Before posting this to the Fedora bugzilla, I thought I should ask if I'm missing something.
I was trying to set up
rsnapshot to do a full backup of my F19 installation to a Passport 1.5Tb USB drive. Here are the settings I was using:
Code:
$ sudo grep -vE '(^[[:space:]]*$)|(^[[:space:]]*#)' /etc/rsnapshot.conf
config_version 1.2
snapshot_root /Passport/rsnapshots/
cmd_cp /usr/bin/cp
cmd_rm /usr/bin/rm
cmd_rsync /usr/bin/rsync
cmd_logger /usr/bin/logger
cmd_du /usr/bin/du
cmd_rsnapshot_diff /usr/bin/rsnapshot-diff
interval hourly 6
interval daily 7
interval weekly 4
interval monthly 3
verbose 2
loglevel 3
logfile /var/log/rsnapshot
lockfile /var/run/rsnapshot.pid
rsync_short_args -aAHX
rsync_long_args --delete --numeric-ids --relative --delete-excluded
du_args -csh
one_fs 1
exclude /proc/
exclude /tmp/
exclude /home/*/tmp/
exclude /dev/
exclude /var/run/
exclude /[[:upper:]]*/
link_dest 1
use_lazy_deletes 1
backup / localhost/Passport/rsnapshots/
When I ran
sudo rsnapshot -V hourly I got this in the log file:
Code:
$ grep '16:42' /var/log/rsnapshot
[21/Jun/2013:16:42:26] /bin/rsnapshot -V hourly: started
[21/Jun/2013:16:42:26] echo 30899 > /var/run/rsnapshot.pid
[21/Jun/2013:16:42:26] mv /Passport/rsnapshots/hourly.0/ /Passport/rsnapshots/hourly.1/
[21/Jun/2013:16:42:26] /bin/rsnapshot -V hourly: ERROR: Could not rename("/Passport/rsnapshots/hourly.0", "/Passport/rsnapshots/hourly.1")
[21/Jun/2013:16:42:26] /bin/rsnapshot -V hourly: ERROR: Error! safe_rename("/Passport/rsnapshots/hourly.0/", "/Passport/rsnapshots/hourly.1/")
[21/Jun/2013:16:42:26] /usr/bin/logger -i -p user.err -t rsnapshot /bin/rsnapshot -V hourly: ERROR: Error! safe_rename("/Passport/rsnapshots/hourly.0/", "/Passport/rsnapshots/hourly.1/")
[21/Jun/2013:16:42:26] rm -f /var/run/rsnapshot.pid
So I tried to see what happened to the Passport drive:
Code:
$ sudo mv /Passport/rsnapshots/hourly.0/ /Passport/rsnapshots/hourly.1/
mv: cannot move ‘/Passport/rsnapshots/hourly.0/’ to ‘/Passport/rsnapshots/hourly.1/’: No space left on device
$ sudo umount /Passport
$ sudo btrfsck /dev/sdd1
Checking filesystem on /dev/sdd1
UUID: 926e48e5-dc5f-4b6b-a137-918474444271
checking extents
checking fs roots
checking root refs
found 368033935360 bytes used err is 0
total csum bytes: 357478568
total tree bytes: 1610190848
total fs tree bytes: 1075429376
btree space waste bytes: 377697383
file data blocks allocated: 366423744512
referenced 367877095424
Btrfs v0.20-rc1
$ sudo mount /Passport
$ df -h /Passport
Filesystem Size Used Avail Use% Mounted on
/dev/sdd1 1.4T 345G 1.1T 25% /Passport
$ df -h /Passport/rsnapshots/
Filesystem Size Used Avail Use% Mounted on
/dev/sdd1 1.4T 345G 1.1T 25% /Passport
$ sudo df -h /Passport/rsnapshots/hourly.0/
Filesystem Size Used Avail Use% Mounted on
/dev/sdd1 1.4T 345G 1.1T 25% /Passport
$ sudo du -csh /Passport/rsnapshots/hourly.0/
59G /Passport/rsnapshots/hourly.0/
59G total
So, does anyone have have any clue why the
mv command (which is only renaming the directory) "thinks" that it's going to run out of space whilst doing the move?
Even if a full copy of the 59Gb directory were to be made (which should not be done), there is, apparently,
1.1 Tb free on the drive.
Strange "no space available" message using btrfs on Fedora 19
