1 |
Benjamin MALYNOVYTCH wrote: |
2 |
|
3 |
>********************************** |
4 |
>JK> Maybe you can just use two A records in your dns. |
5 |
> |
6 |
>JK> Both point to one server, and both servers keep an eye on the other with |
7 |
>JK> hearthbeat, |
8 |
> |
9 |
>JK> As soon as one of the servers gets unresponsive the other claims the ip |
10 |
>JK> as a secundary ip. |
11 |
> |
12 |
>JK> ifconfig eth0 add otherip. |
13 |
>JK> // not forget to arp for it! |
14 |
> |
15 |
>JK> And as soon as the hearthbeat finds the other server to be back online, |
16 |
>JK> it unsets the ip, and lets the other know that it can add the ip again. |
17 |
> |
18 |
>JK> Not so hard, just a couple of shell scripts afaik. |
19 |
> |
20 |
>JK> greets |
21 |
>JK> Jan |
22 |
> |
23 |
>********************************** |
24 |
> |
25 |
>This sounds pretty good, but the customer is running mysql with data |
26 |
>changing all day long. That means that I would have two seperate db, |
27 |
>one on each machine, without being able to merge/keep uptodate those |
28 |
>dabatases :/ |
29 |
> |
30 |
>Best regards, |
31 |
> |
32 |
>Benjamin MALYNOVYTCH |
33 |
> |
34 |
> |
35 |
> |
36 |
> |
37 |
That would be a different problem, |
38 |
|
39 |
You should check mysql failover and master->slave replication. Mysql |
40 |
replication should be handled as a different "project" as its something |
41 |
of a bitch to setup reliably. |
42 |
|
43 |
greets |
44 |
Jan |
45 |
-- |
46 |
gentoo-cluster@g.o mailing list |