[Borgbackup] Regular backup strange behavior (stuck)

[Bzzzz]

On Fri, 26 Jan 2024 13:31:12 +0100
Peter Albrecht <peter at crazymonkeys.de> wrote:

> Hi JY,

Hi Peter,

> how much data do you backup?

Not much, the machine in question only has ~550 GB.

> I recently added a virtual machine (50 GB) to my backup. Although
> most of the vm's virtual disk does not change, borg now always has to
> crunch all those 50 GB to find changes. This takes way longer than in
> times without this virtual disk.

That I know, I reduced the slice size especially to avoid that, but
images have not been touched since the last correct backup, so it
can't be that (and anyway the biggest being 50 GB, it wouldn't take
that much time, any modif adds around 10', not more).

> Maybe some metadata of your files changed, so borg has to check a lot
> of data now.

Again not that I know of and it can't be that, otherwise I would see
these files printing on the console with a leading 'M', but the list
has stopped, iotop doesn't show any borg activity and the HDD isn't
more active than when the machine's idle :/

> A sponatnous idea:
> You can run "borg create" via "strace":
>
>    # strace borg create ...

Ok, I'll launching it in ~15 minutes - I have some errand to run, so
I'll see if it spits something weird when coming back.

Regards,
Jean-Yves


> On 26.01.24 05:20, Bzzzz via Borgbackup wrote:
> > On Thu, 25 Jan 2024 22:48:17 +0100
> > Bzzzz via Borgbackup <borgbackup at python.org> wrote:
> >
> > I made a forced fsck of the repo disk that did not show any problem
> > and followed by a borg check with rc=0.
> >
> > I still can't see what could be the cause of borg being stuck.
> >
> > JY
> >
> >> debian : bookworm
> >> borg   : 1.2.7 (from pip)
> >> repo   : mounted on the client with nfs
> >> 	 (readable from another console,
> >> 	  so not a network PB)
> >> ========================================
> >>
> >> Hi list,
> >>
> >> One machine normally takes something like 30'~35' to backup,
> >> however today borg is up for largely more than 1 hour !?
> >> start time : 21:01:29
> >> this time  : 22:45:09
> >>
> >> The console where my script was launched shows the usual list of
> >> modified files, no recap so it is assumed to be still running,
> >> there is no error message at all and the last repo file shows one
> >> created file with the last access 2' after backup was started and
> >> that is all : $ sudo stat /NFS/BORG/data/3/3992
> >>    File: /NFS/BORG/data/3/3992
> >>    Size: 147182858 	Blocks: 287480     IO Block: 131072
> >> regular file Device: 0,87	Inode: 112984077   Links: 1
> >> Access: (0600/-rw-------)  Uid: (    0/    root)   Gid: (    0/
> >> root) Access: 2024-01-25 21:02:26.104944402 +0100
> >> Modify: 2024-01-25 21:03:07.375370789 +0100
> >> Change: 2024-01-25 21:03:07.375370789 +0100
> >>   Birth: -
> >>
> >> I already tried to kill borg and relaunch it but with the same
> >> result, it stays stuck on the last (?) modified file and nothing
> >> else occurs.
> >>
> >> What could be wrong ?
> >>
> >> Jean-Yves
> >> _______________________________________________
> >> Borgbackup mailing list
> >> Borgbackup at python.org
> >> https://mail.python.org/mailman/listinfo/borgbackup
> >
> > _______________________________________________
> > Borgbackup mailing list
> > Borgbackup at python.org
> > https://mail.python.org/mailman/listinfo/borgbackup
> _______________________________________________
> Borgbackup mailing list
> Borgbackup at python.org
> https://mail.python.org/mailman/listinfo/borgbackup