Hi Dietmar,
yes it returns successfull when all replicas are written to journal /
underlying FS.
Not sure about the performance decrease. But generally ceph only shows
full performance with lots of parallism.
I don't know your configuration may be your journal speed is limited?
Stefan
Am 30.12.2013 11:59, schrieb Dietmar Maurer:
Hi all,
we just notices a big write speed difference when replication size differs:
size = 2: 100MB/s
size = 3: 60MB/s
The network is 10GB, so that should not be the limit.
AFAIK ceph uses primary-copy replication, where copies to second and
third node is done parallel?
So why is that so much slower?
_______________________________________________
pve-devel mailing list
[email protected]
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
_______________________________________________
pve-devel mailing list
[email protected]
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
