Bug#876762: linux-image-3.16.0-4-amd64: tmpf file system sometimes has undeletable directories/files
Package: src:linux
Version: 3.16.7-ckt25-2+deb8u3
Severity: important
Dear Maintainer,
The tmpfs file system sometimes winds up with undeletable files. Every
now and then, deleting a file still leaves its directory entry there,
with an inode number of zero. Please see the following script(1) output
for details.
Regards,
Dianne.
root@colo18:~# rm -rf /var/spool/MIMEDefang/mdefang-v6U3qOmA030183
rm: cannot remove '/var/spool/MIMEDefang/mdefang-v6U3qOmA030183': Directory not empty
root@colo18:~# strace -fv rm -rf /var/spool/MIMEDefang/mdefang-v6U3qOmA030183
[... some output deleted ...]
openat(AT_FDCWD, "/var/spool/MIMEDefang/mdefang-v6U3qOmA030183", O_RDONLY|O_NOCTTY|O_NONBLOCK|O_DIRECTORY|O_NOFOLLOW) = 3
fstat(3, {st_dev=makedev(0, 33), st_ino=4294967295, st_mode=S_IFDIR|0700, st_nlink=2, st_uid=109, st_gid=114, st_blksize=4096, st_blocks=0, st_size=60, st_atime=2017/09/24-21:47:12, st_mtime=2017/07/29-23:52:27, st_ctime=2017/07/29-23:52:27}) = 0
fcntl(3, F_GETFL) = 0x38800 (flags O_RDONLY|O_NONBLOCK|O_LARGEFILE|O_DIRECTORY|O_NOFOLLOW)
fcntl(3, F_SETFD, FD_CLOEXEC) = 0
getdents(3, {{d_ino=4294967295, d_off=1, d_reclen=24, d_name=".", d_type=DT_DIR} {d_ino=14549, d_off=2, d_reclen=24, d_name="..", d_type=DT_DIR} {d_ino=0, d_off=3, d_reclen=32, d_name="HEADERS", d_type=DT_REG}}, 32768) = 80
^^^^ NOTE the above getdents call that returns *three* items!
getdents(3, {}, 32768) = 0
close(3) = 0
unlinkat(AT_FDCWD, "/var/spool/MIMEDefang/mdefang-v6U3qOmA030183", AT_REMOVEDIR) = -1 ENOTEMPTY (Directory not empty)
root@colo18:~# ls -1a /var/spool/MIMEDefang/mdefang-v6U3qOmA030183
.
..
root@colo18:~# rm -f /var/spool/MIMEDefang/mdefang-v6U3qOmA030183/HEADERS
root@colo18:~# rm -r /var/spool/MIMEDefang/mdefang-v6U3qOmA030183
root@colo18:~# rm -r /var/spool/MIMEDefang/mdefang-v6U3qOmA030183
rm: cannot remove '/var/spool/MIMEDefang/mdefang-v6U3qOmA030183': No such file or directory
-- Package-specific info:
Deleted because I am not running reportbug on the actual system experiencing
the problem.
Reply to: