Another pool degraded question.

[bar1]

You'll need to use a different method to replication.

rsync is probably the best option.

ok let me try that...
I was hopping to get my corrupt file from a snapshot.

 

[sretalla]

I was hopping to get my corrupt file from a snapshot.

If it's mounted, look in the .zfs directory and go through the snapshotd there directly.

 

[Arwen]

If the corrupted file has not changed, then the ZFS snapshot(s) will point to the exact same disk blocks. Thus, likely corrupt too.

This block sharing feature of ZFS' snapshots is one way ZFS saves disk space. But, as you may have found out, it means that without enough redundancy or backups, you can loose a file. ZFS snapshots are generally not backups, though they can be part of a backup strategy.

 

[bar1]

If the corrupted file has not changed, then the ZFS snapshot(s) will point to the exact same disk blocks. Thus, likely corrupt too.

This block sharing feature of ZFS' snapshots is one way ZFS saves disk space. But, as you may have found out, it means that without enough redundancy or backups, you can loose a file. ZFS snapshots are generally not backups, though they can be part of a backup strategy.

Ok thanks, the corrupt file did change....but lets see what i can do it wont be the end of the world if i lose it but will be a little bit of extra work for me to recreate it.

 

[bar1]

Ok all my data is recovered....not too sure what to do next, been thinking maybe I should move from truenas to something else (I do really like truenas though).

What is the best practice for a JBOD setup?
Should I create 1 pool with 2 vdevs , and "copies setting greater than 1" on important data? (might be hard to set up as my 8TB is completely full now).

 

[sretalla]

Should I create 1 pool with 2 vdevs , and "copies setting greater than 1" on important data?

I don't think that will get you what you think it does...

Some protection against corruption, yes, but if either disk is lost, all content is gone.

if you just want to do JBOD, then UNRAID probably has you covered the best.

 

[bar1]

What does this setting do exactly?
My set up is, 18TB pool...important stuff replicate to my 8TB pool....
I am actually very happy with that.

I do have an unraid license I got many years ago.

 

[sretalla]

Copies is more-or-less exactly what it says.

Somewhere in the pool, there will be x copies of the files (actually the blocks, but same result) relating to everything in that dataset (or those datasets where you assigned the non-default value).

Nothing in the background is making sure that each of the copies is on a different VDEV (it may or may not be) and individual VDEVs are no good to you if not all VDEVs are present anyway.

In the case of corruption being detected by checksum on one of the copies, the additional copy (or copies) can be used to repair the corruption during a scrub or even a read operation (it's automatic).

 

[bar1]

Ok I understand.
Any recommendations to improve my set up?

- Pass-through the controller (I was really struggling with it, but will try again)
- I can try allocate more RAM

Should I set up a cache drive? I can allocate a virtual drive of maybe 150GB maybe more.

 

[sretalla]

- Pass-through the controller (I was really struggling with it, but will try again)

If you're virtualized, that's a must do.

- I can try allocate more RAM

Generally, more is always better for ZFS/TrueNAS

Should I set up a cache drive? I can allocate a virtual drive of maybe 150GB maybe more.

Not if it's going to be a pool member as a virtual drive. And probably not anyway as you should really have 64GB of RAM before even thinking about L2ARC.

 

[Davvo]

(might be hard to set up as my 8TB is completely full now).

Generally you want to leave 20-10% of free space for the CoW to not struggle.
Usable space is usually considered to be 80% of your total space but it's up to you to decide were the line is up to 90%, you don't want to go any further.
Your performance will die a gruesome death otherwise.

 

[bar1]

Generally you want to leave 20-10% of free space for the CoW to not struggle.
Usable space is usually considered to be 80% of your total space but it's up to you to decide were the line is up to 90%, you don't want to go any further.
Your performance will die a gruesome death otherwise.

Just temorarily so I can recreate my 18TB pool .

 

[bar1]

You can try zpool import -f Bar1-18TB.
Hard drive passthrough is fine if the drives are directly attached to the motherboard iirc.

How can I check that?really struggling with the controller passthrough.

 

[Davvo]

How can I check that?really struggling with the controller passthrough.

No idea, I do not virtualize and my knowledge on the field is based on sporadics readings of others' posts.

Please read the following resource:

"Absolutely must virtualize FreeNAS!" ... a guide to not completely losing your data.

[---- 2018/02/27: This is still as relevant as ever. As PCIe-Passthru has matured, fewer problems are reported. I've updated some specific things known to be problematic ----] [---- 2014/12/24: Note, there is another post discussing how to deploy a small FreeNAS VM instance for basic file...

www.truenas.com

 

[Whattteva]

How can I check that?really struggling with the controller passthrough.

To passthrough properly, your CPU needs virtualization support for directed I/O. I don't think you ever posted what your hardware specs is, so I don't know which CPU you're using. That being said, you need VT-d for Intel or AMD Vi (sometimes can also be called IOMMU).