df in linux not showing correct free space after file removal

Question

I have file servers which are used to store files. Files might reside there for a week, or for a year. Unfortunately, when I remove files from the server, df command doesn't reflect the freed up space. So eventually, the server gets filled up (df shows 99%), and my script doesn't send any more files there, except there might be a few dozen GB of free space on there.

I got noatime flag on the mounted partitions if that makes any difference.

Solution 1

Deleting the filename doesn't actually delete the file. Some other process is holding the file open, causing it to not be deleted; restart or kill that process to release the file.

Use

lsof +L1

to find out which process is using a deleted (unlinked) file.

Solution 2

as Ignacio mentions, deleting the file won't free the space until you delete the processes that have open handles against that file.

Nevertheless, you can reclaim the space without killing the processes. All you need to do is to remove the file descriptors.

First execute lsof | grep deleted to identify the process holding the file

[hudson@opsynxvm0055 log]$ /usr/sbin/lsof |grep deleted
java       8859   hudson    1w      REG              253,0 3662503356    7578206 /crucible/data/current/var/log/fisheye.out (deleted)

Then execute:

cd /proc/PID/fd

then

[hudson@opsynxvm0055 fd]$ ls -l |grep deleted
total 0
l-wx------ 1 hudson devel 64 Feb  7 11:48 1 -> /crucible/data/current/var/log/fisheye.out (deleted)

The "1" will be the file descriptor. Now type "> FD" to reclaim that space

> 1

You might need to repeat the operation if there are other processes holding the file.

Solution 3

If partition has been configured to reserve certain portion of disk space only for root usage, df will not include this space as available.

[root@server]# df -h
Filesystem            Size  Used Avail Use% Mounted on
...
/dev/optvol           625G  607G     0 100% /opt
...

Even after space will be reclaimed by deleting files/directories, non-root user won't be able to write to particular partition.

You can easily check if that's your case by trying to create a file on a device as root and non-root user.

Additionally you can check filesystem configuration by running

tune2fs -l <device> | egrep "Block count|Reserved block count

and calculating actual % on your own.

To change disk % reserved for root-only usage, execute

tune2fs -m <percentage> <device>

Solution 4

The file is still locked by the process opening it. To free up space, do these steps:

1. Run sudo lsof | grep deleted and see which process is holding the file. Example result:

   $ sudo lsof | grep deleted
   COMMAND   PID USER   FD   TYPE DEVICE SIZE/OFF      NODE NAME
   cron     1623 root    5u   REG   0,21        0 395919638 /tmp/tmpfPagTZ4 (deleted)
   

2. Kill the process using sudo kill -9 {PID}. In above sample, the PID is 1623.

   $ sudo kill -9 1623
   

3. Run df to check if space is already freed up. If it's still full, maybe you need to wait a few seconds and check again.

Solution 5

One possibility is that the file(s) you deleted have more references in the filesystem. If you've created hardlinks, several filenames will point to the same data, and the data (the actual contents) won't be marked as free/usable until all references to it has been removed. Before you delete files, either stat them (Entry named Links) or do ls -l on them (should be the second column).

If it does turn out that the files are referenced elsewhere, I guess you'll have to ls -i the file(s) to find the inode-number, and then do a find with -inum <inode-number> to find the other references to that file (you probably also want to use -mount to stay within the same filesystem as well).

View more solutions

Comments

[Khaled]

Is this happening on a single partition or on all partitions?

[Admin]

Well, its happening on my main data partition, which is the only one I care about, since I only write/remove files onto it.

[Muhammad Danish]

What filesystem(s)? DF does a stat of the superblock, it may be that your filesystem is not updating the sb inode. Have you tried flushing cache?

[Admin]

Using ext4. How do you flush caches?

[shoover]

Joined this Stack today just to upvote this question and its answers because I was stuck and this got me unstuck.

[AncientSwordRage]

Does this answer your question? df says disk is full, but it is not

[Admin]

Files that are removed were not accessed in over a month, and the only process that accesses them is nginx, so its doubtful.

[pehrs]

+1. Also, "lsof +L1" will tell you which program is holding the files open.

[ScottZ]

as root run "lsof -n | grep file," you'd be surprised at how long files can stick around due to processes keeping them open for whatever reason. If all else fails, reboot, I feel bad suggesting it but it will definitely make sure nothing is holding on to the file. Per pehrs, lsof +L1 is probably the better way to go.

[Admin]

That returns a single file, which is being encoded with ffmpeg.

[Deer Hunter]

Dear Marcellus, your solution is encompassed by the accepted answer; and sometimes you don't want to do a reboot if you are not forced to...

[Andrew B]

This doesn't really add much; the accepted answer covered the logic behind that in 2001. When you have 50 rep, use comments if you want to add qualifiers to the existing answers.

[Luke Cousins]

You just saved me! Deleted a 93G log file and didn't get the space back and couldn't work out why. Thanks.

[Nick]

Along the same lines and in case this helps others, I erased a large nginx access.log file but was only able to reclaim the space after restarting nginx: service nginx restart

[FilBot3]

what does the > FD do?

[Adrián Deccico]

it removes the file descriptor

[Ashish Karpe]

I have same issue with apache2 access logs, ihad even stopped n started appache still no space returned i am using ext4

[ariera]

does this > command have a name? i had to switch from zsh to bash in order to be able to use it. Is it possible to run it on zsh?

[kontinuity]

Never knew about this. Learning something every day.

[aross]

It helps to know that you need to run this as root/sudo. I could guess and try, but I generally don't try stuff with sudo if I don't know the consequences.

[Aaron Hudon]

it is a output redirect and therefore truncates the file. The long from would be "echo -n > 1" or "true > 1". It does not really remove the FD, it just points to a empty file afterwards.

[Chaminda Bandara]

I got a error -bash: lsof: command not found

[dr0i]

One shouldn't say "thank you" in comments. Here I do it nonetheless.

[chiappa]

I got an easier to read format in MB with this command; lsof +L1 | numfmt --field=7 --to=iec --invalid=ignore

[J.Ko]

... so how do you close those processes?

[Michael Hampton]

This will just kill a bunch of programs. You won't be able to know which ones in advance. It's also not clear why you would want to do this.

[peterh]

Script copy-paste is not a very well going thing, I think. I suggest to explain, what is your command doing and how.

[enharmonic]

Run sudo lsof +L1 to see non-user processes, as well, and then sudo kill [pid] the relevant PID.

[Admin]

+1. Note to ppl like me: Before doing > 1 it's important that we are in the /proc/<PID>/fd dir, as it's mentioned in the answer. I usually avoid going deep in dirs and use appropriate paths in commands, which doesn't work in this case.