Seemingly unable to update 9.3 201412090314 to 201505130355

[Icey]

Hi all

My problem is that I'm just trying to update to the latest but I've hit two bugs:
1. I cannot update either via web or gui
2. with a clean install + db import - iscsi breaks. Possibly related to the change to iscsi daemons but I haven't knowingly switched

History:

I have run FreeNAS-9.3-STABLE-201412090314 for some time. This has been infrequently but continually upgraded since whatever was current in late 2012 (something prior to RC 9.2 I believe).

I attempted to update to 9.3-STABLE-201505130355 via the gui web based update and after that failed I attempted to use the manual update file FreeNAS-9.3-STABLE-201505130355.GUI_Upgrade.txz

Both of these resulted in a broken GUI that gave me a box with
"{"events": [], "message": "Error: no such column: account_bsdusers.bsdusr_microsoft_account", "error": true}".

I found some bugs in the tracker about this but mostly were 'cannot reproduce'. Ok fair enough, fresh install time.

I created a fresh usb stick with the 201505130355 ISO and imported my database from my previously running FreeNAS-9.3-STABLE-201412090314

However this resulted in a system where iscsi is broken. The daemon does not appear to be listening on the port. The only clues I have are: no /etc/ctl.conf exists (no idea if it should?) and the notes in the release notes about optionally switching from istgt - except I don't recall ever changing from the default and can't find any option to now.

Some logs to show the errors when I change things in the iscsi part of the gui:

Jun 8 09:21:34 freenasbox alert.py: [freenasOS.Update:1167] Cached sequence, FreeNAS-9.3-STABLE-201412090314, does not match system sequence, FreeNAS-9.3-STABLE-201505130355
Jun 8 09:21:34 freenasbox alert.py: [freenasOS.Update:789] Cached sequence does not match system sequence

Jun 8 09:43:48 freenasbox notifier: Traceback (most recent call last):
Jun 8 09:43:48 freenasbox notifier: File "/usr/local/libexec/nas/generate_ctl_conf.py", line 297, in <module>
Jun 8 09:43:48 freenasbox notifier: main()
Jun 8 09:43:48 freenasbox notifier: File "/usr/local/libexec/nas/generate_ctl_conf.py", line 136, in main
Jun 8 09:43:48 freenasbox notifier: auth_group_config(auth_tag, auth_list)
Jun 8 09:43:48 freenasbox notifier: File "/usr/local/libexec/nas/generate_ctl_conf.py", line 79, in auth_group_config
Jun 8 09:43:48 freenasbox notifier: auth.iscsi_target_auth_secret,
Jun 8 09:43:48 freenasbox notifier: UnicodeDecodeError: 'ascii' codec can't decode byte 0xa7 in position 0: ordinal not in range(128)
Jun 8 09:43:48 freenasbox notifier: ctld not running? (check /var/run/ctld.pid).


Jun 8 09:43:48 freenasbox notifier: Traceback (most recent call last):
Jun 8 09:43:48 freenasbox notifier: File "/usr/local/libexec/nas/generate_ctl_conf.py", line 297, in <module>
Jun 8 09:43:48 freenasbox notifier: main()
Jun 8 09:43:48 freenasbox notifier: File "/usr/local/libexec/nas/generate_ctl_conf.py", line 136, in main
Jun 8 09:43:48 freenasbox notifier: auth_group_config(auth_tag, auth_list)
Jun 8 09:43:48 freenasbox notifier: File "/usr/local/libexec/nas/generate_ctl_conf.py", line 79, in auth_group_config
Jun 8 09:43:48 freenasbox notifier: auth.iscsi_target_auth_secret,
Jun 8 09:43:48 freenasbox notifier: UnicodeDecodeError: 'ascii' codec can't decode byte 0xa7 in position 0: ordinal not in range(128)
Jun 8 09:43:48 freenasbox notifier: ctld not running? (check /var/run/ctld.pid).

 

[cyberjock]

Did you ever run a RC? If you did that might explain things. As a rule there is no guarantee of future compatibility with config files that ran with RCs because entries may have been added for testing that adds confusion to the system.

 

[Icey]

Possibly, back in the 9.2 days. It seems unlikely that I would survive a major update from 9.2 -> 9.3 only to get caught on a minor.

Also I would not have run anything that gave a warning saying 'you might not be able to upgrade later'.

But OK. If that's the case. What option do I have other than reconfigure a completely clean freenas build by hand? Is there ANY hope of saving parts of the config (accounts/shares/etc).

 

[jgreco]

Not particularly, unless you want to get down and dirty with SQL commands in the database.

 

[Icey]

Is there any point in logging a bug ticket for these issues ? Or am I just completely and unexpectedly SOL on my own here?

 

[jgreco]

There's no harm in submitting a bug report.

 

[Icey]

Ok, well it's been a few days and I have nothing useful from the forum, all I can do it try logging the inability to update in the bug tracker.

I was hoping for some more comprehensive help to be honest.

 

[Ericloewe]

There's not much we can do about seemingly random problems that nobody else has experienced, unfortunately.

The devs, on the other hand, might know what to do about your problem and hopefully fix it or prevent it from happening again.

 

[Tony Fontane]

Actually, I'm having this exact same issue also.

 

[Tony Fontane]

Created
Bug #10250

 

[cyberjock]

I suspect the problem is on your end somehow. I just tried one of my VMs that is running your build, and it worked flawlessly.

I'm not sure I'm buying that this is a "bug" and a bug ticket is going to turn up anything.

 

[jgreco]

Ok, well it's been a few days and I have nothing useful from the forum, all I can do it try logging the inability to update in the bug tracker.

I was hoping for some more comprehensive help to be honest.

As a backdrop to the followups others made, please note that this is a community forum, and as such, members of the user community may try to assist you. However, things going awry in the database are problematic, and while some of us have professionally dealt with complex database based middleware web apps, it'd be better to submit a bug report and see if the devs can provide some guidance on this topic.

I note that SEF already discussed @Tony Fontane 's issue and tracked that down to an external cause. Yours feels a bit different, and if you have a copy of the problematic database, that provides a promising foundation for a bug report. So, please, do consider submitting a problem report.