Ignite recovery from an Ignite cluster doesn't work

[Ivan Rodriguez Sanchez]

Hi all,

I have configured a Serviceguard 11.19 cluster with an Ignite package, following the whitepaper "Configuring an Ignite-UX Server under HP Serviceguard".
The package works correctly, it exports the ignite folders and add the floating IP.
I made an ignite from a server with make_net_recovery without any problem.
But the problem appears when I try to restore that server, I try the restore from the EFI shell, using a dbprofile, but I get a timeout when the client tries to get the files size.

I tried to halt the ignite package, mount and export manually the ignite folders (/var/opt/ignite, /etc/opt/ignite and /opt/ignite/boot), and add the floating IP manually, and I get the same problem.

I tried to manually mount and export the folders, and substitute the physical lan interface IP with the floating IP, and then the restore works. So I think the problem appears only when I try the restore from a floting IP over the physical lan.

Anyone has any idea to solve the problem?

HPUX 11.31
SG 11.19
Last version of Ignite

[Scot Bean]

I run with a similar setup.

So you are saying that the tftp partially fails with the EFI dbprofile lanboot?

Check /etc/inetd.conf. There is a "-s" option for serviceguard.

I have
tftp dgram udp wait root /usr/lbin/tftpd tftpd\
-s\
/opt/ignite\
/var/opt/ignite