Drobo Gen2, RPi hanging on format

[Seth Feinberg]

First I apologize if this info is specifically somewhere in this group, but I looked and couldn't find it and anything I found to be pertinent was from years ago so I thought I'd ask:

I have 4 3TB hdd's in a Drobo Gen2 and I want to connect it to a Raspberry Pi (model B) and use as a sync/backup of a folder (e.g. Movies) of my main NAS (a Synology 1812+), so EXT3 format is obviously a must. Also, I've read all the stuff saying that >2TiB LUN size is impossible with Gen1 and Gen2 Drobo's (though I'm not 100% sure what LUN is or how its different from partition) but those statements always seem to be from around 2010 and to come with the aside that it might/probably be possible with some contingencies (http://bit.ly/15O7IQk and http://drobo-utils.sourceforge.net/#id4). Mainly, that the kernel is later than 2.6.24 and that  "Enable Block Layer-->Support for Large Block Devices" is enabled.  After some Linux learnin', I've discovered that the kernel in my wheezy Raspian OS is well over 3.x.x.x and that that Support for Large Block Devices IS enabled (that was a tough one with some grep proc.d doing the trick).

Yet, when I run the script ("sh /tmp/fmtscript") that "drobom -d /dev/sda format ext3 PleaseEraseMyData" creates ("drobom list" tells me sda is my drobo), it hangs whatever command line shell I'm using (both Terminal on mac and LXTerminal on the RPi).  When I hit enter, the cursor goes to the next line but no output is shown (not even a prompt) and ctrl+c doesn't even stop it.  I have also tried running the individual commands inside the script (e.g. the first one is: "parted /dev/sda mklabel gpt") and get the same result.  I've experienced this when the Drobo was originally formatted to NTFS on PC, HFS+ on Mac and even when I erased the Drobo with the Drobo Dash on my Mac and then formatted to Fat in Mac Disk Utility (since EXT3 didn't seem to be available, but my friend claims this worked for him so I'm gonna try again tonight).  The only slight change in result was when I tried HFS+ and the drive actually mounted (though read only) and I didn't know it.  Running the "sh /tmp/ftscript" at that time actually seemed to do what I thought it would (I saw output) but it was WAY to quick (I heard it would take ~10 mins per TB and it was instantaneous).  Upon closer inspection I noticed that the script couldn't unmount the drive and basically did its thing on the recovery portion of the Drobo.

So I'm sorry this was so long but I am at my wits end and because of the infrequent posts in here, I really wanted to do as much legwork as possible before I posted.  So how do I do this?  What am I doing wrong?  Am I kidding myself with this >2TiB LUN?  My friend hasn't done the formatting on the RPi, but can read write a >TiB LUN Drobo Gen1, so it seems possible.  Perusing this forum, I've seen mention of this mhddfs (http://romanrm.ru/en/mhddfs).  Should I just bail on doing this with conventional means and do this?  How do I do this? What are the drawbacks?  I'm not familiar with this method at all and would really rather not, but am now totally and completely stumped as to how to continue (besides the "hail mary" of trying to erase with drobo dashboard, then format with Mac Disk Utility to Ext3 one more time).  

Any and all help would be GREATLY appreciated.  Thanks!

[Seth Feinberg]

FYI, my second attempt to Erase the drobo on my Mac using the Dashboard and then format using Mac's Disk Utility didn't work.  I forgot there isn't an "Erase" option, just format and reset.  So I reset it but after it was done I had 5 2.2tb volumes (weird since i have 4 3tb disks in it). So I formatted to HFS+ then went over to the Disk Utility and tried to format to exFat (I wanted to use EXT3 but it wasn't an available option).  Also I wanted to format the top level volume in Disk Manager but every time I tried to format that I got a Disk Erase Failed with the error: Invalid Request.  Also plugging in the drobo into the RPi yielded the same results as before.  It basically hangs whenever ANY "parted" command is issued in terminal (even "parted -l").  

[Seth Feinberg]

So I ended up getting it to work by essentially ignoring the drobom script entirely. Or at least any mention of the parted command (shudder).  I ran a pretty simple, "sudo mkfs.ext3 /dev/sdb2" that is a mkfs.ext3 on the storage component of the drobo.  seems to have worked, its mounted and mounts on boot so we'll see how this works going forward.  

[Tom Ostlund]

So I have followed the same advice with a couple of changes and connected the DroboS to a Ubuntu 14.04 box and it is working just fine with USB 3.0. 

I connected the DroboS to a mac and formatted the entire box to HFS+ at 18TB. 

Then I connected the box to a Ubuntu 14.04 machine and from terminal ran mkfs.ext /dev/sdx2 (the 2 being the important part!!) Waited forever and a day for that to finish! ;-)

Then I could pull the blkid from terminal and mount the DroboS through fstab just like a normal drive. 

It kept showing up as a 18 TB Volume in Nautilus. That annoyed me a little.

Then I opened up GParted and changed the label to DroboS for the ext3 partition and remounted the DroboS. 

Worked perfectly! 

[Peter Silva]

The easy way to check that your configuration is safe is to fill the disk mostly up (just copy a lot of data over and over again) at least until you have most of the blue lights lit.  Then remove most of it.   What should happen is that the blue lights will go out after a while indicating it has reclaimed the space.   If that does not happen, then no space is ever reclaimed, and your drobo will just fill up (erased stuff will never really go away.) and when all the lights are filled, you drobo will slow waaayyyy down to the point of being unusable.  At that point, you need to copy the data off to another storage unit, and re-format...

If you do not do this now, you will just find out what it does in a few years, when you fill it up.  As long as the data is somewhere else, the above symptom is not catastrophic.

It would be good to test because the vendor specifically said that 8 TiB was the max for ext3, forever.  16 TiB is only for HFS+, so you are doing an experiment.  If you can put, say 12 TiB of data on there, and then remove 8, and confirm that the lights go out, then you are pretty much good to go.

http://drobo-utils.sourceforge.net/#what-happens-when-i-used-a-bigger-lunsize

--
You received this message because you are subscribed to the Google Groups "drobo-talk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drobo-talk+...@googlegroups.com.
To post to this group, send email to drobo...@googlegroups.com.
Visit this group at http://groups.google.com/group/drobo-talk.

For more options, visit https://groups.google.com/d/optout.