Ronan0's picture

OSError: [Errno 16] Device or resource busy: '/etc/machine-id'

Forum: 
Tags: 
Ronan0's picture

Got the database back by using --skipfiles.

Those two day's inexplicably skipped backups are a serious matter for me. Though could have been much worse.

I am afraid I need to reconsider the lack of control I have over tklbam. The level of obscurity it brings to the underlying processes, just to bring me an easy to use interface.

From a thread before I mentioned memory problems. From monitoring this the last couple of few weeks, along with another identical system that is not running tklbam, there seems to be some interaction there between the JVM and TKLBAM that affects memory, causing it to rise over time.

But you can't really get to the bottom of it with tklbam. It is obscure. Not comprehensively documented. There seems to be a certain instability and unpredictablesness there, that is the last thing you want in a backup system.

 

Jeremy Davis's picture

But I'm glad that you managed to work around it, at least in the short term.

As I think I mentioned in the other thread, the sort of behaviour you have experienced seems to be most likely to occur in low resource scenarios. However, you also noted that it seemed your issues were related to a memory leak.

I'm not sure if I have mentioned previously, but since we released TKLBAM, over the years we have had a handfull (perhaps 5 or 6?) of cases where this sort of behaviour has been reported and not immediately resolved by using a server with increased resources (it is/was a relatively common problem on t1.micro AWS instances). The only common factor seems to be MySQL. One case early on (probably about 5-6 years ago now) ended up being a corner case bug in MySQL itself. Unfortunately we did not have the resources to work with MySQL upstream to discover and resolve the source of the bug. Plus the user disappeared without giving us enough info to proceed.

In fact, that's been the case with other users experiencing the same or similar bug too. Either they haven't shared their data or when they have we've been unable to reproduce the issue. I assume there has been an undocumented factor that we're missing when we try to reproduce. Or perhaps the issue was being caused by something that was removed when the data was sanitised prior to sharing with us?

If you can either provide clear instructions on how to reproduce the issue, and/or share a sanitised dump of your data, then when we have a chance we'd love to look into this further. If you are open to sharing data, that's probably best done privately. The best way to get in touch is via support AT turnkeylinux.org

Ronan0's picture

Anyone any idea what the error I am encountering on trying to restore is about?

 

OSError: [Errno 16] Device or resource busy: '/etc/machine-id'

 

I can restore database but not files.

 

Thanks.

Ronan0's picture

Any ideas on this? Thanks.

Jeremy Davis's picture

Have you tried to see what is making /etc/machine-id busy?

fuser and lsof are 2 utilities that can help in that regard.

Also perhaps if it's trying to overwrite that file, perhaps you could try skipping that particular file (using '--limits' with TKLBAM).

Whilst it may not be ideal, worst case scenario, I suspect that a reboot should resolve it.

Add new comment