[Gllug] disk problems

[Nix]

On Wed, 15 Mar 2006, Andy Smith prattled cheerily:
> On Tue, Mar 14, 2006 at 09:27:30PM +0000, Sean Burlington wrote:
>> Nix wrote:
>> > Everything seems to be stable now, but RAID it is, and because I want
>> > actual *robustness* I'm LVM+RAID-5ing everything necessary for normal
>> > function except for /boot, and RAID-1ing that.
>> 
>> RAID 5 = 3 hard disks + controller...
>>
>> I can't really justify the expense of that even though I have had a
>> couple of failures (and one or two learning experiences a while back)
> 
> Software RAID-5 is 3 hard disks.

Software RAID-5 is n+1 with wandering parity. n is most often two but
can be as low as 1 (in which case the second disk is identical to the
first, so it degenerates into RAID-1), or as high as you like.

In general if you get high enough it makes sense to go to RAID-6, so you
can have more redundancy and more simultaneously dying disks.  (Note
that you don't have to lose two disks *simultaneously* to get in trouble
with RAID-5; if you lose one, and then after you've fitted a new disk
you find bad blocks on the other during the massive read-and-write storm
which is RAID array recovery, you're still in serious trouble.)

>                                  Software RAID-1 is 2 disks and
> still gives you the same level of redundancy.

RAID-1 is N disks, mirrored, with arbitrarily high levels of
redundancy as a consequence. e.g., my site here:

Personalities : [raid1] [raid5]
md2 : active raid5 sdb7[0] hda5[3] sda7[1]
      19631104 blocks super 1.2 level 5, 64k chunk, algorithm 2 [3/3] [UUU]

md1 : active raid5 sda6[0] hdc5[3] sdb6[1]
      76807296 blocks super 1.2 level 5, 64k chunk, algorithm 2 [3/3] [UUU]

md0 : active raid1 sda5[0] hdc1[3] hda1[2] sdb5[1]
      56064 blocks [4/4] [UUUU]

md0 has nothing on it but /boot, but note that it's on *every* disk. Thanks
to

boot = /dev/md0
raid-extra-boot = mbr-only

in /etc/lilo.conf, *every* one of those disks is bootable, so I'll not
lose my system unless I have a *lot* of disks die (well, OK, two will do
the job as one of those RAID-5 arrays houses my root filesystem, but
even then I'll have my initramfs and emergency shell on there, so I can
start putting the arrays back together even without a root
filesystem...)

> If double the disks costs vastly more than your data and time is
> worth then I agree with you however I think that is the case for
> very few people.

Disks really aren't expensive these days, but RAID-5 is less than
`double the disks'; a 3-disk RAID-5 array costs you 1/3 of its
space in redundancy, which is generally *well* worth it. It's been
a while since I had many filesystems 2/3rd full :)

-- 
`Come now, you should know that whenever you plan the duration of your
 unplanned downtime, you should add in padding for random management
 freakouts.'
-- 
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug