HardDrive config

[jaminellio]

I have been using Amahi on a really old machine that has IDE drives for the last few months and have decided to it time to get all my data on to Amahi which means a new machine with SATA drives (or reconfig of existing machines).

I have been reading all I can about how to setup the Amahi sever's Harddrives and am a little confused by it all at the moment. Do I want Greyhole, RAID, LVM or all of them?

Question at the moment is should I have a smaller drive for Operating system, SWAP, Greyhole LZ, etc? (say 320GB or 500GB) Or should I just partition a larger drive (1.5 or 2Tb) with the space for the OS and use the rest for greyhole storage?

I am looking at final server have about 6Tb to start with and will probably increase in size over time.


Thanks

[rkillcrazy]

Question at the moment is should I have a smaller drive for Operating system, SWAP, Greyhole LZ, etc? (say 320GB or 500GB) Or should I just partition a larger drive (1.5 or 2Tb) with the space for the OS and use the rest for greyhole storage?

Generally speaking, the OS (Fedora) doesn't require much room for the installation. In fact, the typical Linux OS can be installed on a partition less than 10-GB and still have room to play with. Personally, I have a 750-GB system drive. I first tried to do this with an 80-GB HDD but it left little room for the initial media dump I did to the server. On the typical system, the data being dumped to the server has to sit in the shared folder for a while before Greyhole begins to deal with it. For that reason, if you have a lot of data to dump at first, you'll need a large system drive. The last install I did was for my buddy who had a full 2-TB HDD and needed to move into a server. I built his HDA with a 1-TB system drive and four 2-TB HDDs in Greyhole's pool. It took a little while to get his data over to the server 'cause we had to dump the files in chunks. So, it really all depends on how much data you have and how much you want to dump to your HDA initially. Plan accordingly with what you want to do and what you have to work with...

I have been reading all I can about how to setup the Amahi sever's Harddrives and am a little confused by it all at the moment. Do I want Greyhole, RAID, LVM or all of them?

I guess you could say you'll want all but one of them. You might want to read up on some of the terminology...

Greyhole
RAID
LVM

Greyhole will be used in lieu of RAID. On a traditional RAID array, you have to provide and carve out the space up front. If you RAID up two 2-TB HDDs in a RAID-1 configuration, you'll have a single volume with about half the space (~2-TB). While that provides fault tolerance, it cannot be dynamically expanded. Meaning, once you fill up that volume, you can't add another HDD to increase the size of it. Greyhole is not a RAID technology per se`. It is, however, a way to get a kind of fault tolerance. Amahi uses Greyhole to work with a drive pool. Meaning, you drop two 2-TB HDDs in the server, tell Amahi to use them in the pool and tell Amahi to allow Greyhole to make X number of additional copies of the data you save on the server. Once you fill that pooled volume, you can add additional storage and the pool can grow and grow. I have four 2-TB HDDs in my server and they are all in the pool. Greyhole is set to make 1 additional copy so, if I lose one of those HDDs, I will still have a backup on another HDD.

LVM will no doubt be used. If you install Fedora with the default options, it will create the LVM and all the partitions. Frankly, most don't need to worry about this sort of thing. Even I've never had to expand an LVM - which can be done but, again, I've never seen a case where it needed to be done. By all means, read up on it and understand what it is and why it is used.

Hope this helps. Amahi is designed to be easy to install by most competent users. It's wizard-driven and there is little config the user has to do once it's set up.

Here's my rough how-to. It's rough so be sure to use this as a mere guide and consult Amahi's documentation.

• Install the OS with just the system HDD plugged in.

• Using the link on the desktop, run the Amahi install once you've installed the OS.

• Shut down and plug in the HDDs you wish to use in the storage pool.

• Using the command line, add the additional HDDs to FSTAB.

• Using the web interface, tell Amahi to use those HDDs in the pool and set your number of extra copies for each share.

• Dump your data to the server.

[moredruid]

Basically it boils down to this:
use what you are comfortable with, follow the defaults if you don't know what everything means.

You can use everything together, sometimes the point is moot but sometimes it may be valid.

Default: start with an OS only disk, install, add the other disks one by one and add them to the greyhole pool (at least that's what bigfoot recommends IIRC).

I went for software RAID (which _is_ expandable BTW) with LVM on top. I could have chosen LVM RAID (so basically have a bunch of disks and have LVM handle the RAID operations), but due to performance reasons this wasn't desirable.
Notice however that you need to think very well ahead for your storage requirements and planning on how to use the storage. And you should at least have a good understanding of LVM & RAID concepts before dabbling with this. If you want to know more about that, fire up a VM host and experiment with it. That's basically how I test everything.

[jaminellio]

Thanks for the relpies

On the typical system, the data being dumped to the server has to sit in the shared folder for a while before Greyhole begins to deal with it. For that reason, if you have a lot of data to dump at first, you'll need a large system drive. The last install I did was for my buddy who had a full 2-TB HDD and needed to move into a server. I built his HDA with a 1-TB system drive and four 2-TB HDDs in Greyhole's pool. It took a little while to get his data over to the server 'cause we had to dump the files in chunks. So, it really all depends on how much data you have and how much you want to dump to your HDA initially. Plan accordingly with what you want to do and what you have to work with...

How much time do you mean for Greyhole to start dealing with the data? Minutes, Hours, Days? I assume it depends on the speed of the machine/components but is there some sort of guide? I have atleast 2tb of data I would like to put onto my server eventually.

[bigfoot65]

Recommend you take a look at the Greyhole web site. It will answer many of your questions.

http://code.google.com/p/greyhole/

[rkillcrazy]

Thanks for the relpies

On the typical system, the data being dumped to the server has to sit in the shared folder for a while before Greyhole begins to deal with it. For that reason, if you have a lot of data to dump at first, you'll need a large system drive. The last install I did was for my buddy who had a full 2-TB HDD and needed to move into a server. I built his HDA with a 1-TB system drive and four 2-TB HDDs in Greyhole's pool. It took a little while to get his data over to the server 'cause we had to dump the files in chunks. So, it really all depends on how much data you have and how much you want to dump to your HDA initially. Plan accordingly with what you want to do and what you have to work with...

How much time do you mean for Greyhole to start dealing with the data? Minutes, Hours, Days? I assume it depends on the speed of the machine/components but is there some sort of guide? I have atleast 2tb of data I would like to put onto my server eventually.

Like bigfoot said, read up and educate yourself. You might not fully understand it all right away but a good layer of background info goes a long way when you get deeper into it....

Having said that, I believe Greyhole waits until the data is finished copying. After SMB is "done" with the file(s), Greyhole takes over. However, if you copy over a mess of files, it may take a few minutes before Greyhole starts doing its magic. Moreover, I believe Greyhole works on a schedule too - kicking off around 00:00 (midnight).

Again, read the official documentation, as I'm basing this on my [very] poor memory.