Tried adding log device, now have a second volume instead, how can I fix this?

[bernieke]

Hi,

I've added two SSD disks to my server, and wanted to use these as mirrored log devices for my zpool. So I tried using the volume manager to extend my zpool.

Instead I ended up with this:

Code:

[root@freenas] ~# zpool status
  pool: zpool
 state: ONLINE
  scan: none requested
config:

	NAME                                            STATE     READ WRITE CKSUM
	zpool                                           ONLINE       0     0     0
	  raidz2-0                                      ONLINE       0     0     0
	    gptid/a8ec8465-7215-11e2-afb8-0025907e9c8c  ONLINE       0     0     0
	    gptid/a9615615-7215-11e2-afb8-0025907e9c8c  ONLINE       0     0     0
	    gptid/a9d4a9ff-7215-11e2-afb8-0025907e9c8c  ONLINE       0     0     0
	    gptid/ab9de51e-7215-11e2-afb8-0025907e9c8c  ONLINE       0     0     0
	    gptid/ae3a6dc9-7215-11e2-afb8-0025907e9c8c  ONLINE       0     0     0
	    gptid/b03b80e0-7215-11e2-afb8-0025907e9c8c  ONLINE       0     0     0
	  gptid/309d1080-78e6-11e2-96d8-0025907e9c8c    ONLINE       0     0     0
	logs
	  gptid/35efb2a0-78e6-11e2-96d8-0025907e9c8c    ONLINE       0     0     0

Apparantly one of the ssd's got added as a log device, and the other as a new volume...

After removing the log device, I see this in the gui:


This looks rather bad to me... How do I fix this, how do I remove that "stripe" volume? And how can I add the log devices properly afterward?

I can't just recreate the zpool, it's currently half full and it would take me several days to get the data off...

 

[ProtoSD]

I can't just recreate the zpool, it's currently half full and it would take me several days to get the data off...

Unfortunately that probably what you're going to have to do unless someone else has a trick up their sleeve that I don't know about. I know, it sucks.

And how can I add the log devices properly afterward?

I haven't tried this myself, but I think this might have to be done from the command line.

 

[paleoN]

Unfortunately that probably what you're going to have to do unless someone else has a trick up their sleeve that I don't know about.

No tricks, although hmm...

This looks rather bad to me... How do I fix this, how do I remove that "stripe" volume? And how can I add the log devices properly afterward?

I can't just recreate the zpool, it's currently half full and it would take me several days to get the data off...

Best thing to do is mirror that SSD with the other one until the data is off and recreate the pool. That will give you redundancy in the meantime.

I haven't tried this myself, but I think this might have to be done from the command line.

I believe the latest BETAs had some improvements with this.

 

[ProtoSD]

No tricks, although hmm...

Best thing to do is mirror that SSD with the other one until the data is off and recreate the pool. That will give you redundancy in the meantime.

Very good suggestion!

 

[bernieke]

Best thing to do is mirror that SSD with the other one until the data is off and recreate the pool. That will give you redundancy in the meantime.

Yes, some googling yesterday pointed that out to me as well, so I quickly did just that.

If I'll have to recreate the pool anyway, I can add the log devices at that point, so that solves my second problem at the same time.

I've arranged for a usb <> sata docking station, which will allow me to move the data to three spare drives. It'll be a pita to move those 6tb to three drives, and back again. But at least I've still got my data...

Thanks for the advice!