Volume manager no longer working?

[Magnus33]

Is it just me or is it no longer working correctly?

 

[cyberjock]

What in particular is wrong? Can you provide a screenshot or error message or something to go on? I'm at iX headquarters so I do have a developer's ear for stuff like this. ;)

 

[Magnus33]

No need firefox was updating in the background and blocked the correct screen changes.

This must be my week for weird hardware and glitches grrrrrrrrrr.

One thing of note is that the newer zfs does not play nice with some seagate drives using older firmware.
If it tries updating the zfs during a install things go out the window fast so a note when upgrade for the masses to check if there is firmware updates may be advisable.

 

[cyberjock]

One thing of note is that the newer zfs does not play nice with some seagate drives using older firmware.
If it tries updating the zfs during a install things go out the window fast so a note when upgrade for the masses to check if there is firmware updates may be advisable.

Wha!? ZFS upgrades:

1. Aren't automatic
2. Shouldn't be affected by disk firmware versions.

Do you have more info on this? A link or something? This seems very odd to me, and I live for odd...

 

[Ericloewe]

No need firefox was updating in the background and blocked the correct screen changes.

This must be my week for weird hardware and glitches grrrrrrrrrr.

One thing of note is that the newer zfs does not play nice with some seagate drives using older firmware.
If it tries updating the zfs during a install things go out the window fast so a note when upgrade for the masses to check if there is firmware updates may be advisable.

This sounds rather curious. Can you elaborate on what's going on? Any error messages, weird behaviors, troubleshooting steps...
Incompatibility with some HDDs would be a serious (and exotic, to say the least) bug.

 

[Magnus33]

I don't think its incompatibility with the drive per say but rather older firmware on the drive.

I know zfs worked fine since it did so on all the other drives but the one seagate which needed a firmware update it failed to work correctly.
This caused a kernel panic in the zfs pool and required the drive to be yanked/updated and then formatted to restore its function in the nas.

Unlike older versions of freenas 9.3 is a lot more critical of the drives needing to have the latest firmware.

So i would suggest a warning for those going to upgrade to it to make sure their drives firmware is up to date.
Not hard to check and its going to be much less painful to update it before hand then fixing the mess it creates later.

 

[cyberjock]

Actually, checking for the latest firmware is more than impossible. Many companies like Seagate and WD do NOT provide any means if identifying disks that need firmware updates. Last time I used Seagate you had to check the firmware version by serial number. Not by model, not by manufacture date, by serial number. They do not release updates that do "large swaths of disks" regularly. To make matters worse, even if you check the latest firmware on their website it can say that no update exists but if you call Seagate and say you found a bug in their firmware they'll "suddenly" have an update for you.

Hard drives, as a general rule, are nearly impossible to determine that an update exists, nearly impossible to validate with any kind of automation, and generally speaking shouldn't be updated unless you *know* there is a problem that warrants an update. It's not like BIOS updates where its something that everyone does "just because". Often a firmware update will result in a change of the disk's internal structure (often adding or removing a few sectors) which can suddenly make all the data on the drive inaccessible.

 

[Magnus33]

Good points cyberjock.

For us that have been doing this for ages its possible to track down and fix but for the average Joe its a nightmare.
Tracking this mess down took me three days since the issue made the crash dump totally unreachable.

Thankfully message during the process gave a hint at what was going on.

It just goes to show just how important it is to make sure you backup and have snap shoots.
Testing is done on a backup of my main nas so i didn't loose anything other then some time.

Better to happen to me then some poor guy who not quite so tech savvy. (or as the wife puts it stubborn)

 

[cyberjock]

Don't get me wrong though.. if there was a way to know when a firmware update was available and how "critical" it would be to get it applied, I'd have a feature request in before you could blink. Unfortunately, it's not really possible. :(

 

[Magnus33]

Sadly the current method of finding out involves $#$%@@%*** and testing.

It works but i don't recommend it for long term sanity.

 

[pjc]

Actually, smartd has a database of problematic firmware versions and will log a warning if it thinks you might have buggy drive firmware, e.g.:

Oct 28 16:33:21 freenas smartd[2789]: Device: /dev/da0 [SAT], WARNING: There are known problems with these drives,
Oct 28 16:33:21 freenas smartd[2789]: THIS DRIVE MAY OR MAY NOT BE AFFECTED,
Oct 28 16:33:21 freenas smartd[2789]: see the following web pages for details:
Oct 28 16:33:21 freenas smartd[2789]: http://knowledge.seagate.com/articles/en_US/FAQ/207931en
Oct 28 16:33:21 freenas smartd[2789]: http://knowledge.seagate.com/articles/en_US/FAQ/207951en
Oct 28 16:33:21 freenas smartd[2789]: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=632758
or
Oct 28 16:33:21 freenas smartd[2789]: Device: /dev/da3 [SAT], WARNING: A firmware update for this drive is available,
Oct 28 16:33:21 freenas smartd[2789]: see the following Seagate web pages:
Oct 28 16:33:21 freenas smartd[2789]: http://knowledge.seagate.com/articles/en_US/FAQ/207931en
Oct 28 16:33:21 freenas smartd[2789]: http://knowledge.seagate.com/articles/en_US/FAQ/223651en

Thankfully, these are just test drives I don't care about, with data I don't care about.

 

[Ericloewe]

Actually, smartd has a database of problematic firmware versions and will log a warning if it thinks you might have buggy drive firmware, e.g.:

Oct 28 16:33:21 freenas smartd[2789]: Device: /dev/da0 [SAT], WARNING: There are known problems with these drives,
Oct 28 16:33:21 freenas smartd[2789]: THIS DRIVE MAY OR MAY NOT BE AFFECTED,
Oct 28 16:33:21 freenas smartd[2789]: see the following web pages for details:
Oct 28 16:33:21 freenas smartd[2789]: http://knowledge.seagate.com/articles/en_US/FAQ/207931en
Oct 28 16:33:21 freenas smartd[2789]: http://knowledge.seagate.com/articles/en_US/FAQ/207951en
Oct 28 16:33:21 freenas smartd[2789]: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=632758
or
Oct 28 16:33:21 freenas smartd[2789]: Device: /dev/da3 [SAT], WARNING: A firmware update for this drive is available,
Oct 28 16:33:21 freenas smartd[2789]: see the following Seagate web pages:
Oct 28 16:33:21 freenas smartd[2789]: http://knowledge.seagate.com/articles/en_US/FAQ/207931en
Oct 28 16:33:21 freenas smartd[2789]: http://knowledge.seagate.com/articles/en_US/FAQ/223651en

Thankfully, these are just test drives I don't care about, with data I don't care about.

The problem I see is the "may or may not" part.
Doubt isn't acceptable when dealing with something as potentially destructive as a firmware change - though this is a start.

 

[pjc]

You're right -- it's limited by the information made widely available by the manufacturer. They know it's one of the ones that MIGHT be affected, but you still have to go to Seagate's web site, punch in the serial number, etc. to see if it's affected.

And sometimes Seagate doesn't even tell you. They just say there's a newer firmware available.

So I think this is about as good an automated warning as you could get.