depend on drbd, then why is it necessary to run in primary/primary mode?
 
DRVTiny said:
DRBD has quite good performance... In comparison to what?? To the speed of an IBM XT floppy drive? It takes me 6 hours to install Mandriva 2010 to DRBD - instead of 30 minutes this process normally taken!!!
Click to expand...

You have done something wrong then (I suggest you contact the DRBD mailing list to verify your configuration).
 
dietmar said:
You have done something wrong then (I suggest you contact the DRBD mailing list to verify your configuration).
Click to expand...

I get maximum of 5,92 Mb/sec on a hardrive in RAID 1+0 that support 400Mb/sec and Gigabit Ethernet LAN, that supports up to 956Mbit/sec. So... "where is my performance"?
 
DRVTiny said:
I get maximum of 5,92 Mb/sec on a hardrive in RAID 1+0 that support 400Mb/sec and Gigabit Ethernet LAN, that supports up to 956Mbit/sec. So... "where is my performance"?
Click to expand...

Try with kernel 2.6.32 on both nodes and check if you have better performance.
 
I am already using 2.6.32 kernel with DRBD 8.3.4 and it working unreally slow :(
On a virtual machine (KVM fully virtualized guest)
[zimbra@mailint ~]$ time ( dd if=/dev/zero of=log/DELME bs=1M count=512; sync )
512+0 records in
512+0 records out
536870912 bytes (537 MB) copied, 190.76 seconds, 2.8 MB/s

real 5m29.890s
user 0m0.013s
sys 0m3.104s
 
I have the same performance problem.

On my host DRBD is quite good :

root@node2:~# time sh -c "dd if=/dev/zero of=/dev/drbd/vm-106-disk-1 bs=8k count=200000 &&sync"
200000+0 records in
200000+0 records out
1638400000 bytes (1.6 GB) copied, 49.8865 s, 32.8 MB/s

In my KVM (a fresh debian lenny) with only one IDE drive stored in DRBD volume:

root@debian08~# time ( dd if=/dev/zero of=/tmp/bf bs=1M count=512; sync )
512+0 records in
512+0 records out
536870912 bytes (537 MB) copied, 147.757 seconds, 3.6 MB/s

For info :

root@node2:~# pveversion -v
pve-manager: 1.5-10 (pve-manager/1.5/4822)
running kernel: 2.6.32-2-pve
proxmox-ve-2.6.32: 1.5-7
pve-kernel-2.6.32-2-pve: 2.6.32-7
pve-kernel-2.6.24-11-pve: 2.6.24-23
qemu-server: 1.1-16
pve-firmware: 1.0-5
libpve-storage-perl: 1.0-13
vncterm: 0.9-2
vzctl: 3.0.23-1pve11
vzdump: 1.2-5
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1dso1
pve-qemu-kvm: 0.12.4-1
ksm-control-daemon: 1.0-3

May be a tweak or something ?

Thanks !

nsc
 
ok i reinstall my vm with virtio as disk and network

Here what i get for local sata :
root@debian08~# time ( dd if=/dev/zero of=/tmp/bf bs=1M count=512; sync )
512+0 records in
512+0 records out
536870912 bytes (537 MB) copied, 8.48 seconds, 63.43 MB/s

And for drbd:
root@debian09~# time ( dd if=/dev/zero of=/tmp/bf bs=1M count=512; sync )
512+0 records in
512+0 records out
536870912 bytes (537 MB) copied, 33.76 seconds, 15.9 MB/s

My First test was done with SCSI VM not VIRTIO ...

Sorry
 
You must log in or register to reply here.
Top Bottom