hi ya charles
yes.... sometimes, if a master disk died... there was no slave...
on that ide controller... need to test for this case...
normally....
ide0 = hda + hdb
ide1 = hdc + hdd
simple case...
primary hda ( master on ide0 )
mirror hdc ( master on ide1 )
-- ie...dont use hdb .... use 2 ide cables...
if you lose hda due to disk crash....
data should still be intact on hdc...unless that data too
has been already mirrored... and erased...
- i dont like automated mirroring...cause it should erase data on
the mirror too that you wanted to retreive...
- daily or hourly mirrors are good... but not automated...
( my silly preference ... and nope...my preference changes
( for real-time apps )
or w/ striping
(primary) md0 = hda + hdd ?? vs hda + hdc
(mirror) md1 = hdc + hdb ?? vs hdb + hdd
if you lose the "master disk" (hda) of the primary data --
we all assume you can get it off the mirror ( hdc + hdb )...
but the big problem is the slave(hdb) might not be accessible
since hda is offline... ( bad ide controllers )
so the trick should be....
- slave ( hdb, hdd ) disks should be accessible even if there is nomaster...
different ide controllers behave differently...
- test it thoroughly by powerdown the ide with hdparm
and or pull the ide cable...
- write fresh new data with each power up/down cycle
have fun
alvin
On Tue, 17 Apr 2001, C. R. Oldham wrote:
> > -- question is... if the "master" disk dies...sometimes
> > the slave dont exist either... than you're sol...
>
> But you probably don't want to put disks on both the master and slave
> channels anyway since it will limit your throughput--only one can be
> accessed at any one time.
>
> --
> / C. R. (Charles) Oldham | NCA-CASI \
> / Director of Technology | Arizona State University \
> / [EMAIL PROTECTED] | V:480-965-8703 F:480-965-9423 \
>
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to [EMAIL PROTECTED]
