On Thu, 24 Jan 2019 11:29:50 +0000
unman <un...@thirdeyesecurity.org> wrote:

> On Thu, Jan 24, 2019 at 01:00:15AM -0500, Chris Laprise wrote:
> > On 01/23/2019 08:15 PM, js...@bitmessage.ch wrote:  
> > > Mike Keehan:  
> > > > Hi,
> > > > 
> > > > I'm using Qubes Backup to save some of my qubes into another VM.
> > > > The backup VM has 18 Gb of storage available, but whenever the
> > > > backup file reaches 3Gb, the backup process just hangs.
> > > > 
> > > > No CPU usage, no error messages, just stops.  The backup window
> > > > shows 40% complete, but never moves any further (different % for
> > > > different combinations of qubes in the backup list).
> > > > 
> > > > After waiting a considerable time (well, 5-10 minutes), hitting
> > > > Cancel in the backup window does cancel it.  The rest of the
> > > > system is continuing to work without problem.  Happens every
> > > > time I try to use Qubes backup.
> > > > 
> > > > The Qubes Disk Space widget shows less than 50% disk used
> > > > overall, the backupVM shows only 18% disk used when the 3Gb
> > > > file has been saved.
> > > > 
> > > > I'm stumped.
> > > > 
> > > > Mike.  
> > > 
> > > Hi,
> > > 
> > > You may have to wait longer than 5-10 minutes. I experience
> > > something similar when doing a full backup, except it's worse
> > > because i'm backing up like 2.5TB. It appears to hang for several
> > > hours at a time (and this happens more than once), but it does
> > > eventually make visible progress again. The whole process takes
> > > over 24 hours. This is why i do full backups very infrequently.
> > > 
> > > For you it shouldn't take nearly as long because it's a lot less
> > > data, but the progress appearing to hang for a while seems to be
> > > normal.
> > > 
> > > I'm using 3.2 tho, and i know they made changes to the backup
> > > mechanism under the hood in 4.0, so i'm not sure if this issue
> > > still applies in 4.0.  
> > 
> > Marek,
> > 
> > Isn't this the null bytes bug in GNU tar?
> > 
> > https://groups.google.com/d/msgid/qubes-users/f4d997d5-7191-06d0-e7bb-ef42745a7db5%40posteo.net
> > 
> > It would be a good idea to update this in dom0. My own backup tool
> > uses GNU tar as well.
> > 
> > -- 
> > 
> > Chris Laprise, tas...@posteo.net
> > https://github.com/tasket
> > https://twitter.com/ttaskett
> > PGP: BEE2 20C5 356E 764A 73EB  4AB3 1DC4 D106 F07F 1886  
> 
> It seems a little early to judge.
> 
> Mike - it looks from your comment as if you have been trying with
> subsets of the qubes? Can you confirm if these are running or stopped.
> 
> Like jsnow, I'm regularly able to backup far more than 3G without
> issue, so it looks as if there's something particular about this
> scenario. It would be helpful if you could confirm the issue when all
> qubes you are backing up are stopped.
> Then try bisecting the qubes backup group - keep bisecting if you hit
> the problem again until you either find the problematic qubes or have
> backed them all up.
> 

OK, progress:)

I can backup a list of stopped qubes, running out to a 10Gb file without
any issues.  They all verify OK too.  However, backing up running qubes
exhibits the problem.  

Starting up one of these qubes and trying to backup causes the progress
indicator to stop at some point, BUT, the data is still being backed up
to disk, and continues flowing for some time.  When the data flow stops,
then I have to cancel the backup operation.  However, verifying the
backup works OK - file size reported is correct, and the verify process
finishes successfully.  I haven't tried to actually restore one of
these yet.

I tried recording the process status of both dom0 and the backup vm
during the backups, but could not see any particular process dying when
the progress updater stopped moving.  The tar processes stopped in dom0
but I guess that was because they had finished creating the private.img
files.

So it looks like backing up a live VM causes the backup process to fail
at some point, but not before the data is actually backed up.  

It doesn't look like a tar null issue as the data does get to disk OK.
It's just the control of the backup process getting out of step
somewhere.

Mike.

-- 
You received this message because you are subscribed to the Google Groups 
"qubes-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to qubes-users+unsubscr...@googlegroups.com.
To post to this group, send email to qubes-users@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/qubes-users/20190125135255.23ddcca0.mike%40keehan.net.
For more options, visit https://groups.google.com/d/optout.

Reply via email to