Proposed FreeNAS Build

[diedrichg]

I thought there was a specific reason that i had to start with the old version in order to get the data off.

I was just trying to save you some config time by getting you back on 9.2 and then upgrading. Sorry for the misunderstanding. Your data is safe as long as the disks are safe. It's nothing more than a simple volume (pool) import when you first boot into 9.10 and you follow the wizard. Be prepared to set up your users and shares during the wizard process. The purpose of the config import is to get your system back up and running quickly, the config knows nothing of your data but knows all the user, shares and jails info. You will not need to recreate your jails, simply point the jails location to where they live on your pool.

 

[thedirtmerchant]

Opened up 9.10 w/ the old drives connected.
Attempted to import the volume.
Step1 recognizes my volume...when I click OK for Step2 in the GUI,
the display on my FreeNAS box which usually shows the Console setup screen starts whizzing by full screens of text
so fast I can't read anything (I was assuming this was some kind of data import/copy process). After a minute or so of that
the box reboots (looks like maybe some hex flying by) and then I was able to read one line "stopped by CPU2". Reboots and when i get back into the GUI there is a message on the Storage tab "sorry, an error occured"

I'm thinking my drives are unrecoverable and I should just move on.

I'm sad about the lost data, but I got most of my stuff in a backup drive and in the cloud...

 

[Dice]

I've not followed the thread in detail but would like to comment on where you are at now with the failed import.
I'd suggest you persist in getting the data out from the drives, rather than quitting and restoring from backup (which is great relief).
My point is that once you run into some really critical scenarios down the line, you'll have a bit more experience under your belt to avoid doing critical mistakes that cannot be undone during recovery.
Many users panic and run and "test" commands to see if they get things to work but instead instill unrecoverable errors in the recovery process.
Even though it consumes time right now, I'd consider that to be well invested.

 

[diedrichg]

My gut says this is a hardware issue. I'm thinking mobo/cpu or a bios setting. Check the bios to see if the drives are set to AHCI mode.

 

[thedirtmerchant]

My gut says this is a hardware issue. I'm thinking mobo/cpu or a bios setting. Check the bios to see if the drives are set to AHCI mode.

You're good.

In the BIOS, SATA Configuration:
SATA Mode Selection: AHCI
I'm guessing this needs to change to RAID?
And it should stay that way w/ the new drives too (is this a duh...obviously! moment)?

Any other BIOS settings that I need to set properly? Never would have even thought about BIOS settings,
is there a thread on that somewhere? Need to open up the manual a bit more...

 

[diedrichg]

You're good.

In the BIOS, SATA Configuration:
SATA Mode Selection: AHCI
I'm guessing this needs to change to RAID?

I don't remember.

 

[diedrichg]

In the BIOS, SATA Configuration:
SATA Mode Selection: AHCI

AHCI is correct, I just rebooted and that's how mine is set (and it makes sense that it should be AHCI so that ZFS's software RAID can run the show).

 

[thedirtmerchant]

Based on some searches and the manual, it should be AHCI.
Ah well...

So, I got a video of the running text...which went on for about 2 minutes, then right at the bottom it stops for a second.
Trying to get a screen cap off the video I took on my phone, which is for some reason proving difficult.
Anyway, it says something about
"Tracing command kernel"
"Text dump complete"
and "cpu_reset: Restarting BSP"
and "cpu_reset_proxy: Stopped CPU4"

 

[Dice]

After reading through the thread and the branch linked I'll reassess my suggestion above.

You've not indicated your propensity to maintain the original configuration.
If you have a boatload of advanced jail setups that you've spent countless hours on setting up you might want to keep that intact as best it can, on the other hand, if you are just looking for the "media data" on the pool, I'd reconsider and skip all of these extra efforts to go via the older install versions to extract configs and do a flat fresh install and reconfiguration from scratch. It should not have any impact on your data on the pool.

The key aspect is whether ZFS has evolved in ways that interrupt a smooth import of the older pool.
Perhaps we could get a confirmation @Ericloewe @DrKK if importing an older pool requires additional precautions?
(I cannot find any such notes http://docs.oracle.com/cd/E19253-01/819-5461/gbchy/index.html )

 

[thedirtmerchant]

After reading through the thread and the branch linked I'll reassess my suggestion above.

You've not indicated your propensity to maintain the original configuration.
If you have a boatload of advanced jail setups that you've spent countless hours on setting up you might want to keep that intact as best it can, on the other hand, if you are just looking for the "media data" on the pool, I'd reconsider and skip all of these extra efforts to go via the older install versions to extract configs and do a flat fresh install and reconfiguration from scratch. It should not have any impact on your data on the pool.

The key aspect is whether ZFS has evolved in ways that interrupt a smooth import of the older pool.
Perhaps we could get a confirmation @Ericloewe @DrKK if importing an older pool requires additional precautions?
(I cannot find any such notes http://docs.oracle.com/cd/E19253-01/819-5461/gbchy/index.html )

I only want the media data from the pool, the jails are basic and I'm ok with setting them up from scratch.
My current issue is that I have installed 9.10.2-u2
When I try to import the volume (2 old drives that were RAIDZ1 connected) via the GUI, it is recognized (in that the option is there to import the volume).
Then it gives me the textdump, restarts the box and says "sorry there was an error" in the Storage menu/window.

Any thoughts on the importing an older pool is very much appreciated.
In that vein, I spent a bunch of time with the extremely helpful diedrichg trying to get 9.2.1.3 installed. I keep hitting the mountroot error 19 issue on the install.
I've tried w/ 2 USB keys and just now via IPMI and a virtual CD iso...which STILL hits the mountroot error.

 

[Dice]

2 old drives that were RAIDZ1 connected

Do you mean mirror?

My current issue is that I have installed 9.10.2-u2

Don't think this is an issue.

When I try to import the volume (2 old drives that were RAIDZ1 connected) via the GUI, it is recognized (in that the option is there to import the volume).
Then it gives me the textdump, restarts the box and says "sorry there was an error" in the Storage menu/window.

I'd suggest you get the machine online, disks connected, into the gui ...and then sshing into the box for some CLI action.
To provide some additional clues:
What are the outputs of

Code:

zpool status -vx

and

Code:

zpool get version <poolname>

.

What was the original FreeNAS version you created the pool on?
If it was prior to 9.1, my google-fu indicates there should have been a ZFS upgrade performed. If you'd missed out on that, I flat out guess it could cause problems. However, I hope not that is the case.
(Idea from this thread: https://forums.freenas.org/index.ph...on-after-zpool-ugrade-on-freenas-9-1-1.15478/ )

 

[Dice]

If you're unsure about what commands to, I'd personally reference the man page to get an idea. Be careful, some of these commands are very potent and cannot be "tested without consequences".
https://www.freebsd.org/cgi/man.cgi?zpool(8)

 

[thedirtmerchant]

Do you mean mirror?

What are the outputs of

Code:

zpool status -vx

and

Code:

zpool get version <poolname>

.

What was the original FreeNAS version you created the pool on?
If it was prior to 9.1, my google-fu indicates there should have been a ZFS upgrade performed. If you'd missed out on that, I flat out guess it could cause problems. However, I hope not that is the case.
(Idea from this thread: https://forums.freenas.org/index.ph...on-after-zpool-ugrade-on-freenas-9-1-1.15478/ )

The only pool appears to be my SSD with the OS.

Code:

[root@freenas ~]# zpool status												 
  pool: freenas-boot															
state: ONLINE																 
  scan: none requested														 
config:																		 
																				
		NAME		STATE	 READ WRITE CKSUM								 
		freenas-boot  ONLINE	   0	 0	 0								
		  ada0p2	ONLINE	   0	 0	 0								 
																				
errors: No known data errors

-vx says all pools healthy
get version, no pools are mounted, so I get "no such pool" if I'm looking for my old pool name.

I'm not sure what version I was using initially, I'm thinking 8.3?
But wouldn't that zfs upgrade have been required on update to a newer version post 9.1?
I was definitely using 9.2.1.3 (and POSSIBLY 9.3) when things started to go south on the old box.

 

[Dice]

ehee.... oh well.
I've done some googling for you and what I've found so far indicates that

Code:

zpool status

should find pools of all sorts, healthy, borked, attached, detached. I may be wrong.
At this point I'm quite concerned for the content on that pool.
Maybe some wizard has an idea @danb35 ..?

 

[thedirtmerchant]

ehee.... oh well.
I've done some googling for you and what I've found so far indicates that

Code:

zpool status

should find pools of all sorts, healthy, borked, attached, detached. I may be wrong.
At this point I'm quite concerned for the content on that pool.
Maybe some wizard has an idea @danb35 ..?

oh...that doesn't sound good.

 

[danb35]

Maybe some wizard has an idea @danb35 ..?

None at all, I'm afraid--if zpool import doesn't show anything for the pool in question, I really wouldn't know what else to try. There's probably something that could be done with zdb, but that command's very much into "there be dragons" territory for me.

 

[thedirtmerchant]

Maybe some hope?

Code:

[root@freenas ~]# zpool import												
   pool: Data1															 
	 id: 4719943691862474623													
  state: ONLINE																
status: One or more devices were being resilvered.							
action: The pool can be imported using its name or numeric identifier.		
config:																		
																				
		 Data1										ONLINE				
		  mirror-0									  ONLINE				
			gptid/22db21f3-06fb-11e7-adcf-94de8001cf84  ONLINE				
			gptid/52f9fa22-d0a8-11e4-a7a0-94de8001cf84  ONLINE

So, when my drives/box failed, I had replaced a drive and it was in a seeming resilver loop.
Then I shut everything down and started on the new box. Maybe it's available now?

Though zpool status does not show anything about Data1

Code:

Shell
[root@freenas ~]# zpool status												 
  pool: freenas-boot															
state: ONLINE																 
  scan: none requested														 
config:																		 
																				
		NAME		STATE	 READ WRITE CKSUM								 
		freenas-boot  ONLINE	   0	 0	 0								
		  ada0p2	ONLINE	   0	 0	 0 

Thoughts? Suggestions? Ideas?

 

[diedrichg]

I highly suspect that

1) 9.2.1.3 can not be installed because it does not have the proper drivers for your late model motherboard

2) 9.10 cannot import your pool because the feature flags were never upgraded beyond 9.2.1.3.

My suggestion would be to keep trying to install from each version higher than 9.2.1.3 to see if one finally installs and imports.

Final suggestion: get in touch with the iX gurus and see what can be done for this situation.

 

[danb35]

Hmmm, I must have been thinking of a different thread where zpool import had returned nothing. OK, so you have a pool, and zpool import is showing it, but when you try to import it, the system crashes? That normally indicates a corrupt pool, which is a Bad Thing (tm). What happens if you try zpool import -F -n Data?

 

[danb35]

9.10 cannot import your pool because the feature flags were never upgraded beyond 9.2.1.3.

This should not be a problem--newer versions of ZFS should be able to import older pools without issue. It's going the other way that's a problem.