Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Mark Watts schrieb:
> On Monday 17 December 2007 12:50:26 Tomasz Chmielewski wrote:
>> Is heartbeat really necessary to use DRBD?
>>
>> I just want to replicate an array/disk/volume from one datacenter to
>> another, but I don't need any heartbeat's features - so perhaps
>> heartbeat is not necessary?
>>
>> On the other hand, documentation mentions heartbeat a lot.
>>
>> Can I use DRBD without heartbeat?
>
> Yes. However heartbeat, when properly configured (which isn't hard), will take
> away some of the management issues from you, such as promoting a device to
> Primary status.
>
> Should you use it without heartbeat? Depends on why you are using DRBD in the
> first place.
>
> Heartbeat will allow you to automagically fail-over to the Secondary node in a
> DRBD cluster, allowing you to continue to provide storage access in the event
> of [certain kinds of] failure.
I just want to replicate a SAN (array, disk etc.) to another physical
location.
There is no sense in using heartbeat for me, as when that primary SAN
dies, the asymmetric link from another datacenter will be too slow
anyway to provide services.
So, if I want to replicate a local array to another machine, should I
use 127.0.0.1 address in drbd.conf?
resource drbd0 {
protocol C;
incon-degr-cmd "halt -f";
on thost1 {
device /dev/nb1;
disk /dev/hda7;
address 127.0.0.1:7789;
}
on thost2 {
device /dev/nb1;
disk /dev/hda7;
address 10.1.1.32:7789;
}
}
> For example, you can place an IP address, DRBD Filesystem and Apache under the
> control of heartbeat. On failure of the Primary node, those services will
> be "migrated" [started] on the Secondary node (and if properly configured,
> the Primary will be fenced out of existance); thus keeping your website live.
And what happens if your heartbeat box dies? ;)
--
Tomasz Chmielewski
http://wpkg.org