Hi,
we (want to) use HA-Linux to migrate a Linux-VServer from a master server
to a backup server in case of a failure. The image of the vserver is in a
LVM volume group which is located in a SAN. Here is the relevant part
of our haresources script:
master SendArp::x.x.x.x/eth0 \
LVM::san_vg \
Filesystem::/dev/san_vg/db-host_lv::/vservers/db-host::ext3::rw \
Filesystem::/dev/san_vg/db-host-data_lv::/data/db-host::ext3::rw \
vserver::db-host \
MailTo::[EMAIL PROTECTED]::[ha]db-host
The takeover of the vserver IP address is managed by the vserver script.
As far as I can see, SendArp should propagate the new mac address to the
caches of the network equippment. But the latter does not work. After a
takeover the vserver does not ping anymore. After sending a gratious arp
(with arping) on the backup master, all is well. So it seems, that SendArp
does not work correctly or my haresource script is wrong.
I sniffed on the network and noticed, that the target mac address of the
SendArp arp packages is 00:00:00:00:00:00, while the target mac address
of other gratious arp packages is ff:ff:ff:ff:ff:ff. Maybe this is the cause
of the flaw?
We are using heartbeat 2.1.2-1 on debian etch.
Thanks for any hint! SY Th. Gebhardt
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
