Help with File Server

[fielder]

Hi,

I hope this is the correct section to ask this question. If not please direct me to the correct section.

I am setting up FreeNAS for my home setup. I am using FreeNAS with vmware. I want to setup a central File Server for all my data.

Here is my setup:

For FreeNAS:
- I created vdev1 with 1 TB of usable space for virtual machines for vmware. I am sharing that out using NFS and adding that as a datastore on vmware.
- I created another vdev2 with 14 TB of usable space for my data.

For VMWare:
On each VMWare host I added the vdev1 NFS share as a datastore. I am using that to store virtual machines.

File Server Option 1:
Setup vdev2 as CIFS and integrate FreeNAS with Active Directory. This will be fine and I will be able to do permissions on each directory. But then how to I backup CIFS share?

File Server Option 2:
Setup vdev2 as NFS and add that as a second datastore on vmware hosts. Then create a Windows File Server virtual machine in vmware and add "second datastore" as second hard drive for it. Using this method I can use number of backup programs to backup.

Pros and Cons to each options:
Option 1:
Pros: Each file is an individual file on the FreeNAS server so when FreeNAS does scrubs each file is independent. If it detects any errors it will only effect those files and all the rest will be ok.
Cons: How do I backup CIFS share?

Option 2:
Pros: Easier to backup since all the data is local to the Windows File Server.
Cons: All my data is one big .vmdk file on vdev2. So when FreeNAS does scrubs it will only have 1 huge file. So if it detects any errors it can potentially corrupt/destroy all my data.

What option is better? Are there any backup solution to CIFS?

 

[SweetAndLow]

- I created vdev1 with 1 TB of usable space for virtual machines for vmware. I am sharing that out using NFS and adding that as a datastore on vmware.
- I created another vdev2 with 14 TB of usable space for my data.

This doesn't make any sense. I suggest you read the noob guide, it might help you get your terminology correct.

For you other question around 1 big file vs multiple files and data corruption. It doesn't matter which one you choose. ZFS see's all the data the same and will protect it the same. Data is checksumed by blocks written to disks not based on the file.

 

[fielder]

Sorry my terminology was completely wrong.

I created a volume called "vol01" using ssds. The total available space on this volume is about 1 TB. Then on this volume I created a dataset called "ssd_dataset01" which I am sharing using NFS to vmware.

I created a second volume called "vol02" using hard drives. The total available space on this volume is about 14 TB. Then on this volume I created dataset called "hdd_dataset01" which I could share using CIFS (to be used as file server on FreeNAS) or NFS (to be used by vmware as a datastore).

About your answer for data corruption... then what happens if there is corruption? Does it make a difference if there is 1 big file vs lots of small files? With 1 big file if there is corruption do I loose everything? With lots of small files if there is corruption then do I just loose those corrupted files or still loose everything on that dataset?

 

[SweetAndLow]

Sorry my terminology was completely wrong.

I created a volume called "vol01" using ssds. The total available space on this volume is about 1 TB. Then on this volume I created a dataset called "ssd_dataset01" which I am sharing using NFS to vmware.

I created a second volume called "vol02" using hard drives. The total available space on this volume is about 14 TB. Then on this volume I created dataset called "hdd_dataset01" which I could share using CIFS (to be used as file server on FreeNAS) or NFS (to be used by vmware as a datastore).

About your answer for data corruption... then what happens if there is corruption? Does it make a difference if there is 1 big file vs lots of small files? With 1 big file if there is corruption do I loose everything? With lots of small files if there is corruption then do I just loose those corrupted files or still loose everything on that dataset?

That makes much more sense or in just not as sleepy now.

I would go with option 1 and backup using what ever method works for you. Zfs replication, rsync, crashplan, ect.

Also why do you need a Windows file server vm? Can't freenas server up your files?

 

[fielder]

Option 1 is where FreeNAS would be my file server using CIFS. I would not need Windows File Server VM in this case. But then how do I do backups? You suggested ZFS replication but I don't have another FreeNAS server to replicate on to. What I have is spare desktop with 4 External Hard Drives attached to it. What backup option can you suggest I can use to backup CIFS shares from FreeNAS?

The only reason Option 2 was even being considered was because with Windows File Server VM it is easier to do backups using Veeam Endpoint Backup Free or something similar. But Veeam Endpoint Backup Free or other similar backup programs do not support backing up CIFS shares. The files have to be on the local server in order for it to backup.

 

[pirateghost]

technically you could use any number of ways to back up. the decision rests on what your backup target is running. windows? rsync(deltacopy) or robocopy would work.