Hello all

[BradTheGeek]

My name, as you can probably tell is Brad. I have bee working in various forms and functions in IT for nearly 20 years now.

I am new to FreeNas (and my unix/sysadmin skills are about 5-6 years rusty). In my current incarnation, I find myself managing the PC repair shop portion of a small IT consulting and PC repair company. Mostly I handle lots of consumer grade boxes coming our way for new builds, repair, virus cleanup, etc. I have lots of standard system images for business customer rollouts, new system sales, etc. Also I find it useful to image incoming repair work from time to time (especially when there is fear of data loss/drive failure). To that end, my little brain was working and I said hey... drive space is cheap, so why not set up a system to image every incoming box (if not prevented by failed hardware).

That is where FreeNAS suits me. I can set up a good drive pool, replicate it or rsync it to a spare server. I can encrypt the drives such that when I pull the boot medium (USB flash drive), customer data is relatively secure. It supports SMB/CIFS which my imaging program can access, and more.

Right now I have set up two servers (one primary, one spare), not really production yet, but to play with configs, learn my way around, develop procedures, etc.

So, that is where I am. There seems to be a wealth of info in the wiki, forum, etc, but sometimes it can be hard to find exactly what you need in the pile of stuff, so if I ask newbish questions, don't cuss me out right away. I will search, at least some before screaming HELP in the forums.

I do have a couple questions burning in my head now, and am more than happy if people point me to the best articles/posts in regards to them.

1. What are best practices for growing drive pools? I plan to use ZFS and provision enough for a while, but usage needs change, sometimes abruptly. I understand I can swap out drives for larger drives one at a time, and also I am reading a little on vDevs and zPools but it still hasn't congealed in my mind what the best way to do it is, and I want to make sure my initial setup is amenable to future expansion.

2. What is the best way to backup/clone configs. I am booting the servers from flash drives, but if a flash drive fails, I want to be able to recover (especially if enc. keys are on there). Should I just regularly clone my boot media for each server? Is there more config/system files stored on my internal disks and not my boot media that I should be concerned with?

3. With 2 servers are ZFS replications via replication tasks the best way to mirror the stored files? What are the implications of this if I grow the storage pool in one server first?

4. Do I sound like a moron? Are there obvious things my config/setup seems to be missing?

 

[Hyperion]

My name, as you can probably tell is Brad. I have bee working in various forms and functions in IT for nearly 20 years now.

I am new to FreeNas (and my unix/sysadmin skills are about 5-6 years rusty). In my current incarnation, I find myself managing the PC repair shop portion of a small IT consulting and PC repair company. Mostly I handle lots of consumer grade boxes coming our way for new builds, repair, virus cleanup, etc. I have lots of standard system images for business customer rollouts, new system sales, etc. Also I find it useful to image incoming repair work from time to time (especially when there is fear of data loss/drive failure). To that end, my little brain was working and I said hey... drive space is cheap, so why not set up a system to image every incoming box (if not prevented by failed hardware).

That is where FreeNAS suits me. I can set up a good drive pool, replicate it or rsync it to a spare server. I can encrypt the drives such that when I pull the boot medium (USB flash drive), customer data is relatively secure. It supports SMB/CIFS which my imaging program can access, and more.

Right now I have set up two servers (one primary, one spare), not really production yet, but to play with configs, learn my way around, develop procedures, etc.

So, that is where I am. There seems to be a wealth of info in the wiki, forum, etc, but sometimes it can be hard to find exactly what you need in the pile of stuff, so if I ask newbish questions, don't cuss me out right away. I will search, at least some before screaming HELP in the forums.

I do have a couple questions burning in my head now, and am more than happy if people point me to the best articles/posts in regards to them.

1. What are best practices for growing drive pools? I plan to use ZFS and provision enough for a while, but usage needs change, sometimes abruptly. I understand I can swap out drives for larger drives one at a time, and also I am reading a little on vDevs and zPools but it still hasn't congealed in my mind what the best way to do it is, and I want to make sure my initial setup is amenable to future expansion.

2. What is the best way to backup/clone configs. I am booting the servers from flash drives, but if a flash drive fails, I want to be able to recover (especially if enc. keys are on there). Should I just regularly clone my boot media for each server? Is there more config/system files stored on my internal disks and not my boot media that I should be concerned with?

3. With 2 servers are ZFS replications via replication tasks the best way to mirror the stored files? What are the implications of this if I grow the storage pool in one server first?

4. Do I sound like a moron? Are there obvious things my config/setup seems to be missing?

 

[Hyperion]

1. As far as I can tell ZFS is a safe storage format, not a lot of errors, but
It depends on the OS u use.

2. Rsync works very well. But it’s a bitch to set up.

3. Pass

4. Pass

 

[Ericloewe]

My name, as you can probably tell is Brad. I have bee working in various forms and functions in IT for nearly 20 years now.

I am new to FreeNas (and my unix/sysadmin skills are about 5-6 years rusty). In my current incarnation, I find myself managing the PC repair shop portion of a small IT consulting and PC repair company. Mostly I handle lots of consumer grade boxes coming our way for new builds, repair, virus cleanup, etc. I have lots of standard system images for business customer rollouts, new system sales, etc. Also I find it useful to image incoming repair work from time to time (especially when there is fear of data loss/drive failure). To that end, my little brain was working and I said hey... drive space is cheap, so why not set up a system to image every incoming box (if not prevented by failed hardware).

That is where FreeNAS suits me. I can set up a good drive pool, replicate it or rsync it to a spare server. I can encrypt the drives such that when I pull the boot medium (USB flash drive), customer data is relatively secure. It supports SMB/CIFS which my imaging program can access, and more.

Right now I have set up two servers (one primary, one spare), not really production yet, but to play with configs, learn my way around, develop procedures, etc.

So, that is where I am. There seems to be a wealth of info in the wiki, forum, etc, but sometimes it can be hard to find exactly what you need in the pile of stuff, so if I ask newbish questions, don't cuss me out right away. I will search, at least some before screaming HELP in the forums.

I do have a couple questions burning in my head now, and am more than happy if people point me to the best articles/posts in regards to them.

1. What are best practices for growing drive pools? I plan to use ZFS and provision enough for a while, but usage needs change, sometimes abruptly. I understand I can swap out drives for larger drives one at a time, and also I am reading a little on vDevs and zPools but it still hasn't congealed in my mind what the best way to do it is, and I want to make sure my initial setup is amenable to future expansion.

2. What is the best way to backup/clone configs. I am booting the servers from flash drives, but if a flash drive fails, I want to be able to recover (especially if enc. keys are on there). Should I just regularly clone my boot media for each server? Is there more config/system files stored on my internal disks and not my boot media that I should be concerned with?

3. With 2 servers are ZFS replications via replication tasks the best way to mirror the stored files? What are the implications of this if I grow the storage pool in one server first?

4. Do I sound like a moron? Are there obvious things my config/setup seems to be missing?

The most important thing is that you thoroughly read the documentation (Cyberjock's guide is a very good introduction).

As for your specific questions:

1. Besides swapping out the disks one by one, you can add vdevs to a pool. Adding a second RAIDZ2 vdev to a pool that has one RAIDZ2 vdev creates a sort of RAID60 setup, for instance. You can also make a new pool with the new vdev. You'll have to figure out what's best for you.
   If you plan on having a ton of storage, you're probably looking at an SAS controller and an expander (or several SAS controllers, depending on pricing and other factors). The big need is to determine upfront if you'll need more than 32GB of RAM, as that implies a more expensive Xeon E5 setup.
2. Backing up the configuration is easily done by backing up the config file. For some mysterious reason, cloning the boot drive doesn't work, according to the local experts. With the pool and the config file, you can start over normally (unless something went very wrong).
3. ZFS replication is the preferred way of backing up ZFS volumes. I don't know how it handles different-sized pools, though.
4. No, just read what there is to read and then feel free to ask any follow-up questions.

 

[BradTheGeek]

Thanks all, I am trying to read up on things. With all my responsibilities, jumping on something like this is a bit like drinking from a fire hose, but I will manage.

As to SAS, that is probably not necessary. At least not for a good long while. Right now I have set up a temp server to play, and part of my process is seeing how much space gets used with my proposed shop procedure. If I am imaging all customer computers and keeping the image for 30 days, it will tell me what that looks like from a capacity provisioning perspective. Then I can plan my final build a little better. Hopefully once I have done a bunch more reading!

 

[Ericloewe]

Thanks all, I am trying to read up on things. With all my responsibilities, jumping on something like this is a bit like drinking from a fire hose, but I will manage.

As to SAS, that is probably not necessary. At least not for a good long while. Right now I have set up a temp server to play, and part of my process is seeing how much space gets used with my proposed shop procedure. If I am imaging all customer computers and keeping the image for 30 days, it will tell me what that looks like from a capacity provisioning perspective. Then I can plan my final build a little better. Hopefully once I have done a bunch more reading!

In any case, proper choice of the motherboard always allows for future expansion. It's just cheaper to buy a motherboard with an LSI SAS2308 on board instead of separate motherboard and controller card - so a general idea of how much you'll need can save some time, money and trouble.

If you don't foresee a need for more than 11-12TB of storage, 6 SATA channels is enough, with 4TB HDDs, RAIDZ2 and 20% free space (nasty stuff happens if you fill up your pools too much, so 20% free space is advisable).