a new kind of "No space left on device" error

a new kind of "No space left on device" error

[Dave]

This is one I have not seen before.

When running a simple, well-tested and well-used script that makes
backups using btrfs send | receive, I got these two errors:

At subvol snapshot
ERROR: rename o131621-1091-0 ->
usr/lib/node_modules/node-gyp/gyp/pylib/gyp/MSVSVersion.py failed: No
space left on device

At subvol snapshot
ERROR: rename o259-1095-0 -> myser/.bash_profile failed: No space left on device

I have run this script many, many times and never seen errors like
this. There is plenty of room on the device:

# btrfs fi df /mnt/
Data, single: total=18.01GiB, used=16.53GiB
System, DUP: total=8.00MiB, used=16.00KiB
Metadata, DUP: total=1.00GiB, used=145.12MiB
GlobalReserve, single: total=24.53MiB, used=0.00B

# df -h /mnt/
Filesystem      Size  Used Avail Use% Mounted on
/dev/sdc2        54G   17G   36G  33% /mnt

The send | receive appears to have mostly succeeded because the final
expected size is about 17G, as shown above. That will use only about
1/3 of the available disk space, when completed. I don't see any
reason for "No space left on device" errors, but maybe somebody here
can spot a problem I am missing.

Re: a new kind of "No space left on device" error

[Henk Slager]

On Mon, Oct 29, 2018 at 7:20 AM Dave <davestechshop@gmail.com> wrote:
>
> This is one I have not seen before.
>
> When running a simple, well-tested and well-used script that makes
> backups using btrfs send | receive, I got these two errors:
>
> At subvol snapshot
> ERROR: rename o131621-1091-0 ->
> usr/lib/node_modules/node-gyp/gyp/pylib/gyp/MSVSVersion.py failed: No
> space left on device
>
> At subvol snapshot
> ERROR: rename o259-1095-0 -> myser/.bash_profile failed: No space left on device
>
> I have run this script many, many times and never seen errors like
> this. There is plenty of room on the device:
>
> # btrfs fi df /mnt/
> Data, single: total=18.01GiB, used=16.53GiB
> System, DUP: total=8.00MiB, used=16.00KiB
> Metadata, DUP: total=1.00GiB, used=145.12MiB
> GlobalReserve, single: total=24.53MiB, used=0.00B
>
> # df -h /mnt/
> Filesystem      Size  Used Avail Use% Mounted on
> /dev/sdc2        54G   17G   36G  33% /mnt
>
> The send | receive appears to have mostly succeeded because the final
> expected size is about 17G, as shown above. That will use only about
> 1/3 of the available disk space, when completed. I don't see any
> reason for "No space left on device" errors, but maybe somebody here
> can spot a problem I am missing.
What kernel and progs versions?
What are the mount options for the filesystem?
Can you tell something about the device /dev/sdc2 (SSD, HDD, SD-card,
USBstick, LANstorage, etc)?
Could it be that your ENOSPACE has the same cause as this:
https://www.mail-archive.com/linux-btrfs@vger.kernel.org/msg81554.html