Thanks.

Frank, I don't want to sound offensive, but I hope that comment and 
document link were for an inexperienced casual reader reading this thread.

The point of my post is that failsafe mode should not have suggested that
I mount one of the slices that make up a mirror - unless the recent failsafe
software can detect it somehow and invalidate the mirror or otherwise cause
its synchronisation. 

Apparently it doesn't (as it never did before). While in failsafe I can't hook 
up 
the metadb replicas so there is little to no "nice and easy" way to fix a 
system 
with mirrored roots. 

The typical approach I saw in some blogs is to mount one of  the slices, then 
reboot into single-user mode mounting root as read-only. This hooks up the
metadb so I can do metadetach/metattach the secondary slice and thus make
the mirror valid. After this I reboot again (or mount root RW) to have the OS
working.

Kinda tedious if you have to pinpoint some config file while repairing a system 
and if it takes you about three reboots and a mirror rebuild each time.

I hoped that this situation is addressed in recent builds (since this unhelpful 
suggestion text appeared in failsafe scripts). So I checked it - and it ain't 
fixed.

I wanted to draw attention and get this problem fixed either way - by removing
the failsafe's suggestion or preferably by getting failsafe to mount 
metamirrored
roots :)

//Jim
-- 
This message posted from opensolaris.org

Reply via email to