LinuxQuestions.org
Latest LQ Deal: Complete CCNA, CCNP & Red Hat Certification Training Bundle
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Newbie
User Name
Password
Linux - Newbie This Linux forum is for members that are new to Linux.
Just starting out and have a question? If it is not in the man pages or the how-to's this is the place!

Notices


Reply
  Search this Thread
Old 08-17-2012, 02:11 PM   #61
Aunnix
Member
 
Registered: Aug 2012
Posts: 126

Original Poster
Rep: Reputation: Disabled

Quote:
Originally Posted by suicidaleggroll View Post
You would want RAID1.

RAID0 is stripped, so anything you write is split between the two drives (half on each). This means that if you lose either drive, you lose everything on the entire array. There is no redundancy, however your array will be the sum of all parts, so two 750GB drives will give you a 1.5TB array.

RAID1 is mirrored, so anything you write is copied to both drives. This means that if you lose either drive, the system will fall back on the other seamlessly. The unfortunate part is because you're mirroring, the array will be the size of one drive (750GB).

Cool thanks. Yeah, I know most people are usually running at least 1TB drives, but I've been running on a 160GB drive for the last 4-5 years and have filled up maybe 50GB of it, lol. That is OS and all, hahah. I'm basically only going to be starting out with around 20GB of data to transfer to the server when it's set up. (perhaps 50GB if I keep the 25-30GB of music I just backed up for a friend of mine, hahah). But, that is also the reason I grabbed the 60GB drive so I can leave as much as possible of the 750GB intact. Right now, my windows machine is only seeing like 690-700GB on the HDDs... not sure if Linux will read them differently or not.
 
Old 08-17-2012, 02:19 PM   #62
Aunnix
Member
 
Registered: Aug 2012
Posts: 126

Original Poster
Rep: Reputation: Disabled
Quote:
Originally Posted by suicidaleggroll View Post
You would probably just set up a SAMBA share on the Linux box. On Windows it will just show up as a shared drive, which you can map to a local drive letter (eg: P:\ will be a remote mount of the Linux SAMBA share, which might be /home/winshare on the Linux system). At that point, you can play it on the Windows machine as if it was local. I have my home set up this way. One Linux server hosting all of the media and serving it through both SAMBA and NFS. I then have several client machines, both Windows and Linux, who remote mount the shares and can play the media as if it's local. Even iTunes has its archive set to the shared location, so any time I buy music through iTunes it automatically moves it onto the Linux box through the share.

THIS! lol. This sounds exactly like what I want to do for "file sharing" on my home network. SAMBA has been mentioned in another post I had online. Will this setup work with smart phones, consoles (such as ps3, or xbox), laptops, etc... they would all be connected wirelessly to the network.
 
Old 08-17-2012, 02:26 PM   #63
Aunnix
Member
 
Registered: Aug 2012
Posts: 126

Original Poster
Rep: Reputation: Disabled
Quote:
Originally Posted by theNbomr View Post
Your ISP may provide this information. I found out about an unpublished, unadvertised way of accessing this information using some backdoor access to, I think, my cable modem. I assume this is also where the ISP gets it's data. At any rate, shortly after my service was cut off, I implemented a primitive system in my server/firewall that makes use of iptables' throughput logging to keep tabs on my usage. It wasn't too difficult and is pretty low overhead. If you have a stand-alone firewall box, it might record such data.

--- rod.

I may want to look into this at some point (depending on what the ISP says). I have no stand alone firewall box (unless my router acts as one, lol) so I don't think I'm recording this info anywhere.
 
Old 08-17-2012, 02:45 PM   #64
Aunnix
Member
 
Registered: Aug 2012
Posts: 126

Original Poster
Rep: Reputation: Disabled
Quote:
Originally Posted by suicidaleggroll View Post
You would probably just set up a SAMBA share on the Linux box. On Windows it will just show up as a shared drive, which you can map to a local drive letter (eg: P:\ will be a remote mount of the Linux SAMBA share, which might be /home/winshare on the Linux system). At that point, you can play it on the Windows machine as if it was local. I have my home set up this way. One Linux server hosting all of the media and serving it through both SAMBA and NFS. I then have several client machines, both Windows and Linux, who remote mount the shares and can play the media as if it's local. Even iTunes has its archive set to the shared location, so any time I buy music through iTunes it automatically moves it onto the Linux box through the share.

Quote:
Originally Posted by Aunnix View Post
THIS! lol. This sounds exactly like what I want to do for "file sharing" on my home network. SAMBA has been mentioned in another post I had online. Will this setup work with smart phones, consoles (such as ps3, or xbox), laptops, etc... they would all be connected wirelessly to the network.

I also thought about setting up a domain and trying to do something like an intranet. This way, users (my girlfriend, lol) can go to our OurMedia.com and browse the files on the server. Of course, I need to look into security for this so the world can't see all of my files and damage, steal, or plant viruses in anything.
 
Old 08-17-2012, 03:19 PM   #65
suicidaleggroll
LQ Guru
 
Registered: Nov 2010
Location: Colorado
Distribution: OpenSUSE, CentOS
Posts: 5,258

Rep: Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947
Quote:
Originally Posted by Aunnix View Post
Right now, my windows machine is only seeing like 690-700GB on the HDDs... not sure if Linux will read them differently or not.
That's normal. Hard drive manufacturers rate them in base 10 (1000). Capacity is generally measured in base 2 (1024).
The notation is technically different, GB for base 10, GiB for base 2, but very few people actually use GiB when referencing it.

So 750GB is actually 698 GiB, which is what your OS is reporting as the capacity. Linux will be the same way.
 
Old 08-17-2012, 03:20 PM   #66
suicidaleggroll
LQ Guru
 
Registered: Nov 2010
Location: Colorado
Distribution: OpenSUSE, CentOS
Posts: 5,258

Rep: Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947
Quote:
Originally Posted by Aunnix View Post
Will this setup work with smart phones, consoles (such as ps3, or xbox), laptops, etc... they would all be connected wirelessly to the network.
Laptops yes, never tried smart phones or consoles.
 
Old 08-17-2012, 03:27 PM   #67
Aunnix
Member
 
Registered: Aug 2012
Posts: 126

Original Poster
Rep: Reputation: Disabled
Quote:
Originally Posted by suicidaleggroll View Post
That's normal. Hard drive manufacturers rate them in base 10 (1000). Capacity is generally measured in base 2 (1024).
The notation is technically different, GB for base 10, GiB for base 2, but very few people actually use GiB when referencing it.

So 750GB is actually 698 GiB, which is what your OS is reporting as the capacity. Linux will be the same way.

I knew HDDs never actually have the amount they state on them. I knew you always lose a few GBs of storage, but I didn't know why (the base thing). I just didn't realize you lose so much, lol. I only lost like 10GB on my 160GB drive while my 320GB drive (in my windows machine) loses like 30GB.
 
Old 08-17-2012, 03:29 PM   #68
suicidaleggroll
LQ Guru
 
Registered: Nov 2010
Location: Colorado
Distribution: OpenSUSE, CentOS
Posts: 5,258

Rep: Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947
It comes out to about 7% when you're in the GB range, closer to 9% in the TB range.
 
Old 08-17-2012, 03:29 PM   #69
Aunnix
Member
 
Registered: Aug 2012
Posts: 126

Original Poster
Rep: Reputation: Disabled
Quote:
Originally Posted by suicidaleggroll View Post
Laptops yes, never tried smart phones or consoles.

Ok. I figured laptops, because they can recognize it as a "local drive" like the desktop. But, I'm not sure if smart devices and consoles are able to have that same recognition. That's where the intranet like setup comes to mind because all of these devices can access the web.
 
Old 08-19-2012, 05:55 AM   #70
sjreilly
Member
 
Registered: Oct 2007
Location: Edinburgh, Scotland
Distribution: Debian, CentOS, RHEL, Mint, Ubuntu
Posts: 79

Rep: Reputation: 15
Quote:
Originally Posted by frieza View Post
actually i would put all 3 on the 750 gig, depending on what you are doing with the machine, a LAMP server (linux apache mysql php for instance) will use /var for mysql and can be configured to have the Document root (the top level folder where the your are stored) either as /var/www (default) or somewhere in individual user's /home directory

I would shy away from this configuration for a couple of reasons;

a) If the 750GB drives come up after the OS drive the system might not boot as the partitions are not mounted yet, especially /etc.

b) It is pretty straightforward to redirect MySQL and apache root dirs to places with more space; /home/mysql and /home/www

c) Backup - if two of the OS partitions are on the RAID'd disks and the rest are on the 60G drive it gives Clonezilla a harder job if you need to reinstate the drive

d) /etc changes very little in size over the life of an OS - my web server /etc currently stands at 28M

Mirroring - there a number of packages in Linux that can RAID drives without the need for a hardware RAID solution; dmraid, LVM. They will use more of your CPU to do the RAID function than dedicated hardware.
 
Old 08-19-2012, 07:18 PM   #71
Aunnix
Member
 
Registered: Aug 2012
Posts: 126

Original Poster
Rep: Reputation: Disabled
Quote:
Originally Posted by sjreilly View Post
I would shy away from this configuration for a couple of reasons;

a) If the 750GB drives come up after the OS drive the system might not boot as the partitions are not mounted yet, especially /etc.

b) It is pretty straightforward to redirect MySQL and apache root dirs to places with more space; /home/mysql and /home/www

c) Backup - if two of the OS partitions are on the RAID'd disks and the rest are on the 60G drive it gives Clonezilla a harder job if you need to reinstate the drive

d) /etc changes very little in size over the life of an OS - my web server /etc currently stands at 28M

Mirroring - there a number of packages in Linux that can RAID drives without the need for a hardware RAID solution; dmraid, LVM. They will use more of your CPU to do the RAID function than dedicated hardware.

I'll have to look into the directing of PHP and mySQL directories. I like the idea of everything for the OS on the one drive...

As for the mirroring using the CPU... would my 3.2 dual core hold up? or would hardware for the RAID be needed? If the CPU can handle it well, that'd be great, lol.
 
Old 08-19-2012, 07:20 PM   #72
suicidaleggroll
LQ Guru
 
Registered: Nov 2010
Location: Colorado
Distribution: OpenSUSE, CentOS
Posts: 5,258

Rep: Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947
Hardware RAID is only really beneficial (from a performance standpoint) for the architectures that require parity calculation, namely RAID levels 5 and 6. For RAID levels 0, 1, and 10 it really won't make much of a performance difference.

mdadm is the standard software RAID controller in Linux, I would just stick with that.
 
Old 08-20-2012, 11:23 AM   #73
Aunnix
Member
 
Registered: Aug 2012
Posts: 126

Original Poster
Rep: Reputation: Disabled
Quote:
Originally Posted by suicidaleggroll View Post
Hardware RAID is only really beneficial (from a performance standpoint) for the architectures that require parity calculation, namely RAID levels 5 and 6. For RAID levels 0, 1, and 10 it really won't make much of a performance difference.

mdadm is the standard software RAID controller in Linux, I would just stick with that.

Is RAID 1 the only RAID "level" that does the mirroring? Above, someone mentioned RAID 0 will split my saved file information between the two 750GB HDDs so I won't use it. What can I do with RAID 10?
 
Old 08-20-2012, 11:40 AM   #74
suicidaleggroll
LQ Guru
 
Registered: Nov 2010
Location: Colorado
Distribution: OpenSUSE, CentOS
Posts: 5,258

Rep: Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947Reputation: 1947
Quote:
Originally Posted by Aunnix View Post
Is RAID 1 the only RAID "level" that does the mirroring? Above, someone mentioned RAID 0 will split my saved file information between the two 750GB HDDs so I won't use it. What can I do with RAID 10?
RAID 0 - data is stripped across all drives in the array. Total array capacity will be the sum of the individual drives. Speed goes up, but reliability goes down since losing just a single drive will wipe the entire array and everything on it. Processor overhead is low.

RAID 1 - data is mirrored across all drives in the array. Total array capacity will be the size of ONE of the drives. Write speed stays the same as single, read speed goes up, reliability goes up since you won't lose any data until you lose ALL of the drives in the array. Processor overhead is low.

RAID 10 - combination of 0 and 1. It requires 4+ drives, essentially you're setting up two 2-drive RAID 0 arrays (for a 4-drive RAID 10), and then putting those RAID 0 arrays in RAID 1. Total array capacity will be the sum of half of the drives (2TB for four 1TB drives, for example). Write speed goes up, read speed goes up, reliability goes up. You can lose at least one drive, possibly two (depending on which one it is) without losing any data. Processor overhead is low.

RAID 5 - basically like RAID 0, except one drive in the array is reserved for parity calculations. Total array capacity will be the sum of the individual drives minus one (3TB for four 1TB drives, for example). Write speed goes up, read speed goes up, reliability goes up. You can lose any ONE drive from the array without losing any data, but as soon as you lose a second drive you lose everything. Processor overhead is high.

RAID 6 - basically the same as RAID 5, except TWO drives are used for parity. Total array capacity will be the sum of the individual drives minus two (6TB for eight 1TB drives, for example). Write speed goes up, read speed goes up, reliability goes up. You can lose any TWO drives without losing any data, but as soon as you lose a third drive you lose everything. Processor overhead is very high.


All of my RAID systems run levels 1, 10, or 6, depending on the capacity and speed required (1 being the smallest, 6 the biggest). For levels 1 and 10 I usually just use mdadm, for level 6 I use a dedicated hardware controller because of all of the parity calculations involved...usually Adaptec since they're stupid fast, at least compared to LSI.

Last edited by suicidaleggroll; 08-20-2012 at 11:45 AM.
 
Old 08-20-2012, 01:00 PM   #75
Aunnix
Member
 
Registered: Aug 2012
Posts: 126

Original Poster
Rep: Reputation: Disabled
Quote:
Originally Posted by suicidaleggroll View Post
RAID 0 - data is stripped across all drives in the array. Total array capacity will be the sum of the individual drives. Speed goes up, but reliability goes down since losing just a single drive will wipe the entire array and everything on it. Processor overhead is low.

RAID 1 - data is mirrored across all drives in the array. Total array capacity will be the size of ONE of the drives. Write speed stays the same as single, read speed goes up, reliability goes up since you won't lose any data until you lose ALL of the drives in the array. Processor overhead is low.

RAID 10 - combination of 0 and 1. It requires 4+ drives, essentially you're setting up two 2-drive RAID 0 arrays (for a 4-drive RAID 10), and then putting those RAID 0 arrays in RAID 1. Total array capacity will be the sum of half of the drives (2TB for four 1TB drives, for example). Write speed goes up, read speed goes up, reliability goes up. You can lose at least one drive, possibly two (depending on which one it is) without losing any data. Processor overhead is low.

RAID 5 - basically like RAID 0, except one drive in the array is reserved for parity calculations. Total array capacity will be the sum of the individual drives minus one (3TB for four 1TB drives, for example). Write speed goes up, read speed goes up, reliability goes up. You can lose any ONE drive from the array without losing any data, but as soon as you lose a second drive you lose everything. Processor overhead is high.

RAID 6 - basically the same as RAID 5, except TWO drives are used for parity. Total array capacity will be the sum of the individual drives minus two (6TB for eight 1TB drives, for example). Write speed goes up, read speed goes up, reliability goes up. You can lose any TWO drives without losing any data, but as soon as you lose a third drive you lose everything. Processor overhead is very high.


All of my RAID systems run levels 1, 10, or 6, depending on the capacity and speed required (1 being the smallest, 6 the biggest). For levels 1 and 10 I usually just use mdadm, for level 6 I use a dedicated hardware controller because of all of the parity calculations involved...usually Adaptec since they're stupid fast, at least compared to LSI.

Thanks for the break down. I was wondering if there was another RAID level for mirroring where one might be better to do than the other... Looks like I'll be sticking with RAID 1 when I begin to set up to the hard drives.
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off



Similar Threads
Thread Thread Starter Forum Replies Last Post
LXer: Easy way to Build FTP server in Ubuntu Server 12.04 LXer Syndicated Linux News 0 05-29-2012 09:10 AM
jenkins continuous build server not starting application userbrian Programming 2 03-31-2012 03:12 PM
Starting to build a cluster martins48 Linux - Software 4 10-21-2011 06:33 PM
Building my first home server - atom based router/file server build need advice mothergoose729 Linux - Networking 5 03-23-2010 04:19 AM
starting xprogram without starting X server. bruse Linux - Newbie 9 04-26-2005 08:05 PM


All times are GMT -5. The time now is 08:27 PM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration