further restoring / send - receive of lost pool - howto?

[ipanini]

As this post started to get looooong and the subject did not fit the bill, I chose to open a new post.

Where I'm at?

- bootable eon system release 151a9-071215

- two previously pools ready to be imported with -f

What I tried to do?

- re-create the previous pool0 and pool00 with their respective disks (succeeded)

- did updimg.sh hoping the newly created pools would survive reboot 

They did not survive reboot.

What am I missing?

[Andre Lue]

should survive reboot if updimg succeeds. check hostid before and after reboot(should not change). Also check /mnt/eon0/.exec if POOL= is defined accurately(no typos)

did you try a zpool import -a without the -f switch as a sanity check?

--
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 https://groups.google.com/group/eonstorage.
For more options, visit https://groups.google.com/d/optout.

[ipanini]

Thx for following up;

- it doesn't survive a reboot

- in the mean time I found your wiki article about /mnt/eon0/.exec 

- previously I did not adapt .exec

- but I did after finding the article

- typos:

looks like this:

----------------------------------------- 

eon:2:~>cat /mnt/eon0/.exec

# ident     :  @(#).exec  ver 1.0.08   c:12.08.07       m:07.04.15   ASL

# add (non-interactive) commands here to run at boot (rc3: run level 3)

# commented lines are ignored

# define zpool name for binary kit, zfs swap, dump

export POOL=pool0

### update intel CPU microcode

/usr/sbin/ucodeadm -u /platform/i86pc/ucode/intel-ucode.txt 

- I did not try zool import -a

- maybe I've forgotten too much, but isn't there some analogy to mounting on unix? meaning that it needs a dir to be mounted to? 

[ipanini]

This is what import -a yields:

------------------------------

eon:4:~#zpool import -a

cannot import 'pool00': pool may be in use from other system, it was last accessed by eon (hostid: 0x63929) on Wed Apr  6 19:36:17 2016

use '-f' to import anyway

cannot import 'pool0': pool may be in use from other system, it was last accessed by eon (hostid: 0x63929) on Wed Apr  6 19:36:17 2016

use '-f' to import anyway

On Wednesday, 6 April 2016 20:17:06 UTC+2, ipanini wrote:

[Andre Lue]

sounds like /etc/hostid is not being preserved or updimg/zpool import -f is being run out of sequnece.

Try the following

-Make sure you rail the latest updimg.sh

-import the pool(s), most likely will be -f

-run setup again, to create /etc/hostid that is now mated with the import step above

-run hostid and note what it is so you can check/compare after reboot.

-run current version of /bin/updimg.sh /mnt/eon0/boot/x86.eon

-reboot and run hostid

If all goes well the pool will automount and hostid before and after will be the same.

[ipanini]

- I forgot to check hostid, this has nothing to do with hostname? I'm checking and rebooting as I write

It's changed: before 004f3aaf 

after 00cbff4e 

- What do you mean with:

"Make sure you rail the latest updimg.sh"

Aren't we missing something with me not working with a usb stick but persistent ssd? I don't see where, but...

On Wednesday, 6 April 2016 20:17:06 UTC+2, ipanini wrote:

[Andre Lue]

since hostid differs before and after it is not being preserved and is the source of your force import issue.

persistence is stored on whatever device is mounted in /mnt/eon0 as long as it is writeable.

rail latest updimg.sh means

rail in updimg.sh

[ipanini]

checking once more:

--------------------------

eon:2:~#passwd admin

New Password: 

Re-enter new Password: 

passwd: password successfully changed for admin

eon:3:~#updimg.sh /mnt/eon0/boot/x86.eon

 updating files in /mnt/eon0/.backup to x86.eon

 backup list /mnt/eon0/.backup

 uncompressing /mnt/eon0/boot/x86.eon -> /tmp/x86.1168 ... done

 mounting /dev/lofi/1 /mnt/upd ... done

 updating /etc/svc/repository.db -> /mnt/upd/etc/svc/repository.db

 umounting /mnt/upd

 backup /mnt/eon0/boot/x86.eon.0 /mnt/eon0/boot/x86.eon.1

 backup /mnt/eon0/boot/x86.eon /mnt/eon0/boot/x86.eon.0

 compressing /tmp/x86.1168 -> /mnt/eon0/boot/x86.eon

 verifying /mnt/eon0/boot/x86.eon:       OK

eon:4:~#

----------------- 

- For what it's worth: I changed the password for admin and it survived the reboot

On Wednesday, 6 April 2016 20:17:06 UTC+2, ipanini wrote:

[ipanini]

This is ruby I guess?

Nothing has been added to this system since the virgin install today.

I tried this:

-------------

eon:1:~#rail in updimg.sh

 ERROR: no such directory /usr/local

eon:2:~#rail install rail

 ERROR: no such directory /usr/local

eon:3:~#

On Wednesday, 6 April 2016 20:17:06 UTC+2, ipanini wrote:

[ipanini]

Am I not running in circles?

/usr/local is a symlink to a pool that is not online?

------------------------

ERROR: no such directory /usr/local

eon:3:~#cd /usr/

eon:4:/usr#ls -la

total 72

drwxr-xr-x  14 root     sys          512 Apr  6 20:21 ./

drwxr-xr-x  21 root     root         512 Apr  6 20:21 ../

lrwxrwxrwx   1 root     root          10 Jul 13  2015 adm -> ../var/adm/

drwxr-xr-x   4 root     bin         6144 Apr  6 20:21 bin/

drwxr-xr-x   4 root     bin          512 Jul 13  2015 ccs/

drwxr-xr-x   3 root     bin          512 Jul 13  2015 gnu/

drwxr-xr-x  10 root     sys          512 Jul 13  2015 kernel/

drwxr-xr-x  38 root     bin         8192 Jul 13  2015 lib/

lrwxrwxrwx   1 root     root          14 Apr  6 20:21 local -> ../pool0/local

lrwxrwxrwx   1 root     root          11 Jul 13  2015 mail -> ../var/mail/

drwxr-xr-x   4 root     sys          512 Jul 13  2015 net/

lrwxrwxrwx   1 root     root          13 Apr  6 20:21 perl5 -> ./local/perl5

drwxr-xr-x   3 root     sys          512 Jul 13  2015 platform/

lrwxrwxrwx   1 root     root          15 Jul 13  2015 preserve -> ../var/preserve/

drwxr-xr-x   3 root     bin          512 Jul 13  2015 proc/

drwxr-xr-x   3 root     bin          512 Jul 13  2015 sadm/

drwxr-xr-x   4 root     bin         4096 Jul 13  2015 sbin/

drwxr-xr-x   4 root     sys          512 Jul 13  2015 share/

lrwxrwxrwx   1 root     root          12 Jul 13  2015 spool -> ../var/spool/

lrwxrwxrwx   1 root     root          10 Jul 13  2015 tmp -> ../var/tmp/

drwxr-xr-x   3 root     bin          512 Jul 13  2015 xpg4/

eon:5:/usr#

On Wednesday, 6 April 2016 20:17:06 UTC+2, ipanini wrote:

[Andre Lue]

paste the output of
cat /etc/hostid

[ipanini]

Not gonna like it:

eon:5:/usr#cat /etc/hostid

cat: cannot open /etc/hostid: No such file or directory

On Wednesday, 6 April 2016 20:17:06 UTC+2, ipanini wrote:

[Andre Lue]

run setup and after cat /etc/hostid

if it is there, which it should be if setup completes ok

run  updimg.sh and you should be set.

[ipanini]

eon:8:/etc#sysdef -h

*

* Hostid

*

  00851b5f

eon:9:/etc#setup

EON Storage setup

 

Proceed with setup (WARNING: resets current network setup) [y,n?] y

Hostname: library

Hostname: library correct [y,n?] y

Configure network interface [e1000g0] via DHCP

 (Dynamic Host Configuration Protocol) [y,n?] y

Hostname: library

DHCP: YES

Is this correct [y,n?] y

Initializing interfaces ... 

eon:10:/etc#cat /etc/hostid

# DO NOT EDIT

"_I__gd`3d7"

eon:11:/etc#

It's there after init 6.

What was wrong? I ran setup and tried to name my system eon.panini.com which is my local group.

I assume it must have gone wrong with checking DNS etc?

On Wednesday, 6 April 2016 20:17:06 UTC+2, ipanini wrote:

[ipanini]

They're online after reboot!

Thanks a lot!!

On Wednesday, 6 April 2016 20:17:06 UTC+2, ipanini wrote:

[ipanini]

But I'm missing my data in the 6 disk pool?

It only has public_html in it?

On Wednesday, 6 April 2016 20:17:06 UTC+2, ipanini wrote:

[ipanini]

I still don't have /usr/local

It seems both pool0 and pool00 are empty?

On Wednesday, 6 April 2016 20:17:06 UTC+2, ipanini wrote:

[Andre Lue]

check "export POOL=abyss" in /mnt/eon0/.exec. This controls what /usr/local is symlinked to?

I think in a previous listing it was linked to ../pool0/local

[ipanini]

export POOL=pool0

But as said:

I did not install binary kit, zfs swap or such (not yet)

So assuming the recovered pool0 is / was intact, it should refer to the not-up-to-date /pool0/local...

But since both the recovered pools came online appearing empty..

there is no /pool0/local

Is it possible to lose zpool data using zpool import -f?

Can it be there's a permission issue between old users / new OS rendering the data not viewable?

I have a real hard time the pool would have been empty already...

On Wednesday, 6 April 2016 20:17:06 UTC+2, ipanini wrote:

[Andre Lue]

import -f should not lose data. restore from backup?

[ipanini]

That is a nogo after all these years..

Well I guess the trip ends here then...

On Wednesday, 6 April 2016 20:17:06 UTC+2, ipanini wrote: