Recomendation for home server with centos: raid/file system

E

Euphoria

New Member
Reaction score
0
Hi all.

I actually have a Dell T620 server with 1x500gb hdd and 4x3000Gb hdd Raid10. Have installed Ubuntu server 14.04 on the 500Gb hdd and have personal file on the raid (movies, picture, documents, etc...) All got Ext4 format. the server are full and im buy other 4x3000Gb hdd for upgrade the raid and want to change the SO to Centos (more long time support). I use it with plex.tv, samba for sharing with family, pxeboot, and on the future make some virtual machine and counter server for friends.

- With Raid you recomand me? my option are 0,1,5,10,50 (raid 10 if 2 from same array crash lost data and with raid 5 or 50 1 drive fail but maybe i can use 1 drive for hot swap with raid 5 or 50)
- do i need to put 1 old drive in one bay and one new drive or just add the 4 new drives?
- with format for S.O and Raid? maybe lvm2 with some new format and the next time need upgrade raid dont need to erase all first.

Any recomendation are welcome!

Thx.
 
For a home server like that, basically just storing tons of file for read access, I'd probably go with RAID 5 or 50 ...as you can extend it easily down the road if need be (depending on which RAID controller you have that PE).

Maybe consider a tiny boot volume for the *nix host...bootable USB or if it has the bootable SD Card port...use one of those...think you can get up to 8 gigs...maybe larger. Ubuntu server should fit on that? (haven't looked to see what minimum partition sizes required is). This way you could built 2x USB drive copies or SD Card copies...if the first one fails, just do a quick swap.

Or just carve out a partition on the big RAID volume you're building. Normally I prefer the host OS on it's own separate spindle..for performance purposes, but for just a home server doing sharing..putting it on a slice of the big RAID volume would be fine.
 
Euphoria said:
Hi all.

I actually have a Dell T620 server with 1x500gb hdd and 4x3000Gb hdd Raid10. Have installed Ubuntu server 14.04 on the 500Gb hdd and have personal file on the raid (movies, picture, documents, etc...) All got Ext4 format. the server are full and im buy other 4x3000Gb hdd for upgrade the raid and want to change the SO to Centos (more long time support). I use it with plex.tv, samba for sharing with family, pxeboot, and on the future make some virtual machine and counter server for friends.

- With Raid you recomand me? my option are 0,1,5,10,50 (raid 10 if 2 from same array crash lost data and with raid 5 or 50 1 drive fail but maybe i can use 1 drive for hot swap with raid 5 or 50)
- do i need to put 1 old drive in one bay and one new drive or just add the 4 new drives?
- with format for S.O and Raid? maybe lvm2 with some new format and the next time need upgrade raid dont need to erase all first.

Any recomendation are welcome!

Thx.
Click to expand...

To make things a bit easier, I would just do a RAID5. However, Backup your /etc/mdadm/mdadm.conf file. I would look into setting up the raid so that it emails you if a drive dies. It could email you a RAID fitness report daily for that matter.
 
Thx soo much for help!

1º My raid is my backup :( im read thats is not the best, maybe i can also have some simple drive and sync some importants folders.

2º My server have 2 inside sd card (normaly for vmware install) but i made speed test and bad result.
Code: 
user@server:~$ sudo hdparm -Tt /dev/sda

/dev/sda:
Timing cached reads:   14590 MB in  2.00 seconds = 7301.14 MB/sec
Timing buffered disk reads: 416 MB in  3.01 seconds = 138.35 MB/sec
user@server:~$ sudo hdparm -Tt /dev/sdb

/dev/sdb:
Timing cached reads:   13128 MB in  2.00 seconds = 6569.09 MB/sec
Timing buffered disk reads: 782 MB in  3.00 seconds = 260.53 MB/sec
user@server:~$ sudo hdparm -Tt /dev/sdc

/dev/sdc:
Timing cached reads:   11592 MB in  2.00 seconds = 5799.56 MB/sec
Timing buffered disk reads:  30 MB in  3.02 seconds =   9.94 MB/sec
sda are the 500gb for S.O (im no is big but is the one come with the server) sdb are the actual raid 10 and sdc is ext4 sdcard. I prefer to keep the SO and files/raid separed.

4º i forget to say is hardware raid (perc H310) and dosent finde the file you say but dell server have idrac and i can read the manual and made some config for email when something fail. I keep the idea :)

Thats whats i do it:
- 500Gb hdd for SO (centos) too slow for the sd option
- 7 hdd for raid 50, more speed that raid 5 and not lose soo much.
- keep 1 hdd if someone fail for raid.

With format you recomend me for the raid? some lvm2 with ext4 or some new format (xfs,btrf,etc...)

Thx. again
 
My opinion is this:

We are not talking a major corporation backup scheme here. Therefore, KISS (Keep It Simple Stupid). Setup a RAID5 and be done with it. Keep in mind that the drives will eventually die. So, Sure - have a hot spare if you want. But do not make it so overly complicated that you are back here looking for help when a drive dies. The bottom line should be that you have a RAID to use with some redundancy and any maintenance is going to be somewhat easy to do.
 
I receive RAID data recovery projects frequently because of multiple drive failures. RAID is absolutely not an alternative to backup.
 
lcoughey said:
I receive RAID data recovery projects frequently because of multiple drive failures. RAID is absolutely not an alternative to backup.
Click to expand...

Ive had them before also. The multiple failures are usually because they never noticed the first disk had died and now the second one has. Normally, This was because they laid off their in house I.T. and depended on a 'break fix' approach.

IMHO, There is no fool proof way of securing you data. Everything has its drawbacks.

I use a RAID5 for my business and just monitor it closely. For over 8 years it has not let me down. If I were a business I would still do RAID but would incorporate offsite NAS or cloud.
 
Here are a few scenarios for multiple drive failure (or just the need for a backup):

1. RAID hits bad sectors on drive A, then while rebuilding a new drive A, drive B hits bad sectors and goes offline too
2. Drive A bows circuit board and causes a surge damaging all the other circuit boards too
3. Fire or water damage takes out entire unit
4. System theft
5. Physical system damage from being bumped or dropped
6. Accidental or intentional deletion of data
7. External power surge
8. Power surge from power supply
9. Virus issues (ie Cryptolocker)
10. and the list goes on
 
Thx. all for help!!! i keep all the recomendation and start the setup this week when i got time.

thx. you.
 
You must log in or register to reply here.

Similar threads

Metanis
[REQUEST] Yet another Optane horror story
Replies
2
Views
369
labtech
labtech
Big Jim
ideal drive capacity and RAID type
Replies
6
Views
646
Big Jim
Big Jim
Big Jim
Office storage upgrade and Kerio migration
Replies
13
Views
1K
Big Jim
Big Jim
Big Jim
Dell T320 Server and RAID question
Replies
18
Views
2K
nlinecomputers
nlinecomputers
Big Jim
Another server question
2
Replies
26
Views
3K
Big Jim
Big Jim
Back
Top