[BUG] Delete /dev/null

[Nik-1989]

Last night, when I ran "(/etc/cron.daily) finished netdata-updater" on several of my servers, the file /dev/null: no such file or directory was deleted.
This caused me to lose access to the server via the ssh. Could you explain the cause? So I can prevent this from happening again.

[limebond]

FYI:
We had the same problem, where almost 800 instances were affected. What was particularly "nice" was that the SQL services and SSHD services died because the process could no longer access the device.

During my investigation, I found the following, which may explain the problem:

[root@XXXX ~]# stat /dev/null
  File: /dev/null
  Size: 0         	Blocks: 0          IO Block: 4096   character special file
Device: 0,6	Inode: 675         Links: 1     Device type: 1,3
Access: (0666/crw-rw-rw-)  Uid: (    0/    root)   Gid: (    0/    root)
...

[root@XXXX ~]# /etc/cron.daily/netdata-updater
Tue Dec  9 05:29:29 CET 2025 : INFO: netdata-updater:  Checking if a newer version of the updater script is available.
Tue Dec  9 05:29:30 CET 2025 : INFO: netdata-updater:  Downloading newest version of updater script.
....
Metadata cache created.
Tue Dec  9 05:29:35 CET 2025 : INFO: netdata-updater.sh:  Checking if native packages are still being published for this platform.
curl: (22) The requested URL returned error: 404
--2025-12-09 05:29:35--  https://repository.netdata.cloud/repos/stable/centos/10.1/x86_64/.currently.published
Resolving repository.netdata.cloud (repository.netdata.cloud)... 104.20.22.2, 172.66.170.216, 2606:4700:10::ac42:aad8, ...
Connecting to repository.netdata.cloud (repository.netdata.cloud)|104.20.22.2|:443... connected.
HTTP request sent, awaiting response... 404 Not Found
2025-12-09 05:29:35 ERROR 404: Not Found.

Tue Dec  9 05:29:35 CET 2025 : ERROR: netdata-updater.sh:
Tue Dec  9 05:29:35 CET 2025 : ERROR: netdata-updater.sh:  NETDATA CANNOT BE UPDATED ON THIS SYSTEM!
Tue Dec  9 05:29:35 CET 2025 : ERROR: netdata-updater.sh:
Tue Dec  9 05:29:35 CET 2025 : ERROR: netdata-updater.sh:  Native packages are no longer being published for this platform.
Tue Dec  9 05:29:35 CET 2025 : ERROR: netdata-updater.sh:
Tue Dec  9 05:29:35 CET 2025 : ERROR: netdata-updater.sh:  To update to the latest version of Netdata, you will need to switch to a different install type.
Tue Dec  9 05:29:35 CET 2025 : ERROR: netdata-updater.sh:  For details on how to do so, see https://learn.netdata.cloud/docs/netdata-agent/installation/linux/switch-install-types-and-release-channels
Tue Dec  9 05:29:35 CET 2025 : ERROR: netdata-updater.sh:
Tue Dec  9 05:29:35 CET 2025 : FATAL: netdata-updater.sh: FAILED TO UPDATE NETDATA:  Unable to update due to native packages no longer being published for this platform

[root@pXXXX ~]# stat /dev/null
  File: /dev/null
  Size: 0         	Blocks: 0          IO Block: 4096   regular empty file
Device: 0,6	Inode: 674         Links: 1
Access: (0644/-rw-r--r--)  Uid: (    0/    root)   Gid: (    0/    root)

It is particularly interesting to see that today (as things stand) the same script can be run again without any problems.

How can this happen?
I spent the whole night and morning correcting servers using cron.

[ktsaou]

Hi @Nik-1989 and @limebond,

We owe you an apology - both for the bug itself and for taking so long to respond here.

This was a serious bug in our updater script that should never have happened. When a download check failed, our cleanup logic mistakenly deleted /dev/null instead of a temporary file. We understand the impact this had on your servers.

The issue was also reported in netdata/netdata#21427, where we:

• Fixed the root cause on December 9, 2025 (PR #21428)
• Added automatic recovery for affected systems on December 11, 2025 (PR #21432)

We should have responded here sooner to let you know. Thank you for reporting this, and again - we're sorry.