I've been playing with boot-from-multipath in the 5.1 beta today, and
I've very pleased to see that things finally seem to be mostly working,
using the "mpath" boot-time option in anaconda.
I'm using an IBM DS4400, so there are a few glitchy things that I'm
looking for insight on. For starters, this is an active/passive array,
and it looks like anaconda defaults to trying more of a round-robin
approach during install. So, every time it accesses the disk, it has to
error/timeout on each disk several times, causing waits of 1 minute or
two between each step when partitioning a drive. This is *painful*. Is
there any way (even a command behind the scenes) to tell it so pick one
path and stick to it?
Also, since I want to use a failover config and mpath_prio_rdac, but the
multipath.conf is loaded into the initrd, is there a way to provide
anaconda with an alternative multipath.conf to use when it makes the
initial ramdisk. I can change it after the fact, but the errors and
disk failures that popped up in the meantime and during the initrd
update weren't at all pretty, so I'd rather have the correct settings in
place before the first boot.
Hope this is all somewhat clear. Any insights would be greatly
appreciated.
Thanks
DC
_______________________________________________
rhelv5-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/rhelv5-list
