Looking For RAID Hardware/Software Advice v1

[Lisa Kachold]

I have built and maintained linux production servers under both hardware RAID 5 on HP MSA's, 
and hardware RAID 1+0 on both HP Proliant's and Dell 2950's/1950's (with a variety of software disk management [depending on the server farm standards]).
   
I have also configured LVM and md under SATA or iSCSI systems in both RAID 5 and RAID 10.

[I have also built Solaris servers under RAID 5/RAID 10 using SVM, and later replaced with N1 over multi-path I/O on Sun 2450's (zfs); and worked with NetApps/Redhat (xfs).]

I am a proponent of md/LVM over hardware RAID because Linux md does not handle bad block relocation; I love simply rebuilding the array, especially since drives heat up, get torged by power, and are simply not build with QA these days.

Since disk is so incredibly cheap these days (by comparison in your lower level [non enterprise] solution), the popularity of RAID 5 is offset by the full mirror and rebuild protection of RAID 1+0.  

Here's the technical descriptions of each level:

RAID 0 is not redundant at all but offers the best
throughput of all levels here. Data is striped across a number of
drives so read and write operations take place in parallel across
all drives. On the other hand if a single drive fail then
everything is lost. Did I mention backups?
RAID 1 is the most primitive method of obtaining redundancy
by duplicating data across all drives. Naturally this is
massively wasteful but you get one substantial advantage which is
fast access.
The drive that access the data first wins. Transfers
are not any faster than for a single drive, even though you might
get some faster read transfers by using one track reading per
drive.

Also if you have only 2 drives this is the only method of achieving
redundancy.
RAID 2 and 4 are not so common and are not covered
here.
RAID 3 uses a number of disks (at least 2) to store data
in a striped RAID 0 fashion. It also uses an additional redundancy
disk to store the XOR sum of the data from the data disks. Should
the redundancy disk fail, the system can continue to operate as if
nothing happened. Should any single data disk fail the system can
compute the data on this disk from the information on the redundancy
disk and all remaining disks. Any double fault will bring the whole
RAID set off-line.

RAID 3 makes sense only with at least 2 data disks (3 disks
including the redundancy disk). Theoretically there is no limit for
the number of disks in the set, but the probability of a fault
increases with the number of disks in the RAID set. Usually the
upper limit is 5 to 7 disks in a single RAID set.

Since RAID 3 stores all redundancy information on a dedicated disk
and since this information has to be updated whenever a write to any
data disk occurs, the overall write speed of a RAID 3 set is limited
by the write speed of the redundancy disk. This, too, is a limit for
the number of disks in a RAID set. The overall read speed of a RAID
3 set with all data disks up and running is that of a RAID 0 set
with that number of data disks. If the set has to reconstruct data
stored on a failed disk from redundant information, the performance
will be severely limited: All disks in the set have to be read and
XOR-ed to compute the missing information.
RAID 5 is just like RAID 3, but the redundancy
information is spread on all disks of the RAID set. This improves
write performance, because load is distributed more evenly between
all available disks. Parity data is rotated across all disks so
total net storage equals all disks minus 1.
RAID 6 is similar to RAID 5 except that there is twice the
redundancy and the array can survive 2 failed drives.
Parity data is also rotated across all disks so
total net storage equals all disks minus 2.


There are also hybrids available based on RAID 0 or 1 and one other
level. Many combinations are possible but I have only seen a few
referred to. These are more complex than the above mentioned
RAID levels.
RAID 01 combines striping with duplication
as mirrored arrays of striped arrays
which gives very high transfers combined with fast seeks as well as
redundancy. The disadvantage is high disk consumption as well as
the above mentioned complexity.
Also a single disk failure turns the array into RAID 0.
RAID 1+0 combines striping with duplication
as striped arrays of mirrored arrays
which gives very high transfers combined with fast seeks as well as
redundancy. The disadvantage is high disk consumption as well as
the above mentioned complexity.

When you are scrounging disk against money you are sure that two of your disks are not going down, so choose not to use 4 disk 1+0.  But believe me, THEY DO, and that's it!   Also RAID 1+0 out performs in a fast good way.  What use is a huge cheap slow system?

Full discussion available here that includes a complete analysis of md systems, and all the types of data loss that generally occur.

http://linas.org/linux/raid.html

www.Obnosis.com |  http://wiki.obnosis.com | http://hackfest.obnosis.com (503)754-4452
PLUG HACKFESTS - http://uat.edu Second Saturday of Each Month Noon - 3PM

Date: Mon, 19 Jan 2009 07:30:08 -0700
From: mark at phillipsmarketing.biz
To: plug-discuss at lists.plug.phoenix.az.us
Subject: Re: Looking For RAID Hardware/Software Advice

Eric
Thanks for the summary, and thank-you to everyone for their ideas.
Based on NewEgg prices, here is some more information:

Option A
Single Disk IDE Drive - 500 GB and backups, keep OS on existing drive = $69.99

Use existing controller and just add another drive. No redundancy

Option B
RAID10 with 500 GB backup capacity and redundancy, keep OS on existing drive = $179.97
2 500 GB SATA2 Drives, new SATA2 controller


Option C
RAID10 with 750 GB backup capacity and redundancy, keep OS on existing drive = $239.97
Two 750 GB SATA2 Drives, new SATA2 controller

Option D
RAID5 with 1,000 GB backup capacity and redundancy, keep OS on existing drive = $239.97

Three 500 GB SATA2 Drives, new SATA2 controller

I am leaning towards Option C based on less power consumption with fewer drives. However, I have to rethink my budget...
After some more reading, I am a little confused about the debate between RAID5 and RIAD10. I am interested in the group's opinions on which is better - RAID 5 or RAID 10 and why? What experiences have you had regarding installation, maintenance, and fixing problems? I am running Debian testing.

Thanks!
Mark
On Sat, Jan 17, 2009 at 6:35 AM, Eric Shubert <ejs at shubes.net> wrote:

Mark Phillips wrote:

> I am running out of room for my backups. I use backuppc and I have

> almost filled a 150GB drive with backups from 7 computers, and I need to

> add another 2 computers to the set. I have an old Dell Poweredge 1300

> server (Pentium III 550 Mhz, 500 MB RAM, PCI 33.3Mhz) that I could turn

> into a backup server. I am looking for suggestions/thoughts on how to

> set this up. I need to keep the cost down as much as possible; under $150.

>

> My initial thoughts:

>

> * Keep current 72 GB drive for OS (debian testing, about 68% full)

> * Add two 500 GB SATA drives and a PCI SATA controller ~$130

> * Software RAID and LVM for the two drives

> * Move current 150 GB of backups to the RAID

> * Backuppc now runs on this machine and slowly fills up the RAID

>

> My questions:

>

> 1. Should I keep the 72 GB drive for OS, or put it on the RAID?

>

> 2. I can add another CPU (P III 550 MHz) processor to the box - is it

> worth the effort to find one? I found one source for $5/CPU, I just need

> to find the heat sink and mounting hardware. Will this improve performance?

>

> 3. The box has a built-in SCSI 68-pin Ultra2/wide bus/controller, but

> SCSI drives are more expensive, at least from a cursory google search.

> Is this correct? I don't think I can use SCSI drives within my budget

> constraint.

>

> 4. Would upgrading the memory to 1GB improve performance - top shows:

> Mem: 646676k total,      639300k used     7376k free,      64548k buffers

> This would add another ~$60 to my cost.

>

> 5. Should I look at hardware RAID cards - they seem very cheap, so

> perhaps software is better?

>

> 4. Does this plan make sense, or is there a better way to proceed for

> about the same cost?

>

> Thanks!

>

> Mark

>



Good replies, all. To sum things up, I think a SATAII PCI card (2 or 4

port) and 2 drives is all the HW you need to add to the backup box you

currently have. Set up the drives with SW RAID-1 (mirrored) and you're

good to go. Migrate the data to the raid device, and keep the OS on the

existing drive.



With KeepItSimpleStupid in mind, I recommend using RAID-1 as opposed to

RAID-5. With the price of drives these days, the additional space you

get with RAID-5 isn't worth the headache you'll get when there's a

problem. With RAID-1, each drive can be mounted (and used) individually

if necessary. Not so with RAID-5.



--

-Eric 'shubes'



---------------------------------------------------

PLUG-discuss mailing list - PLUG-discuss at lists.plug.phoenix.az.us

To subscribe, unsubscribe, or to change your mail settings:

http://lists.PLUG.phoenix.az.us/mailman/listinfo/plug-discuss



Windows Live™: Keep your life in sync.  Check it out.
_________________________________________________________________
Windows Live™ Hotmail®: Chat. Store. Share. Do more with mail. 
http://windowslive.com/explore?ocid=TXT_TAGLM_WL_t1_hm_justgotbetter_explore_012009
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.PLUG.phoenix.az.us/pipermail/plug-discuss/attachments/20090119/cbe01a90/attachment.htm