> Dedicated replication link?
>
> Maybe the additional latency is all that kills you.
> Do you have non-volatile write cache on your IO backend?
> Did you post your drbd configuration setings already?
There is a dedicated 10GB Ethernet replication link between both nodes.
There is also a cache on the IO backend. I have started some additional
measurments with dd and oflag=direct.
On a remote host I get:
- With enabled drbd link: 3 MBytes/s
- With disabled drbd link: 9 MBytes/s
On one of the machines locally:
- With enabled drbd link: 24 MBytes/s
- With disabled drbd link: 74 MBytes/s
Same machine but a parition without Drbd and LVM:
- 90 MBytes/s
This is our current drbd.conf:
global {
usage-count yes;
}
common {
syncer {
rate 500M;
}
}
resource lfs {
protocol C;
startup {
wfc-timeout 0;
degr-wfc-timeout 120;
}
disk {
on-io-error detach;
fencing resource-only;
}
handlers {
fence-peer "/usr/lib/drbd/crm-fence-peer.sh";
after-resync-target "/usr/lib/drbd/crm-unfence-peer.sh";
}
net {
max-buffers 8000;
max-epoch-size 8000;
}
on d1106i06 {
device /dev/drbd0;
disk /dev/sda4;
address 192.168.2.1:7788;
meta-disk internal;
}
on d1106i07 {
device /dev/drbd0;
disk /dev/sda4;
address 192.168.2.2:7788;
meta-disk internal;
}
}
Thanks
Christoph
_______________________________________________
Pacemaker mailing list: Pacemaker@oss.clusterlabs.org
http://oss.clusterlabs.org/mailman/listinfo/pacemaker
Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://bugs.clusterlabs.org
