Incorrect Disk Usage Reporting on Ubuntu 22.04 Server clean install

In an ordinary scenario , full disks space means actually full disk space with a matter of hidden/small large files somewhere. just like this form states RecoverLostDiskSpace

I know my disk uses around 4GB out of sda (50GB formatted) and 12GB on sdb (100GB formatted) but i get No space left on device

Previously -

I have a GCP VM with 2 drives, 1 with 30GB boot + an additional 100GB mounted. main drive only has ubuntu 22.04 + mysql 8.0 installed. mysql disk is pointed to the mounted disk at /mnt/disk/mysql_data

when i df -Th it reports 100% (30GB availabile + 30GB used) used on /, when i use du and ncdu it reports that only 15GB (combined with mounted disk's data) has been used at / .

so i already found out that df can confuse the mounted drive as internal.

if so with that included i should have 50% free. i tried many solutions and ended up resizing my drive from GCP to 50GB by taking a snapshot of the 30GB disk and creating a new 50GB disk with the previous image.

Before resizing

/dev/root      ext4    29G   29G     0 100% /
/dev/sdb       xfs    100G   12G   88G  12% /mnt/disks/data_disk

After resizing

/dev/root      ext4    49G   3G     0 16.3% /
/dev/sdb       xfs    100G   12G   88G  12% /mnt/disks/data_disk

Now the reading is suddenlycorrect ?? so case closed.

Now 2 weeks later im facing the same issue with 100% usage. the data on the mounted disk is still 12GB. when i try to create a file i get No space left on device

i cannot find where the missing space went??

i tried

• looking large files
• cleaning small files
• cleaning logs
• cleaning backups
• cleaning cache
• suggested solution

here is my df -h out put

df -ih
Filesystem     Inodes IUsed IFree IUse% Mounted on
/dev/root        6.2M  122K  6.1M    2% /
tmpfs            931K     1  931K    1% /dev/shm
tmpfs            800K   633  800K    1% /run
tmpfs            931K     4  931K    1% /run/lock
/dev/sda15          0     0     0     - /boot/efi
/dev/sdb          50M  8.8K   50M    1% /mnt/disks/data_disk
tmpfs            187K    20  187K    1% /run/user/0

Funny thing, it seems to be a bug. /temp is read incorrectly. before my solution, here is the df -h out put

df -h
Filesystem      Size  Used Avail Use% Mounted on
/dev/root        49G   49G   17M 100% /
tmpfs           3.7G     0  3.7G   0% /dev/shm
tmpfs           1.5G  692K  1.5G   1% /run
tmpfs           5.0M     0  5.0M   0% /run/lock
/dev/sda15      105M  5.3M  100M   5% /boot/efi
/dev/sdb        100G   13G   88G  13% /mnt/disks/data_disk
tmpfs           745M     0  745M   0% /run/user/0

As you can see tmpfs uses less than 6GB, so i never borthered to clean it since its temp files which will get deleted automatically, plus cleaning 6GB didnt make sense that it would solve the reading error.

i just stopped mysql and ran the following,

sudo rm -r /tmp/*

now

df -h
Filesystem      Size  Used Avail Use% Mounted on
/dev/root        49G  3.3G   46G   7% /
tmpfs           3.7G     0  3.7G   0% /dev/shm
tmpfs           1.5G  692K  1.5G   1% /run
tmpfs           5.0M     0  5.0M   0% /run/lock
/dev/sda15      105M  5.3M  100M   5% /boot/efi
/dev/sdb        100G   13G   88G  13% /mnt/disks/data_disk
tmpfs           745M     0  745M   0% /run/user/0

this explains why the storage showed free after the disk upgrade, because the /tmp gets wiped on reboot.

i'm closing this as question since its resolved using the following methods under the conclusion that its a calculation bug.

the version is

Description:    Ubuntu 22.04.1 LTS
Release:    22.04
Codename:   jammy

Update

as @Macro mentioned, /tmp is a directory on the filesystem which is not necessarily the same as tmpfs, which is a filesystem. Previously i thought that /tmp is tempfs.