From: Helmut Wollmersdorfer (helmut.wollmersdorfer_at_gmx.at)
Date: Mon 22 Aug 2005 - 15:25:55 BST
Herbert Poetzl wrote:
> On Mon, Aug 22, 2005 at 01:18:33PM +0200, Helmut Wollmersdorfer wrote:
>>With interfaces/0/name defined it seems to work now:
>>
>># ip addr ls
>>[...]
>>2: eth0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 1000
>> [...]
>> inet 192.168.0.22/24 brd 192.168.0.255 scope global eth0
>> inet 192.168.0.20/24 brd 192.168.0.255 scope global secondary eth0:0
>> inet 192.168.0.27/24 brd 192.168.0.255 scope global secondary eth0:vs27
> so you are telling me, that with an entry like this:
> inet 192.168.0.27/24 brd 192.168.0.255 scope global secondary eth0
> (note the missing :vs27 alias, which would be the default
> without specifying the 'name')
> does not work on your system, while the above (with alias) does?
No, without specifying the 'name' it works nice - without heartbeat
establishing some other aliases.
As I understand the IPaddr script of heartbeat, it uses ifconfig to
search a free name for ethx:<subdevice> and establish it. With heartbeat
alone, this will result in e.g. eth0:0. But there was some conflict with
vserver, as I had these 'ping-pong' failovers, and disconnections
shortly after ssh login into the vserver.
The truth is, that I cannot say exactly, what the reason was/is. That's
why I asked 'How ... handled exactly'. The docs say something like
'ethx:0' if not 'named', which will be a problem, if heartbeat does not
see it via 'ifconfig'.
Now I reconfigured (and recabled, isolated) my test cluster completely,
doing the tests step-by-step to avoid other reasons. Maybe I can tell
more in some hours.
Helmut Wollmersdorfer
_______________________________________________
Vserver mailing list
Vserver_at_list.linux-vserver.org
http://list.linux-vserver.org/mailman/listinfo/vserver