[Linux-cluster] IPv6 enabled: out of memory!

[Kadlecsik Jozsi]

Hi,

Something really strange happened here: adding one global IPv6 address to 
an ("external") interface of one node caused "out of memory" on *all* of 
our nodes.

All nodes have got two interfaces, one dedicated to GFS and the other is 
for servicing the users. We configured IPv4 addresses for GFS, otherwise 
all interfaces have got link local IPv6 addresses. Now we wanted to 
introduce services over IPv6 as well. Adding an IPv6 address to one 
interface and rebooting the machine for a clean start caused the disaster.

Is it an unsupported configuration? Should we convert to IPv6 exclusively 
for GFS? (As we had trouble with IPv4 multicast for GFS, that'd be a major 
modification.)

What do you suggest, how could we enable IPv6 for services but keeping 
IPv4 for GFS?

We run cluster-2.01.00 over 2.6.23-rc4 kernel. 

Best regards,
Jozsef
--
E-mail : kadlec at sunserv.kfki.hu, kadlec at blackhole.kfki.hu
PGP key: http://www.kfki.hu/~kadlec/pgp_public_key.txt
Address: KFKI Research Institute for Particle and Nuclear Physics
         H-1525 Budapest 114, POB. 49, Hungary