Prev: Solaris licensing after Oracle takeover
Next: How to tell the Solaris Version
From: jay on 15 Apr 2010 10:50
greetings,

i have a test v120 for a little while. so i can act
w/ impunity. i'm using zfs on one or two things
but i've never tested what happens during a
failure so that's my project.

installed solaris 10, using zfs for root file system.
once i got up, i mirrored / to the second disk.
halted, pulled second disk out, rebooted. the
machine came up, IIRC nothing seemed wrong
until i did zpool status which showed the missing
disk. (there might have been a notice that things
were amiss but if so,i overlooked it. that's okay,
though, this is the kind of question i was trying
to answer. push comes to shove i can always
write a startup script to check zpool status.)

so far, so good. i'm officially all happy. now
i want to re-make the mirror. next i made a
mistake.
zpool online rpool _missingdisk_
i __thought__ i had re-seated the disk, turns out
i hadn't. the problem is that command hung the
machine. i had to reboot it.

that seems very bad, even though it was my
fault. (i know now that what i should have done
was
zpool detach rpool _missing_
zpool attach rpool _old_ _reseated_
).

okay, it's my fault not reseating the disk,
but the punishment seems excessive for
the crime. i didn't try reseating the disk
while it was hung. but the OS hung when
i think it should have reported an error.

did i have options? i couldn't get in through
the console.

j.
 | 
Pages: 1
Prev: Solaris licensing after Oracle takeover
Next: How to tell the Solaris Version