ROCK - Database Issue

Hi @Brice_Lang Thank you for touching base with me and verifying that information, the insight is greatly appreciated! Apologies for the wait.

I must admit that this is the first time myself (and the team) has seen a DB get into this state, twice. In light of the fact this is such an unusual situation, I am going to discuss this with our CTO and senior devs to try to come up with some next steps for you so we can minimize the amount of “back and forth” while trying to troubleshoot the issue. The first guess would be that there is some kind of hardware problem but before we begin to venture down “that road”, let me see what the team says so we can offer the BEST advice possible.

If I am not mistaken (please due correct me if I am wrong :innocent:), your MacBook Pro had once hosted your Roon core, correct? Remind, were there any issues when Roon was hosted on that device?

-Eric

Yes, for my first year or so of using Roon, my 7 year old Macbook Pro hosted the core. While performance was kind of slow, I never had any database crashes or any big problems I can recall. After the first year, I moved the core to my bran new NUC. Both database crashes have occurred with that device.

Hi @Brice_Lang ---- Many thanks for your patience here, and my sincerest apologies for the slow response.

Moving forward, I wanted to first check in with you to see if any progress or new observations have been made since we last spoke. Secondly, as mentioned in my previous, I did have a chance to touch base with our CTO and senior devs to discuss this issue you have been experiencing with your Roon DB.

Our CTO and devs all agree that this is certainly one of the rarest cases we’ve run into (i.e a DB becoming corrupted twice) and the team is concerned that these issues you’ve encountered are in fact due to a hardware problem in the NUC (i.e SSD or RAM). In light of this the team has proposed the following:

  • If the NUC is currently “on” please power it down and leave it offline.

    • The NUC should remain in this state during this test.
  • Next, the team would like you to take restore a recent back up (preferably one just before you ran into this issue) on the MacBook Pro that had been originally hosting your Roon core.

  • With the restored backup in place please monitor how the DB behaves. If things remain stable, then this would support the team’s theory that there is a hardware issue occurring on the NUC. If you run into the same problem, then this would suggest an issue with the the install and not the hardware.

-Eric

Okay I restored my Roon database from Dropbox on my MacBook Pro, which is the device I used as Core before moving to the currently suspect Intel NUC. The database seems to be from early October. I didn’t really notice the most recent rash of bugs until Thanksgiving so this state should be a good test. So far so good. I’m playing tracks from Tidal and my laptop’s hard drive successfully. The MacBook Pro takes a couple seconds to start a song just because it’s not the best computer these days, but tracks have been playing as they should.

I have only encountered one hiccup. On just one occasion tonight, a Tidal track failed to play; I got this error message: "This track is not currently available" TIDAL error message (at 10:28pm Eastern). That is a known issue that us being worked on separately.

Since my last post, I have only once encountered the error “Tidal: This track is not currently available”. Besides that, everything is performing well and seems stable. I’ve been playing music for hours each day. Shall we give it more time, or is this enough to tell us there is a hardware issue occurring on the NUC? My NUC, SSD hard drive, and memory card should all still be under warranty.

Hi @Brice_Lang ---- I hope the holiday season has been good to you and your loved ones. All the best in 2018!

Moving forward, based on your most recent observations with the DB being hosted on the previous device (i.e the MacBook Pro) I would say that this does appear to be some kind of hardware issue with the NUC. Before proceeding forward, I have placed a feedback request with our tech team just to be 100% certain that there are no other variables here for us to explore before we begin to looking into replacing the NUC (SSD, memory card, etc). As soon as I hear back from the team I will be sure to follow up with you immediately.

-Eric

1 Like

Hey Brice – sorry for the slow response here. I just want to first reiterate again how uncommon this kind of corruption is, and how unlikely it is that you’ve experienced this twice.

The tricky part going forward is that we don’t really have a clear way to determine why this has happened. It could be bad luck, but more likely it’s something related to the hardware or environment. I assume if you were experiencing frequent power cuts, you would have mentioned that by now, but that’s the kind of thing I’d be looking for if this was triggered by environmental factors.

If this is a hardware issue, my guess would be the hard drive or memory. At this point, I think it’s probably best to move forward on one of two tracks:

  • Set up a new database and configure regular daily backups with the maximum number of backups kept (99). With a fresh database, you can run until something goes wrong, and hopefully that never happens again.

  • Replace one piece of hardware (such as the hard drive) and then follow the instructions above – new database, regular backups, see how it goes.

In either case, if something goes wrong again, we know something is still off in your hardware or environment.

I know this sounds like a lot of trial and error, but again, this is extremely uncommon, and not something we can make guarantees about. What we can guarantee is that we’ll be working with you on this until things are completely stable, even if this issue crops up again months or years from now.

Let me know how that sounds, or if you have any questions, and we’ll go from there. And thanks again for your patience here Brice!

I breathed a long sigh just now. Yes, this is a real bummer to deal with all this stuff, but you all have tried your best to help me fix this. Thank you for your effort.

On the topic of “environmental” factors, I don’t believe we have anything like “power cuts” going on. I’ve never noticed anything like that in our home. As I’ve mentioned before, I do keep my NUC inside an IKEA drawer while it runs continuously. The drawer is slightly cracked open about half an inch and also has an opening out the back. I considered this carefully and even submitted a ticket here: Long term care for my NUC/ROCK. Ultimately, I decided that wasn’t a factor that would lead to a database corruption. Should I not do that anymore? I could just decide to be ultra safe and keep the NUC out in the open.

The first of two tracks you proposed…do you mean set up a new database on my existing NUC without changing any of the hardware? I’m not sure that’s even worth doing since my database on that NUC has already failed twice. I feel like it will just happen again unless something changes. So I think I’ll go down track number two, especially since my NUC, hard drive, and memory stick, are all still under warranty. I might as well replace while I can still do it for free.

Say I replace the hard drive in my NUC with a brand new one, do I have to go through that whole process I did when I first got the NUC of updating the bios and hooking up the NUC to a mouse and monitor and keyboard? I can’t remember what all that was for but it took me so long and had to borrow a bunch of stuff like mouse, keyboard, monitor. It took hours! Or do I just snap a new hard drive in and my network will recognize the NUC and I can use a Mac laptop to set up a new database on the NUC?

Hi @Brice_Lang ---- Thank you for touching base with us.

Moving forward, we agree that “track number two” is the most logical approach for you (i.e replacing the NUC “hardware”). In regard to your question about replacing the HD and then re-setting up ROCK. When the HD is replaced you will not need to setup BIOS again, however you will need to go through the process of flashing a jump drive with the Roon OS image and installing the necessary codecs as well.

-Eric

Okay, good to know. It may a week or two (or more?) to get my hard drive replaced and to set up my NUC again. I’ll chime in then.

1 Like

I got my replacement hard drive in the mail. Same model as before from the same manufacturer, from the same seller. If I get my NUC back up and running today, I’ll chime in.

I replaced my SSD hard drive in the NUC, flashed a jump drive with the Roon OS image, installed the codecs, started Roon from scratch. It took about 1 hour and 45 minutes. I started streaming a song from Tidal for a few minutes, but before the song finished, the Roon database crashed again. Same error as before: “There was an issue loading your database”. Huge bummer.

So…what’s the next step? Replace the memory stick just like I replaced the hard drive and try this all over again? If the database fails again with a new memory stick, do I replace my NUC? Do you wish to look at my logs or run any diagnostics?

I’m just grasping at straws here, but I noticed that this webpage outlines best practices for USB Drive music storage, noting that the format called exFAT is preferred. I have never heard of USB drive formats before, but I looked my USB drive and it evidently formatted as “Mac OS Extended (Journaled)”. Is the formatting of my USB hard drive a problem? I figured not because I have always used the same drive for music storage and I didn’t run into database problems until months later.
USB Drive info

Wow, thanks for continuing to work on this with us on this Brice.

Apologies if it sounds like I am repeating myself, but just to reiterate how rare this is – I’m pretty sure we’ve seen single digit number of corrupt databases on ROCK, perhaps even less than five ever… and you’ve now seen this happen 3 times :frowning:

My advice at this point would to replace the memory and the NUC – if they’re under warranty and you can have them swapped do it, since clearly something is wrong here.

I completed the warranty claim on the Crucial memory card yesterday. The replacement should arrive Friday. Should I go about replacing the NUC now or wait till I can try again with the new memory card this weekend. I guess it does make me feel somewhat better knowing this is super rare. Cause man I am really scratching my head here.

No problem with the format of my USB music storage drive as outlined above?

FWIW, I never put my NUC back in that IKEA drawer during this latest round of use. I wanted to eliminate that question. The database still crashed even though the NUC was out in the open air.

Another thing I’ve not asked before (again, I’m grasping at straws) is it easy to install the hard drive and memory card incorrectly? I was very careful and watched a handful of YouTube videos to see where and how these components are inserted. I felt like I had solid contact at the connections points. And I guess it wouldn’t have worked successfully for months if I’d installed them incorrectly.

How rare do you think it is for a NUC to become faulty after 2-3 months of light use?

I don’t think the storage is related.

In my opinion, I would just swap the NUC is it’s possible. Your experience is such an outlier here that if starting fresh is an option, I would do it.

Okay. I’ll start the warranty return process for the NUC.

The warranty processer asked me a number of questions about my issue and one of them is “what operating software are you using?”. I told him, “Roon Optimised Core Kit, which is based on Linux”. The response I got from Intel thus far was that they don’t support any such operating system so they can’t help me out. I asked to escalate this case to a manager who will be calling me back later today. I’ll keep you up to speed, but I just wanted to let you know that however rare this case is (the database crashing), if Intel won’t honor a warranty because the NUC won’t successfully run ROCK, that’s a huge ding to the customer value position for using ROCK.

The guidance here on your website (https://kb.roonlabs.com/Roon_Optimized_Core_Kit) says:

More about ROCK’s hardware support…
We have worked with Intel to support their Intel NUC line of products. They are low power high performance easy to install units that work very nicely as a Roon Core Server. The models we support are:◦NUC5i3XXX
◦NUC5i5XXX
◦NUC6i3XXX
◦NUC6i5XXX
◦NUC7i3XXX
◦NUC7i5XXX
◦NUC7i7XXX

•Any hardware configurations other than the above are unsupported. Although they may work now, they may also stop working at any time due to updates.

Hi @Brice_Lang ---- Checking in with you to see what the manager whom you spoke to said. If there is still a problem replacing the device please let me know and I will make sure we go through the proper channels to get this figured out for you.

-Eric

1 Like

The manager gave me approval to process a warranty return. She stressed again that they do not support running any OS besides Windows and said if I run into problems again based on an OS besides Windows, they won’t help me out. So…mission accomplished for now I guess. I should get my new NUC in about 10 business days.

1 Like