After upgrade 9.1.1 to 9.2.1.7 no pools

[KostaM]

Hi,

I using FreeNas now since more than 1 year without any problems.
My hardware is a Intel core i3 and a external Raid USB Storage (Fantec QB-35US3R) with 4 Disk each 2TBytes in Raid 5.
I see in some emails a few new features and so I decide to upgrade my System from my first and current version 9.1.1 to the newest Version 9.2.1.7.
I use the webexplorer the make the upgrade.
After the upgrade is done and I restart the PC I get an yellow alert (atteched the screenshot incl. the Volumescreen).
If I open SSH session and start command
# zpool status
no pools available

also not Disks where found.
I check the disk all Disk are Ok and healthy.

I somebody who can help, I use freenas as MediaCenter Storage and Homecloud, I've some important contents on the disks...

Thanks in advance
Kosta

 

[SweetAndLow]

Since you are using a raid card and not following the best practices of zfs you have an unsupported configuration. Zfs needs direct access to the disks it's writing data to not the virtual disk presented to it by the raid card. Your data is most likely gone if you are indeed using raid.

 

[mjws00]

I'd try a rollback first. You should be able to hit F2 at the boot prompt and run the previous version. If that fails, I'd try a fresh 9.1.1 and import the pool.

With no dice there, try a freebsd cd, or one of the solaris zfs flavors, but check in before you try anything interesting so you don't blow up your pool. Nothing interesting has happened with the disks yet. So don't panic or do anything crazy.

Kind of looks like drivers changed or something along those lines. I have a version of that enclosure that is USB 3.0... don't know if when they changed usb support they nuked compatibility. But it wouldn't surprise me if that is where the error starts. I'd actually be tempted to pull all 4 disks to keep the pool absolutely safe. Then work on the connectivity issue with a spare disk. When you get the enclosure seen by rolling back or loading a driver... then reinstall the disks and import or just mount the pool. Then reconsider using raid like this.

Good luck, you still have options as your hardware hasn't failed. Unfortunately a wrong move here and the whole thing goes boom.

 

[KostaM]

Hi,

I try to fix the USB3 problem and now Disks are working.
The access to the shares are faster and works fine.
The only thing is the red Alert with the messages:

• KRITISCH: The following multipaths are not optimal: disk1
• KRITISCH: Der Status von Volume Ext.1 (ZFS) ist DEGRADED

Has anybody an Idea what this is??

thanks in advance
Kosta

 

[danb35]

You should not be using a USB external disk enclosure to store important data--USB is simply not reliable enough for this application. eSATA isn't going to help either, since it relies on the use of a port expander, which are known to be unreliable. Fortunately, from what I can see on the product page at fantec.eu, your enclosure isn't doing any RAID work, otherwise your situation might not be recoverable. It appears that one or more of your disks is faulty and/or offline, which is why your volume status is DEGRADED.

Log into the shell, type "zpool status", and post the output.

In the very near future, if your data truly is important, you will need to plan to migrate to appropriate hardware for FreeNAS.

 

[mjws00]

Kosta, it may be you just need to clear the checkboxes on your alerts?

Dan35, there is a raid version of that enclosure (R). My guess was that he setup raid and has 1 big drive in FreeNAS.

'zpool status' will let us know.

Reality is you are going to need to get your data off that enclosure. The current USB 3 drivers are flaky enough that they disabled them. It is a bit of a shame, as they are a very nice enclosure under other OS's. Very reliable and much faster than 1GBe.

Glad to hear you can access your data, Kosta. Get it off of your current pool and hardware. Your risk of loss is HUGE with no good options if the smallest thing goes wrong.

 

[danb35]

I doubt it's one big drive presented to the FreeNAS system, otherwise it wouldn't have a way of knowing the pool is degraded. But in any case, yes, zpool status will let us know. I also doubt clearing the checkboxes will have any effect here--the "degraded" alert only appears if a volume is actually degraded. When it's fixed, the alert goes away.