Replaced Drive, Can't Get Back Online

[hermanpeckel]

Folks,

I was hoping to get some help with this issue. I'm fairly new to FreeNAS. Well, to problems with FreeNAS at least. Last week I had a drive fail in my RAIDZ1 array. I went through and replaced it using a guide I found here -

http://dcprom0.blogspot.com.au/2013/09/freenas-replacing-failed-disk.html

All seem to go well. I took the old disk offline, replaced it and resilvered it. Now it seems that the array is still degraded. I'm hoping I've not done anything catastrophically wrong, but I've tried a few things and nothing seems to work (and I'm also hoping I haven't stuffed things up further).

Below is the status -

Code:

scan: resilvered 864G in 27h25m with 0 errors on Sat May  3 01:19:05 2014   
config:                                                                       
                                                                               
        NAME                                            STATE    READ WRITE CKS
UM                                                                             
        freenas                                        DEGRADED    0    0   
0                                                                             
          raidz1-0                                      DEGRADED    0    0   
0                                                                             
            gptid/9c95ded0-bcea-11e1-9f82-e4115bad3382  ONLINE      0    0   
0                                                                             
            replacing-1                                DEGRADED    0    0   
0                                                                             
              18377946949923537714                      OFFLINE      0    0   
0  was /dev/gptid/9dc2e151-bcea-11e1-9f82-e4115bad3382                       
              ada1                                      ONLINE      0    0   
0                                                                             
            gptid/9eef854b-bcea-11e1-9f82-e4115bad3382  ONLINE      0    0   
0                                                                             
            gptid/a018b8e0-bcea-11e1-9f82-e4115bad3382  ONLINE      0    0   
0                                                                             
                                                                               
errors: No known data errors                

Anyone got any ideas cause obviously I don't

Thanks in advance
HP

 

[DrKK]

Not sure why you would be following some blog post, when the procedure for replacing a failed drive is clearly outlined in our docs.

http://doc.freenas.org/index.php/Volumes#Replacing_a_Failed_Drive

 

[solarisguy]

@DrKK: that part of documentation is not up to date, thus that blog is more relevant for newer FreeNAS installation. The blogger was fair and immediately referenced the FreeNAS original documentation, and only then proceed to describe his own steps.

@hermanpeckel: Could you please tell us how many hard-drives your RAID-Z1 originally had? Also please post the output of glabel status

 

[warri]

@hermannpeckel:
The blog post actually is based on incorrect assumptions and probably insufficient knowledge of FreeNAS internals. FreeNAS always creates a swap partition on the drives, uses GPTIDs instead of whole disks, and does some 4k sector alignment - all of which the manual procedure in the blog does not cover!
This is exactly the reason why your zpool status now shows ada1 instead of a gptid for the new disk.
This might get you in trouble in the future, I'd suggest you do disk replacements from the GUI only in the future.
As DrKK states, the manual is an excellent source of information, not sure why you used some random blog post..

To your problem: The new disk has resilvered successfully and now you need to detach the old, off-lined disk to get the pool back into a healthy status. This would be point 4 in the manual:

4. If the replaced disk continues to be listed after resilvering is complete, click its entry and use the Detach button to remove the disk from the list.

You can do that manually from the CLI as well with the zpool detach command.

@solarisguy:
Can't see where the Wiki is outdated in this section. If the disk is already missing, one can directly use the replace button without the need of off-lining the disk in the first place.


EDIT: I've just added a comment to the blog post with essential the same details as outlined above, so that people actually can see the drawbacks of this approach.

 

[cyberjock]

1. I've used that manual to flawless replace more than 20 disks. I also am the one that added the section for encrypted pools. So claiming that a blog is better than our manual is a bit of a stretch in my opinion.

Second, I can guarantee you that the blog is NOT going to be kept up to date. If we change something in FreeNAS with how disk replacement *should* be done, you might find out later you screwed up badly by following that guide. We've had quite a few people lose pools because they thought they "knew better" and followed some random guys post somewhere.

Third, I don't recommend you do replacements from the CLI. Do what the manual says. You don't need the FreeNAS config file to end up out of sync with your config and later you may be unable to remove a disk. Not a good place to be at all. ;)

Fourth, using the whole disk is so f*cking stupid I won't even discuss it further. If you end up kissing your data goodbye later I will laugh.. in your face... on purpose.... when you tell me you used a blog instead of the manual. Nothing pisses me off more than people that are going to use a damn blog that says to do things structurally different than the mother f*cking manual!

Use it or get the hell out of the forum. I have MUCH better things to do than tell people to do what they should be doing anyway.. USING THE MANUAL!

closing this thread as there is nothing to be gained from continuing a conversation about one of the dumbest ideas in all of FreeNAS.