Jump to content
RESET Forums (homeservershow.com)

Drobo Beware


revengineer
 Share

Recommended Posts

Hello everyone,

It has been a while since my last post. My WHS 2011 build still works great and the BIOS corruption issue I encountered still has not resurfaced. Knock on wood, I no longer think it will return. I still have not turned the WHS 2011 box into my production server because I am still waiting for the drive extender replacement. At this point Drivepool is my favorite...

 

The reason for my posting is that I blow steam every time I hear the word Drobo on the Podcast. So I wanted to tell my story with the Drobo S, which may be help others in deciding whether this product is right for them.

 

I have used a Drobo S heavily for the past 6 months in my business and I feel I have now sufficient experience to judge the product. First off, I have no issues with the $$$. My business paid the expense for the unit, so my judgement is based purely on performance.

 

I chose the product because in my business I am not permitted to hook an arbitrary device such as a NAS, WHS or other server on the network. Once you exclude these great choices, there is not a whole lot to choose from. The Drobo S is a fast USB3 device that can hold lots of data while being directly connected to my PC. It also offers dual-disk redundancy. These were sellers to me so I purchased and loaded mine with 5x 2TB disks.

 

My first hurdle came loading my 4 external drives onto the Drobo. I work with lots of scientific data, some file are large (100s of MB) others are small (few kB). Of the latter kind I have hundreds of thousands. Arguably not the best way to store data, but the format was not my choice to be made. I just have to work with them. Transferring the data took days (not surprising) and I used a program called Fastcopy, which can transfer files much faster than the Windows Explorer.

 

During the copy process I noticed on three occasions that the Drobo disconnected and was no longer accessible. The repeated problem was alarming to me, especially since every time it occurred I had to run chkdsk to fix the Drobo file system. On the third time I encountered the problem, I ran a copy over night. This was a copy of the kB-size files I mentioned above and I ended up corrupting the file system to the point that I could not recover from it using chkdsk. Chkdsk when run would only fix a few files at the time and then it would barf. The problem occurred both on Windows XP and on Windows 7.

 

Not being able to fix the problem I contacted customer service. Once I got to a higher level support person (getting the PEBKAC issue out of the way) the service was friendly and prompt. Unfortunately, after two weeks of back and forth customer service was unable to save the corrupted Drobo volume. The volume had to be reformatted. Fortunately, they were able to solve this issue without affecting the remaining volumes.

 

Here was my big lesson. All this dual-disk redundancy is worth cr*p if there is a single point of failure with the file system itself. This is exactly where WHS v1 differed. With mirroring enabled, data are independently stored on two drives and two different file systems. Since the file system is just NTFS, these disks can be read anywhere unlike the Drobo disks having a proprietary format. I was happy to have my WHS at home, which is much more reliable. For this reason alone, I will never let a Drobo into my house. (I probably could not afford it anyways.)

 

I minor issue happened a few weeks ago, when DroboCopy stopped working with the new Dashboard version 2.x. I reported the problem but never got this escalated to a higher level support person. I tried to get more forceful to reach more competent help, but the rep would not escalate without me responding to a total of 18 questions, some of which obviously had nothing to do with my problem. Instead I took the reps initial advise to use synctoy instead. Not sure if Drobo has a problem standing behing their own product, but synctoy has been working fine ever since.

 

My second lessons thus was that the tech support quality I experienced was rather variable even while under warranty. I cannot imagine that the prospects are any better once the warranty expires. After this experience I have doubt that I can get the help I need if life of my data is ever in danger. There is also the question is to what more substantial help would cost when out of warranty. With WHS not only have I never had any such issues, but I also feel I understand the system well enough that I can always help myself, something that is impossible dealing with a proprietary system.

 

I am not a big writer, often because I am short on time. I have written this post because I feel that the scope of the testing as reviewed on the Homeserver Show podcast is too limited. I hope that you found this somewhat insightful. I do not hate this product, it is still the best solution in my environment. But I think for the home there are better products out there.

 

Cheers,

rev

Edited by revengineer
Link to comment
Share on other sites

Nice write up and thanks for sharing. It is things like this that establish brand loyalty or in this case brand disloyalty. Glad to hear my predictions that your problem would not retun is coming true. Now we have to work on getting you out of your quest for a drive extender. I am glad to hear it is not too late.

Link to comment
Share on other sites

Very interesting issues. I'm sure Mario will want to address these when he sees them.

 

You make some great points here and bring up some things that I have not really tested well with my reviews. I do take a bit of issue (just a very small one) with the term "limited" testing. I've been running a Drobo in some form or fashion in my home for the last six months now and have moved large data sets on and off of it as a part of my testing. I can't say that I haven't had issues at times, but the unit does see a lot of action at my place. It's been working on both V1 and 2011 boxes as well as just Windows 7. I do understand your point however, and the units are not perfect, for that matter, neither is my testing. I guess I have a few things to fix with my testing.

 

That being said, I guess I need to post more about the actual process and loads that I put the Drobo's thru when I review. I tend to post just on the set up and usage, but leave out some of the testing and data. That bit of advice helps me in the future. I will keep that in mind for some future posts. Good call. If you have some ideas on how to make this testing better, let me know.

 

The debate is a good one and I gravitate towards Drobo because I have a test box sitting on my desk. If you have some suggestion for better test options, please DM me and I will write about them. I would be more than happy to run it through it's paces. If PopDrive or IcyDock had been more aggressive with me, I might be doing more testing with their product.

 

Thanks again for the post. It's important that we hear from everyone.

 

Jim

Link to comment
Share on other sites

I have always liked the idea of running iso image files across the LAN in order to test speeds. These are usually large files so you can get a sense of how your network is performing. So, run on over to TechNet, download 20 or so images and start moving them around.

Link to comment
Share on other sites

Thanks Mr. Wills. I will give that a shot. I also set all my Music, Photos, Videos (which includes every podcast I have ever done both RAW and Final) as well as all my Documents to sync with Drobo Copy in the Dashboard V2. Thanks for the suggestion.

Link to comment
Share on other sites

Thanks Mr. Wills. I will give that a shot. I also set all my Music, Photos, Videos (which includes every podcast I have ever done both RAW and Final) as well as all my Documents to sync with Drobo Copy in the Dashboard V2. Thanks for the suggestion.

I think it's important to have 3 kinds of data sets: one with all large files, one with all small ones, and one with a mix. Often, a device that performs well with one or two data sets will do poorly with the 3rd, etc, etc. etc.

Link to comment
Share on other sites

I think it's important to have 3 kinds of data sets: one with all large files, one with all small ones, and one with a mix. Often, a device that performs well with one or two data sets will do poorly with the 3rd, etc, etc. etc.

 

So I if I have a folder of .iso files all larger than 1GB files (say 500GB worth), a folder of music files all <5MB (another 500GB) and a folder that has documents and ripped CDs with files of different sizes (another 500GB), should that be enough?

Link to comment
Share on other sites

So I if I have a folder of .iso files all larger than 1GB files (say 500GB worth), a folder of music files all <5MB (another 500GB) and a folder that has documents and ripped CDs with files of different sizes (another 500GB), should that be enough?

sounds good. In the small file data set I would want to include a bunch of files under 10K as well. I'm sure you've noticed how they can make a computer 'churn' when copying. They can really show the effect of copying to drives with different cluster sizes. :)

Link to comment
Share on other sites

sounds good. In the small file data set I would want to include a bunch of files under 10K as well. I'm sure you've noticed how they can make a computer 'churn' when copying. They can really show the effect of copying to drives with different cluster sizes. :)

 

Any good idea where I could get a bunch of those without having to create them?

 

This has given me a great idea to have a constant data set that I use for load, transfer and storage testing. Thanks for your input.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...