replicate JBOD of 6 HD's to 1 HD

[erwin...@fastmail.fm]

I'm trying to free up my AOC-SAT2-MV8 to build a new pool.
Originally, a couple of years ago, I grabbed together all disks I had and threw these together into pool0. All together some 2.6 TB. It's really a mix, there's even an IDE drive.
These are the sizes in GB, just to give an idea.
232,84
465,76
1360
233,76
189,92
250

I'd like to try to add a single 3TB drive, create a new pool, and replicate, mirror or copy the old pool to this new pool.

Is this possible, and if yes, what command would I need to use?
I would need to do this using a live DVD, either osol 0906 or Openindiana oi-dev-151a7-live.

Problem is that I can't get EON to boot anymore on this system.
(Asrock AliveNF7G-HDready with AMD cpu) I had EON 0.600-130-64-cifs running on this machine, tried to install EON 1.0b-151-64-cifs-min but it won't reboot.

What I did here is:
Take out AOC controller, disconnect IDE and disable in bios.
Boot EON 1.0b CD. This comes up ok. Did the tests like mentioned on the wiki.
No problem. Then install.sh. All goes well, but no way to boot from the stick.
(yes, checked bios, USB was marked bootable in priority)
Tried all 4 grub methods. The OEM versions error out, the EON versions hang.
That's why I'd like to get rid of this Asrock mobo and replicate my zpool.

Thanks a lot for your insight!

Erwin

[andr...@gmail.com]

Here are some options:
Pre-req: all disks for current pool and new pool must be connected properly on the system this will be done on.

Boot:
-You can boot on another motherboard/system if you have one(using same usb that owns the pool, so no forced import required). You are not tied to the asrock.
-If you can boot on the Asrock motherboard with the OI live cd, do so. However, the downside of the live cd, is you will have to import the pool using the force switch. The way your pool was built is out side the guidelines of the rules, so I'd try the force switch with the read only import switch to test for errors first. Do not remove the read only switch until you are all set to roll the dice. This is the most critical point.
-best chance is to be able to boot without having to do a forced import.

If you pass boot and pool mounts successfully:
-build the new pool and from there you could rsync or zfs send/receive the data from old pool to new pool.

Let me know if anything is unclear.
Sent via BlackBerry by AT&T

--
You received this message because you are subscribed to the Google Groups "EON ZFS Storage" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage+...@googlegroups.com.
To post to this group, send email to eonst...@googlegroups.com.
Visit this group at http://groups.google.com/group/eonstorage?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.

[andr...@gmail.com]

Does it make it to the grub EON boot screen?
If so can you boot with the verbose switch and post a screenshot?
Are you saying EON 1.0 boots and install but after setup it won't boot/re-boot?

Sent via BlackBerry by AT&T

-----Original Message-----
From: eonstorage+noreply-APn2wQfNxQbZC...@googlegroups.com
Sender: EON ZFS Storage <eonst...@googlegroups.com>
Date: Thu, 7 Feb 2013 15:00:28
To: <eonst...@googlegroups.com>
Reply-To: EON ZFS Storage <eonst...@googlegroups.com>
Subject: [EON Storage] replicate JBOD of 6 HD's to 1 HD

[erwin...@fastmail.fm]

On Friday, February 8, 2013 12:35:02 AM UTC+1, dre2kse wrote:
> Does it make it to the grub EON boot screen?
>
> If so can you boot with the verbose switch and post a screenshot?
>
> Are you saying EON 1.0 boots and install but after setup it won't boot/re-boot?

Yes, that's exactly what I'm saying. As a matter of fact I did the whole install process twice with the same result. I'll also try to run this stick on another machine to see how it goes.

I'll run GRUB verbose and post back.
I also happen to have a copy of the first install history as I did this using ssh, this lists all the test commands as listed in the wiki.

Erwin

[erwin...@fastmail.fm]

Believe it or not, I put the system back to its state (controller, hd's)
I didn't have the USB the pool belonged to anymore. (It had a corrupted part somehow, I tried to salvage following my previous post of january 4 & 5.
I could only get it to boot using the 3rd GRUB option: OEM.)

Booting threw the same error. Frozen.
Next I edited GRUB to boot verbose.
And the system continued to boot successfully!!

To make sure, I repeated this scenario a couple of times, but all was the same.
Next I made the GRUB -v edit persistent editing /mnt/eon0/boot/grub/menu.lst

Next I setup my new user, ran ./slinky r and ./slinky c
Ran setup etc.

I still have some messages like:
SMF Initialization problems..svc:/network/nfs/server:default
SMF Initialization problems..svc:/network/nfs/nlockmgr:default
SMF Initialization problems..svc:/network/nfs/nlockmgr:default
SMF Initialization problems..svc:/network/nfs/nlockmgr:default
SMF Initialization problems..svc:/network/nfs/nlockmgr:default
SMF Initialization problems..svc:/network/nfs/server:default
SMF Initialization problems..svc:/network/nfs/server:default
SMF Initialization problems..svc:/network/nfs/client:default
SMF Initialization problems..svc:/network/nfs/client:default
SMF Initialization problems..svc:/network/nfs/server:default
SMF Initialization problems..svc:/network/nfs/mapid:default

I guess these are a remnant of NFS running on the previous USB stick.

I did not follow the way of the read-only import, as a)I was just unaware of this feature and b) I already forced import yesterday on a couple of occassions booting with the live DVD's

So finally where am I now?
- Added 3TB drive and created pool00
Purpose is to zfs send the existing pool0 to create a replica.
I've peeked into the commands, but I'm not very sure.
It seems I need to create a snapshot first? As you can see, I still have a lot to learn, but it's an extremely interesting voyage!
- Also added an extra 2TB drive.
Purpose is to use a partition and then as play / test ground for zfs learning.

What are your suggestions?

Erwin

[andr...@gmail.com]

Your guess on the smf nfs errors are correct.

You are also correct on needing to create a snapshot before zfs send/receive. Remember you can also use rsync between the pools.

Its weird that the oem image works fine then after running updimg that system does not boot. Can you screenshot how far the
boot with -v -m verbose options go with the configured image?

Make sure you update to the latest updates using rail.

You are building some risky pools :) please make sure you have a backup or please add a mirror 3TB.

Sent via BlackBerry by AT&T

-----Original Message-----
From: eonstorage+noreply-APn2wQfNxQbZC...@googlegroups.com
Sender: EON ZFS Storage <eonst...@googlegroups.com>
Date: Fri, 8 Feb 2013 09:14:11
To: <eonst...@googlegroups.com>
Reply-To: EON ZFS Storage <eonst...@googlegroups.com>

[erwin...@fastmail.fm]

Your guess on the smf nfs errors are correct.

> What is the way to get rid of these?
> I'd like to gain an understanding as to where this information resides? It must be 'inside' the pool and not in some /etc/config file?

>
You are also correct on needing to create a snapshot before zfs send/receive. Remember you can also use rsync between the pools.

> Here's a snippet of my history log trying to create a snapshot.
> My reasoning being that I want to make a snapshot of pool0 and want to name it the date of 08-02-2013:
** Create snapshot
library:8:~#zfs snapshot pool0@08-02-2013
library:9:~#zfs list -r pool0
NAME USED AVAIL REFER MOUNTPOINT
pool0 2.28T 358G 20.2G /pool0
pool0/data 394G 358G 394G /pool0/data
pool0/public 1.88T 358G 1.35T /pool0/public
pool0/swap 2.13G 361G 16K -
library:10:~#

*** Hmm, I don't see anything?
library:10:~#zfs snapshot pool0/data@08-02-2013_data
/Perhaps I need this 'sub-dir'?/

**** library:11:~#zfs list -r pool0/data
NAME USED AVAIL REFER MOUNTPOINT
pool0/data 394G 358G 394G /pool0/data
/Don't see anything yet../
/It seems I'd need to do recursive?/

**** library:12:~#zfs snapshot -r pool0/data@08-02-2013_data
cannot create snapshot 'pool0/data@08-02-2013_data': dataset already exists
no snapshots were created
/Don't see nothing yet, but at least a sign snapshot is alive, and I'm doing something's wrong../

**** Ok, so it seems I created something, but probably something empty
library:13:~#zfs list
NAME USED AVAIL REFER MOUNTPOINT
pool0 2.28T 358G 20.2G /pool0
pool0/data 394G 358G 394G /pool0/data
pool0/public 1.88T 358G 1.35T /pool0/public
pool0/swap 2.13G 361G 16K -
pool00 468K 2.68T 144K /pool00
library:14:~#zfs list -r pool0/data
NAME USED AVAIL REFER MOUNTPOINT
pool0/data 394G 358G 394G /pool0/data
library:15:~#zfs list -r pool0/data@08-02-2013_data
NAME USED AVAIL REFER MOUNTPOINT
pool0/data@08-02-2013_data 0 - 394G -
library:16:~#zfs destroy pool0/data@08-02-2013_data
library:17:~#zfs list -r pool0/data@08-02-2013_data
cannot open 'pool0/data@08-02-2013_data': dataset does not exist

>
>
> Its weird that the oem image works fine then after running updimg that system does not boot. Can you screenshot how far the boot with -v -m verbose options go with the configured image?

> We have a misunderstanding here.
The oem boot succeeding is about the previous 0.6 version. Somehow the usb image got dammaged and could not boot anymore. The error was error 23 or 24 and I think it was GRUB related.
I used the same stick and installed 1.0 overwriting the 0.6 version. (At least I assume it's being formatted during install)
It's with this fresh 1.0 install that I could only boot when adding the -v switch.

Now I must admit, at some point I've thought the system froze, and as I was continuing working on another PC, while I thought it it just stood waiting, all of the sudden after 'minutes'it continued booting while I'm not aware of what has been processing in the background, perhaps only some unresponsive network traffic??
So I'm not 100% sure. I can't exactly pinpoint to what situation the attachd screenshot is related.

At the mement I've put the system into a less accessible place and it's headless. I'd like to edit my GRUB and copy the first boot entry (without -v) so that I can choose with or without -v.
I'll then retest booting and post back.
When it hung, it froze immediately after the SunOS Release message, nothing more after "Üse is subject to licence terms."

>
>
Make sure you update to the latest updates using rail.

> I have to look into this, but I've not installed this before.
> Do I need to install the binary kit first?

>
You are building some risky pools :) please make sure you have a backup or please add a mirror 3TB.

> I realize this, but in fact at a certain point I already considered the pool > lost. It would be a shame but nothing life threatening.
> I have some questions on this issue, but I'll open a new topic.

Erwin

[dre2kse]

On Saturday, February 9, 2013 5:11:20 AM UTC-5, (unknown) wrote:

Your guess on the smf nfs errors are correct.
> What is the way to get rid of these?
> I'd like to gain an understanding as to where this information resides? It must be 'inside' the pool and not in some /etc/config file?

To get rid of these requires a fix in the illumos source. Even though NFS services are not imported in SMF it is somehow being checked. These are harmless and should be ignore for now. The other reason is usually NFS was configured on the previous image/pool but is not on the current booted image. Based on which of these is the cause it would be located in the  SMF database /etc/svc/repository.

 

>
You are also correct on needing to create a snapshot before zfs send/receive. Remember you can also use rsync between the pools.
> Here's a snippet of my history log trying to create a snapshot.
> My reasoning being that I want to make a snapshot of pool0 and want to name it the date of 08-02-2013:
** Create snapshot

   library:8:~#zfs snapshot pool0@08-02-2013  correct

   library:9:~#zfs list -r pool0
   NAME           USED  AVAIL  REFER  MOUNTPOINT
   pool0         2.28T   358G  20.2G  /pool0
   pool0/data     394G   358G   394G  /pool0/data
   pool0/public  1.88T   358G  1.35T  /pool0/public
   pool0/swap    2.13G   361G    16K  -
   library:10:~#

*** Hmm, I don't see anything? zfs list -t snapshot to see them

I don't see verbose data in the screenshot. Try these guidelines https://sites.google.com/site/eonstorage/faq#TOC-My-computer-hangs-or-reboots-while-booting-EON-

If you have a DB9 + cisco cable and hyperterminal on win you can redirect console levels https://sites.google.com/site/eonstorage/faq#TOC-How-do-I-enable-console-redirection-on-EON-

Now I must admit, at some point I've thought the system froze, and as I was continuing working on another PC, while I thought it it just stood waiting, all of the sudden after 'minutes'it continued booting while I'm not aware of what has been processing in the background, perhaps only some unresponsive network traffic??
So I'm not 100% sure. I can't exactly pinpoint to what situation the attachd screenshot is related.

 

Sounds like a low level hardware issue somewhere, as it sounds like if given time it times out and comes to login?

At the mement I've put the system into a less accessible place and it's headless. I'd like to edit my GRUB and copy the first boot entry (without -v) so that I can choose with or without -v.
I'll then retest booting and post back.
When it hung, it froze immediately after the SunOS Release message, nothing more after "Üse is subject to licence terms."
>
>
Make sure you update to the latest updates using rail.
> I have to look into this, but I've not installed this before.
> Do I need to install the binary kit first?

No, the binary kit does not need to be installed but it seems we need to focus on these issues and can back burner these for later. Link here https://sites.google.com/site/eonstorage/updating-eon 

[dre2kse]

The zfs send/receive commands you could use are. (receive switches can also be -Fduv, -d to maintain naming scheme and -u to make sure the newly created zfs does not get mounted)

# send/receive between local pools

zfs send -Rv old_pool/zfs_name@snapshot_name | zfs receive -Fv new_pool/new_or_same_zfs_name

# send/receive across the network using ssh (could also use mbuffer and netcat included in binary kit)

zfs send -Rv local_pool/zfs_name@snapshot_name | ssh root@remote_eon_box zfs receive -Fv new_pool/new_or_same_zfs_name

[dre2kse]

Here's a link to the zfs send/receive howto http://sites.google.com/site/eonstorage/syncing-data