[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Causes, cures and prevention of orphaned inodes?



On Sun, 2018-02-04 at 23:06 -0200, Henrique de Moraes Holschuh wrote:
> On Sun, 04 Feb 2018, Stephen P. Molnar wrote:
> > I am running Debian Stretch on am eight thread AMD GPU platform.
> > Lately, it seems if I have been plagued by surfeit of orphaned
> > nodes.
> 
> That means:
> 
> 1. "unlinked" files or directories were still open when the
> filesystem
>    had to be shutdown/made read-only.  Uncommon, unless you did
> something
>    like breaking the system by having /run on something other than
>    tmpfs, etc.
> 
> 2. Filesystem corruption, most likely due to bad RAM.
> 
> 3. Really unusual workload and usage (or *mis*usage ?) pattern that
>    results in (1), perhaps involving VMs.
> 
> This assumes XFS or ext3/ext4, which seems to be your case.
> 
> > At first I though it might be impending hard drive failure as I
> > started
> > receiving warning messages from the OS, but installing a brand new
> > drive has not solved the problem,  They seem to happen when I am
> > running  four or more apps at the same time. I have 8GB of RAM in
> > the
> > system.
> 
> Test that RAM.
> 
> > Model name:            AMD FX(tm)-8320 Eight-Core Processor
> 
> Ensure you are running the latest AMD microcode on that processor:
> Update the motherboard firmware (BIOS/UEFI) to the latest version
> available from your vendor, and *also* install package amd64-
> microcode
> from non-free.  Unless you're on the latest AMD microcode on the
> FX-8300, you cannot even run VMs safely.  Who knows what else could
> be
> broken...
> 

Thanks for your reply.

I installed memtest86+ and ran it with all of the defaults.  It took
over an hour, but no errors were reported.

I am rather hesitant about updating the BIOS/UFEI.  In fact I can't
seem to find an upgrade for the FX-8320 on the AMD web site.

Also, I'm rather hesitant about installing the amd64-package.

-- 
Stephen P. Molnar, Ph.D.
Consultant
www.molecular-modeling.net
(614)312-7528 (c)
Skype: smolnar1


Reply to: