On 09/01/06 01:59, Dennis Clarke wrote:
I have SVM metadevices defined in the usual way.
# metattach d7 d27
d7: submirror d27 is attached
metattach: mars: /dev/md/dsk/d0: not a metadevice
#
That last message makes no sense. I already have d0 defined and setup
fine. So .. where did that weird message come from ?
Have a look at your /etc/lvm/md.tab. I have a hunch that something
updates that for you nowadays (never used to).
First place I look. As far as I know the /etc/lvm/md.tab file never gets
updated and its typically hand edited by the sysadmin.
That's correct, I believe; SVM updates the /etc/lvm/md.cf and mddb.cf
files, but not md.tab.
The output of metastat would probably be useful, if you want somebody to
take a look, but (as you ask below) the code is open. Look in
usr/src/lib/lvm/libmeta/common/, usr/src/cmd/lvm/util/metattach.c, and
(possibly) usr/src/uts/common/io/lvm/ if you don't find it there.
Start by finding the error text "not a metadevice" in
libmeta/common/meta_error.c. That tells you some code path is throwing
MDE_NOT_META.
It might also help to know with finer granularity when the problem was
introduced; there were SVM putbacks into snv_40, snv_42, snv_43, and
snv_44.
--Mark
In my case its real real simple :
d10 1 1 /dev/dsk/c0t0d0s0
d0 -m d10
d20 1 1 /dev/dsk/c0t1d0s0
d11 1 1 /dev/dsk/c0t0d0s1
d1 -m d11
d21 1 1 /dev/dsk/c0t1d0s1
d17 1 1 /dev/dsk/c0t0d0s7
d7 -m d17
d27 1 1 /dev/dsk/c0t1d0s7
Been doing this forever ( head rolls here ) and this is the same config that
was on this machine when it ran snv_36 last week.
This new message is just plain weird and it may be time to go to the sources.
If they are open.
Dennis
_______________________________________________
opensolaris-discuss mailing list
[email protected]
_______________________________________________
opensolaris-discuss mailing list
[email protected]
