Score:2

My inode count is drastically larger than file and folder count in a mounted partition

in flag

I've been dealing with an issue whereby df is reporting a much larger usage of an exfat partition mounted onto my system from a USB drive.

df -h gives:

Filesystem      Size  Used Avail Use% Mounted on
/dev/sdb3       4.6T  4.2T  396G  92% /media/smb

Whereas du -ch reports:

3.7T    .
3.7T    total

This led me to investigate inode counts using the same methodology.

From df -i I get:

Filesystem     Inodes IUsed IFree IUse% Mounted on
/dev/sdb3         19M   17M  1.6M   92% /media/smb

and du -ch --inodes shows:

3.5K    .
3.5K    total

Intuitively, knowing the number of files on that partition, 3.5K reads as the correct number, whereas 17M from df seems wildly inflated.

How can I figure out why there's such a large disparity in inodes used?

I've restarted the machine and it persists, and similarly I've restarted all the docker containers on my machine that mount folders from this drive as volumes.

Running lsof +L1 (and lsof +L1 /media/smb) shows very few orphaned files

root@couch:~# lsof +L1 /media/smb
lsof: WARNING: can't stat() fuse.gvfsd-fuse file system /run/user/123/gvfs
      Output information may be incomplete.
COMMAND      PID USER   FD   TYPE DEVICE SIZE/OFF NLINK    NODE NAME
networkd-    845 root  txt    REG  179,2  5490352     0 2228342 /usr/bin/python3.8 (deleted)
unattende    963 root  txt    REG  179,2  5490352     0 2228342 /usr/bin/python3.8 (deleted)
unattende    963 root    3w   REG  179,2      113     0 1181000 /var/log/unattended-upgrades/unattended-upgrades-shutdown.log.1 (deleted)
Xorg        1198  gdm   28u   REG    0,1        4     0   33475 /memfd:xshmfence (deleted)
pulseaudi   1342  gdm    6u   REG    0,1 67108864     0   33746 /memfd:pulseaudio (deleted)
none        1523 root  txt    REG    0,1    17032     0   37099 / (deleted)
container 431005 root  txt    REG  179,2 10556952     0 2232052 /usr/bin/containerd-shim-runc-v2 (deleted)
container 431020 root  txt    REG  179,2 10556952     0 2232052 /usr/bin/containerd-shim-runc-v2 (deleted)
container 431046 root  txt    REG  179,2 10556952     0 2232052 /usr/bin/containerd-shim-runc-v2 (deleted)
container 431089 root  txt    REG  179,2 10556952     0 2232052 /usr/bin/containerd-shim-runc-v2 (deleted)
container 431100 root  txt    REG  179,2 10556952     0 2232052 /usr/bin/containerd-shim-runc-v2 (deleted)
container 431109 root  txt    REG  179,2 10556952     0 2232052 /usr/bin/containerd-shim-runc-v2 (deleted)
container 431144 root  txt    REG  179,2 10556952     0 2232052 /usr/bin/containerd-shim-runc-v2 (deleted)
container 431203 root  txt    REG  179,2 10556952     0 2232052 /usr/bin/containerd-shim-runc-v2 (deleted)
mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.