[2 days left] What’s wrong with your cloud strategy? Learn why multicloud solutions matter with Nimble Storage.Register Now


What would happen to hdd and data in case of this raid 5 configuration ?

Posted on 2012-03-11
Medium Priority
Last Modified: 2016-12-08
My  storage server has 12 HDDs.

and  RAID5 group X2

RAID group1 is on the hdd 1 - hdd 7
RAID group2 is on the hdd 8 - hdd12

According to control panel,
for each usage,
hdd 1 is system
hdd 2 is system
hdd 3 is data
hdd 4 is global hot spare
hdd 5 is data
hdd 6 is data
hdd 7 is data

hdd 8 - hdd 12 are data.

My question is  this "system" meaning.
Normally, What do you predict this "system" mean ? RAID5 parity ?
Can global hot spare become alternative hdd instead of the not only failured hdd of RAID5 group1 but  RAID5 group2 ?
If hdd 7 and hdd 6 would be broken at the same time, the data can be recovered automatically after replacment of new hdds ?
what would  happen if hdd 1 and hdd 2 would be broken at the same time ?

Best regards
Question by:shouichi
Welcome to Experts Exchange

Add your voice to the tech community where 5M+ people just like you are talking about what matters.

  • Help others & share knowledge
  • Earn cash & points
  • Learn & ask questions
  • 2
LVL 37

Expert Comment

by:Neil Russell
ID: 37707116
In a RAID 5 array, the loss of ANY 2 disks would mean a TOTAL loss of the data and the array. You would nee to be paying specialist recovery companies to recover data for you.
LVL 37

Expert Comment

by:Neil Russell
ID: 37707123
So in answer to your question, if 1 and 2 or 6 and 7 die at the same time, you have lost your entire array. same if 1 and 6 or 3 and 7 or ANY TWO DISKS IN A SINGLE RAID5 ARRAY
LVL 20

Expert Comment

ID: 37707238
A general answer as you have not given the make of the server / raid controllers etc :

You have two raid arrays, 1-7 and 8-12.
As mentioned above if you loose 2 disks in either of these array groups then you would have a fail.
If disks 1 and 6 went bad then that raid group (disks 1 - 7 and everything stored on them would be gone).  the other set - disks 8-12 would carry on working fine.
And the same thing the other way round.

Disks 1,2 are called 'system' and disks 3,5,6 and 7 are called Data.
The likely reason for this is your system is partitioned so you have a first drive for your OS (for example a C drive for windows files) and then the rest of that group is data (maybe D drive for user data) and the other group would be anothe drive again (maybe E:)
This is quite a strange way of doing it.  I wouldhave made those into two seperate raid groups - but then there may have been a reason for it.

Disk 4 is a hot spare.  As long as it has been setup correctly it will become the first failed disk.  So if disk 12 goes wrong, disk 4 will take over.  The data that would have been on disk 12 gets rebuilt off the other disks in its array. After an hour or so the rebuild is finished and you can then look at buying a new disk for the slot 12 disk and making that the new hot spare.

It is unlikely that two disks would go on the same day so this setup should be able to survive a disk fail without even rebooting.

Note - you should also keep tape backups in case this did fail - or in case of corruption or files being deleted by accident etc.
LVL 47

Accepted Solution

David earned 2000 total points
ID: 37707522
Well, even with RAID5 you can lose data in event of a single disk failure under numerous scenarios.  Removing the firmware bugs; multiple failure scenarios; power-loss; the most common causes of data loss are

1. If your controller doesn't automatically do consistency checks and sweep/repair unreadable blocks, then if you have a bad block on disk "A", and disk "B" fails, then you WILL lose data corresponding to a chunk of data that aligns with the data on disk A.

Example, say chunk size is 64KB and you have 5 disks in a R5 (to keep math easy), and you lose physical block 0 on the first disk.  Then that means you lose 64KB x 4 = 256KB of contiguous data.  So real world, your system will lose all partitioning, and won't boot, and you'll lose pretty much all the directory information.  [Despite the fact this is RAID5, and you only lost 1 disk. With RAID5, then there is only one chunk of parity data - the redundant stuff, per any block; so losing a drive means you lost the parity data.  Having a bad block means you lost the redundancy, but then if all 5 disks are working, you would have to have the bad block in the 64KB area on two different drives, which is low odds, but not impossible .. I've seen it]

This is why people go to RAID6 for important data, and why RAID-anything is never a substitute for backing up.

(Premium controllers that have battery backups for write cache and do background scrubbing and RAID6 are worth the extra money if the data is important)

2. Other common scenario is that people don't realize that most RAID controllers are architected for enterprise class drives, especially with SATA disks, and almost always in RAID5+RAID6.  So get  desktop class drives and they don't have the right timing settings.  If a disk goes into deep recovery, the RAID controller may fail the drive.  If you have not been scrubbing your drives to do data consistency check/repairs, then even choosing the wrong drives almost guarantee data loss.

The "HDD is system" type stuff is a function of your RAID controller/software.  RAID controllers in general have no idea what the data is, so the fact that it says "system" means you have some software that was told, either by the person who set it up, or some management software that extrapolated it.

It is also incorrect to assume that a global hot spare CAN work for any RAID group.  This is a function of your architecture.  In your situation, simply no way can anybody tell you YES or NO, based on what you supplied.

If, for example, you have 2 chains of disks, where each chain is attached to it's own port on the adapter, and half the disks are on each, and the HS is on bottom chain along with 2nd LUN ..

Then some RAID controllers won't let any disk on the bottom chain be associated with a LUN built on the top chain.   Many dual-ported controllers are effectively two DIFFERENT controllers, that share nothing, so it would be like the disks on the bottom chain is attached to a different enclosure and different controller.  No way could it possibly serve as a spare to a different controller card, since there is no syncing of data between the two different controllers.

It would be foolish to make an assumption to tell you definitive yes or no w/o knowing exact make/model of controller and details of physical arrangement.

Featured Post

On Demand Webinar - Networking for the Cloud Era

This webinar discusses:
-Common barriers companies experience when moving to the cloud
-How SD-WAN changes the way we look at networks
-Best practices customers should employ moving forward with cloud migration
-What happens behind the scenes of SteelConnect’s one-click button

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

Patch Management is administrated by page 129 of  the following document:     http://support.dell.com/support/edocs/software/smdmc/11/en/ug/ug.pdf However, there are issues with Patch Management with FTP (for instance, a slow Internet connectio…
Every server (virtual or physical) needs a console: and the console can be provided through hardware directly connected, software for remote connections, local connections, through a KVM, etc. This document explains the different types of consol…
In this video, Percona Director of Solution Engineering Jon Tobin discusses the function and features of Percona Server for MongoDB. How Percona can help Percona can help you determine if Percona Server for MongoDB is the right solution for …
In this video, Percona Solutions Engineer Barrett Chambers discusses some of the basic syntax differences between MySQL and MongoDB. To learn more check out our webinar on MongoDB administration for MySQL DBA: https://www.percona.com/resources/we…
Suggested Courses

656 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question