ada4/ada5 always have a corrupt GPT table

[mka]

Hello, I've a strange problem with my RaidZ2. I've six Western Digital Red and Green drives and at boot up the following message is displayed on my Intel S1200V3RPS Haswell Board using the internal controller:

Code:

GEOM: ada4: the primary GPT table is corrupt or invalid. 
GEOM: ada4: using the secondary instead -- recovery strongly advised. 

GEOM: ada5: the primary GPT table is corrupt or invalid. 
GEOM: ada5: using the secondary instead -- recovery strongly advised. 

The problem is: I've already made two replaces of the specified drives but with new drives... the error messages remains exactly the same. The drives at port ada4/5 have a corrupt gpt table.

The other odd part: the drives did work many boot up times but now both drives are marked at zpool status as unavailable with this...

Code:

[root@nas] ~# zpool status
  pool: tank0
 state: ONLINE
status: One or more devices could not be used because the label is missing or
        invalid.  Sufficient replicas exist for the pool to continue
        functioning in a degraded state.
action: Replace the device using 'zpool replace'.
   see: http://illumos.org/msg/ZFS-8000-4J
  scan: resilvered 983G in 4h8m with 0 errors on Sat Oct 19 17:23:36 2013
config:

        NAME                      STATE     READ WRITE CKSUM
        tank0                     ONLINE       0     0     0
          raidz2-0                ONLINE       0     0     0
            ada0                  ONLINE       0     0     0
            ada1                  ONLINE       0     0     0
            ada2                  ONLINE       0     0     0
            ada3                  ONLINE       0     0     0
            15179309252833590570  UNAVAIL      0     0     0  was /dev/gptid/79666f48-d25a-4ddb-9b02-52042ab79a0d
            9018550940036114596   UNAVAIL      0     0     0  was /dev/gptid/3cd50f0d-37eb-4f97-9e1b-d7906e7dc527

errors: No known data errors

Why are only ada4 and ada5 affected and no matter if I replace the drives or not? What is the problem with this setup?

 

[cyberjock]

Well, start with the forum rules... post your hardware, software, and anything else that is applicable that is listed in the forum rules.

Since you have devices that are disks and not partitions, where did this pool come from? Was it created in FreeNAS 8.x or 9.x? If so, have you had to replace any disks. Your configuration is NOT the standard FreeNAS config, so something hasn't been doing "properly".

 

[mka]

FreeNAS 9.1.1 x64, Intel Pentium G3220, 16GB ECC Ram, Intel ServerBoard S1200V3RPS (internal SATA HBA Controller), 4x WD Red 4.0TB, 2x WD Red 3.0TB

Because to my understanding FreeNAS WebGui doesn't support a RaidZ2 setup of different sized hdds to one pool. For the moment a RaidZ2 capacity of 6x3.0TB does suffice but I'm planing to add more 4.0TB devices... but did't wanted 6 drives with virtually identical production dates.

According to a lot of threads in the forum (http://forums.freenas.org/threads/freenas-9-1-zfs-volume-of-different-drive-size.13549/), I did create the pool from the command line, did an zpool export tank0 and an import in the WebGui. This did work fine but after some reboots this error above appeared.

What I do find strange, the pool is still considered "ONLINE" and not "DEGRADED". And if I let the drives spin down and access the pool the former ada4 and ada5 do wake up. There is currently no real data on it, just some testing stuff files.

EDIT: I did some "zpool replace" commands to a different drive and it seemed to work after resilvering, but the error reoccurred after a few boots on the new drive as well.

Thanks

 

[mka]

Any ideas what went wrong here?

I'm currently think about booting up FreeNAS 8.x and create the pool—since I've read creating pools with that layout was support back then... and again import in FreeNAS 9.1.1

 

[cyberjock]

Any ideas what went wrong here?

I'm currently think about booting up FreeNAS 8.x and create the pool—since I've read creating pools with that layout was support back then... and again import in FreeNAS 9.1.1

^^^ That.

Despite your attempt to use the CLI, I really recommend against it because FreeNAS expects things to be a certain way. If you don't replicate those exact things perfectly you can have problems. Generally, if the GUI can do what you want, you should use it. Do disk replacements from the GUI since you can, etc. Obviously, your pool config makes the volume manager give you the middle finger, so the best advice I can give is to use 8.3, then go to 9.1. It'll easily import. I actually still have v28 pools on my 9.1.1 system in case I need to do something with the volume manager. I can go back to 8.3.2 and do what I want, then go back to 9.1. If you do a v5000 pool you can't go back to 8.x, so no volume manager for you. And as I said above, the CLI thing really is a bad idea.

 

[mka]

Thank you very much, I will try that!

Are there any plans to support the creation of such pool configurations by the WebGUI again in an upcoming FreeNAS release?

 

[mka]

I'm still trying to get this working. FreeNAS 8.3.2 doesn't recognize my NIC on my Intel C222 Haswell Board. I don't have a compatible spare NIC which isn't PCI but PCIe ...so I'm waiting for the one I ordered, just 10 bucks and might get useful after this. :-/