Drobo

Empty rebuild of Drobo v2 takes 12 hours?

I thought the rebuild time was supposed to be proportional? I’m upgrading a v2 drobo from 4 1.5TB drives to 4 2TB drives. I left about 500 gb of data on there just to see how it would behave (everything is backed up). It took two days to rebuild one drive replacement!

Now I deleted all the data on the drive and emptied the trash can, and I replaced another 1.5TB drive. The rebuild still says it’s going to take 12 hours? What’s up with that?

in dashboard it tells you that only 500GB is in use?

unusually long rebuild times or very slow performance are sometimes indicative of a failing drive. it may be one of the drives is having some read errors or timing out and that is dramatically increasing your rebuild time.

once it has finished with the current rebuild send DRI your diagnostic logs and they will tell you what is going on

It’s an HFS formatted volume attached to my Mac mini.

The dashboard says 1.89GB are in use right now. I could obviously just reset the whole volume b/c there’s no real data on the drobo, but I wanted to see the data protection in action. This is really weird and not behaving at all like I expected.

do you mean 1.89TB?

if so that is your issue! dashboard says there is 1.89TB in use - then its protecting 1.89TB of data.

make sure that the trash is emptied… or run a file system check, eitherway there is a lot more on that volume than you think and drobo is protecting it all for you - that is why your rebuild is taking longer than you thought

No no … literally 1.89GB … or almost nothing. I assume it must be simple file system overhead that prevents it from being zero.

@rkuo

Please run a diagnostic on your Drobo and open a case. If you reset your Drobo unit you will clear out the diagnostic log and DRI will not be able to tell you what happend

It sped up the estimate considerably in the middle of the second drive swap … going to around 3 hours total. The subsequent drive swaps took half an hour or so.

However, I just had a scary experience where the Drobo was not allowing me to copy off data … it was spontaneously disconnecting. Even after rebooting both my computer and the Drobo, the problem persisted. I seem to have solved it for the moment by switching to Firewire with my Mac Mini. I hope this isn’t some sort of pattern … it’s not funny to have a few terabytes on something that spontaneously disconnects from your computer.

Is this a good time to ask YET AGAIN that we get to read at least a portion of our OWN LOGS on our OWN EQUIPMENT without having to get DRI to tell us what is happening?

Feature Requests:
User-readable logs: http://www.drobospace.com/forums/showthread.php?tid=654
SMART Info: http://www.drobospace.com/forums/showthread.php?tid=1045
More info in Dashboard: http://www.drobospace.com/forums/showthread.php?tid=508

Your best bet is to put that in as a feature request. When any case closes with DRI you get a customer survey. In that survey there is a section for feedback on the unit, put your request in there as well.

towards the end of last year Jennifer (i think - there’s a chance it may have been mark F) said that DRI was looking in to providing logs split into two-parts readable part and a proprietary DRI on part, and that it was expect to be delivered early in the new year (i.e. at the start of this year!)

What has happened to that project?

How many more feature requests would you like? I pointed to three related ones in my post…

I"m gonna echo a few things here after my current experience with the Drobo v2.

The USB disconnection issue is unacceptable. I have no idea why switching to Firewire fixed it for now, but this is the kind of thing you cannot have in a fault tolerant environment.

The rebuild times for 500GB on 1.5TBx4 to 2TBx4 border on unacceptable. Being forced to endure a days long layout is a real recipe for disaster. At the very least the dashboard needs to show me some sort of fixed progress about what it is actually doing (aka x of y bytes processed) rather than some wildly inaccurate time estimate. That way I can tell whether it is actually doing what it is supposed to be doing.

The dashboard software itself is weird … i don’t understand the need for two windows. It should just be the one window with 3 tabs. Not fatal, but really unnecessary.

The lack of self diagnostic information is not good in a product like this. The UI says nothing about the logs being encrypted, and it is not logical for a user to think a diagnostic run would not show him the diagnostics!!! Add a message saying the logs are encrypted for technical support, or better yet, actually show us the data we’re allowed to see! Drive health indicators are extremely important in a product like this.

That’s all for now.

I agree 100%. The odd way that Data Robotics tries to keep customers dependent on their paid support was thrown into sharper relief for me recently when I worked on a ReadyNAS. The NetGear tech was happy to answer questions about how their product works, and happy to help me learn to support my client better.

There is a script to decode a diagnostic file. (They aren’t encrypted, merely obfuscated.) I just tested it with my Drobo V2 running Firmware 1.3.7. This was posted to the old DroboSpace forums. Wish I remembered who originally posted it.


<?php
$fp = fopen("$argv[1]", r);
while(!feof($fp)) {
$array = str_split(fgets($fp));
foreach($array as $char) {
echo chr(ord($char) ^ 0xC3);
}
}
?>

Save that to a file called, say, “drobo_decode.php”. Then run it from the command line with
“php drobo_decode.php DroboDiag_20101218_050600.log > diags.txt”.
Of course, this requires that you first install PHP.

Thank you all for your feedback. In regards to the making the diagnostic files public. We do appreciate your feedback and invite you to let us know specifically what you would like to see more of in the way of information displayed in the Drobo Dashboard. Your comments will certainly be forwarded to our product development team.

It would be useful to have a function where if you click on a drive in the graphic of the drobo - then it presents a window with the main information about that drive model/serial/operating temp. Possibly with a further “smart” button at the bottom to access the full SMART information of a disk (or run a basic smart test).

something similar to this: http://www.synology.com/us/products/features/images/dsm3.0_640/hdd_smart_1.jpg

with the basic disk info shown in the back window - then more advanced smart info if you click the button to load that

additionally - information on WHY a disk was ejected from an array - sudden disconnect/increasing re-allocated sectors, etc etc

Those sound like excellent additions.

if theres a way to see basic details of each individual drive, eg not just capacity (which is shown on the main bay icon), but things like serial number, or drive firmware version, that would be handy if possible.

plus any info about which drive had any bad sectors and how many (even if just basic info where drobo might have marked a sector as bad to play safe even if not truly indicative of drive failues, just for info)

actually does anyone have a diagnostic file (after php’ing etc) which they would be happy to upload, just to see what sort of actual info it contains? (if theresn anything private in there delete it first etc) :slight_smile: