you REALLY dont want to build a raid across the same physical harddrive. this is
a bad idea in terms of performance, and what it could do to the drives life
(just think of the heads popping back and forth between two partitions for every
chunk written).
if you can get a second disk in there, then do that. otherwise, an UPS will
prevent fs corruption as well as anything else. of course thats true x86 or
otherwise.
also, you REALLY should upgrade the kernel to the newest alpha drivers and raid
tools. dump the md commands that are on the system now.
try ftp.us.kernel.org/pub/linux/daemons/raid/alpha or something like that.
al
ps- what do you mean they wont let you change? you can have two diff. mail
clients on your machine.
Phillip Jaenke <[EMAIL PROTECTED]> said:
> I do apologize for the use of Exchange. While I may be the guy in charge
> of all things unix, excepting MAS90 (thank gods), they won\'t let me
> switch over from exchange. Anyways...
>
> I\'m working on some production machines that need to be at least
> *somewhat* fault tolerant. (for x86 hardware. *puke*)
>
> I figured I could at least prevent filesystem corruption by using a
> RAID1 setup. These are *tiny* rackmount PCs on \"industrial\"
> motherboards. So I only have room for the NIC, the modem, the Equinox
> SST8P, and the \"motherboard,\" if it could even be called that. (can you
> tell I *hate* x86?)
>
> Well, I\'m using Debian 2.1 (slink), 2.0.36 monolithic and minimized. (No
> extra drivers or debugging stuff.) And I\'ve done the RTFM thing.
>
> And I\'m positively lost. So, I\'m hoping someone out there can help. I\'ve
> got the syntax of mdadd and such figured out. (I\'m not using raidtools,
> FYI.) Here\'s how it\'s setup.
>
> Industrial Automated Computers AP-540 motherboard
> Pentium 233MMX, 32M EDO
> Equinox SST8P PCI 8 port Serial Board (using Equinox drivers, separate
> modules - not a kernel patch.)
> USR Sportster 33.6k Internal ISA FaxModem (Jumpered, ttyS1)
> 3Com 3C905 \"Vortex\" 10/100bT NIC
> 4.3G Quantum Fireball SE IDE HDD
>
> The 4.3G is partitioned like this:
>
> 1971.2M
>
> hda1 - 750M - Linux Native, ext2
> hda2 - 1221M - Linux Native, ext2
> hda3 - 750M - Linux Native, ext2
> hda4 - 1447M - Extended
> hda5 - 1221M - Linux Native, ext2
> hda6 - 128M - Linux Swap
> hda7 - 128M - Linux Swap
>
> Here\'s what I\'d like to do:
>
> hda1 -> / -> Soft RAID1 -> hda3 -\\
> >-< NFS Mounted Update Server
>
> hda2 -> /usr -> Soft RAID1 -> hda5 -/
>
> So, in words:
>
> hda1, mounted on /, is mirrored via software RAID1 on hda3.
> hda2, mounted on /usr, is mirrored via software RAID1 on hda5.
> The machine updates automatically via crontab off an NFS mounted update
> server.
>
> So far, NFS is working. But I ran into a problem with RAID1. I *did*
> achieve RAID1 mirroring on /usr, but /usr would not mount.
>
> Is there a way to RAID1 both / and /usr, yet still be able to mount
> them, without adding a second drive?
>
> Thanks muchly.
>
> -Phillip R. Jaenke
> Unix Systems Administration - Unicent Technologies
>
<br>
<b>Warning</b>: Sybase: Server message: Changed database context to \'twig\'.
(severity 10, procedure N/A) in <b>features/sybase.db.inc</b> on line
<b>33</b><br>
--
