Paul Krizak a écrit :
iSCSI boot enablement != iSCSI HBA. All that will do is allow your NIC
to mount an iSCSI device as a boot drive and load the first few blocks
from it (i.e. a bootloader). When the kernel actually loads into memory
and boots, it will need to be able to do the following before it can
actually load the OS past the initrd:
1. Start up the NIC
2. Run DHCP (or set a static IP)
3. Load the iSCSI daemon
4. Initiate a connection to the iSCSI shelf which should create
/dev/<something> (I think...been a while since I did iSCSI)
5. Mount the iSCSI LUN (/dev/something) as root.
Whereas in the situation with an iSCSI HBA (as we've all been
recommending) the kernel only needs to do the following:
1. Load the driver for the iSCSI HBA which creates /dev/<whatever>
2. Mount root on /dev/<whatever>
Obviously the second method is much better supported since it's exactly
the same as booting from a locally-attached disk or a FC SAN or something.
I see. But I'm stuck with a spiny situation (see
http://thread.gmane.org/gmane.linux.redhat.release.rhel5/5863), since I
can only get my HBA to connect its both ports to the same Blade
interconnect, making it non-redundant, while my onboard NICs can have
one port on one interconnect, and the other on another interconnect,
allowing for redundancy.
Any ideas?
Regards,
Ugo
_______________________________________________
rhelv5-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/rhelv5-list
