Re: Support for DRDB

On 1/16/2015 8:26 PM, Digimer wrote:

... CentOS aims to repackage RHEL
> RPMs exactly, then changes only trademarks. Gluster is, as I understand
> it, fully open source so I can see no reason, from a general position,
> why gluster's RPMs on CentOS should be treated any differently.

Because the straight open source RHEL gluster rpm does not include any 
daemons. Those are RHEL's own, they are called "Red Hat Storage Server", 
they are a for-pay "add-on" and so they are not in centos.

>> More or less. IRL they apparently shine on top of a 10Gb network. Or
>> better, three 10Gb networks. DRBD works just fine over a crossover piece
>> of cat-5e.
>
> I use it extensively on both 1 Gbps copper (always through a switch) and
> 10 Gbps SFP+ (again, always through a switch). In all cases, I use dual
> links in simple mode=1 (active/passive) bonding with each leg to a
> different switch in a stacked pair.
>
> In both cases, I can get replication speeds very very close to the
> maximum the underlying network is capable of. So in this light, I am not
> really sure what issues you might be referencing are. Can you expand on
> your comments?

What do you mean by "it"? If "it" is DRBD, I've several active/passive 
pairs running it over cat5e crossover cable it shares with heartbeat. 
I've no complaints about replication speeds either.

The 3x10Gbps was a reference to OpenStack, actually, now that I think 
about it. Ceph just says 10Gb is "worth considering" and gluster sez 
"oh, we could list a few success stories..." and then don't.

Dima

_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems