Fw: [Bug 208563] New: After detaching the HDD from the machine, DHCP starts requesting every possible address available.

Fw: [Bug 208563] New: After detaching the HDD from the machine, DHCP starts requesting every possible address available.

[Stephen Hemminger]

Concerning but does not look like a kernel bug.

Begin forwarded message:

Date: Wed, 15 Jul 2020 10:24:34 +0000
From: bugzilla-daemon@bugzilla.kernel.org
To: stephen@networkplumber.org
Subject: [Bug 208563] New: After detaching the HDD from the machine, DHCP starts requesting every possible address available.


https://bugzilla.kernel.org/show_bug.cgi?id=208563

            Bug ID: 208563
           Summary: After detaching the HDD from the machine, DHCP starts
                    requesting every possible address available.
           Product: Networking
           Version: 2.5
    Kernel Version: 4.19.0-9-amd64
          Hardware: All
                OS: Linux
              Tree: Mainline
            Status: NEW
          Severity: blocking
          Priority: P1
         Component: IPV4
          Assignee: stephen@networkplumber.org
          Reporter: vcc.nicolas@gmail.com
        Regression: No

Created attachment 290289
  --> https://bugzilla.kernel.org/attachment.cgi?id=290289&action=edit  
Router DHCP Page

Hello everyone.

Yesterday i shutdown my NAS which hosts the VMS HDD, without shutting the VMS
down (forgot).
Today, i woke up with my wifi unable to give my phone a ip on the network.
Same happened for my computer, and all the other devices which had to connect.

After logging in to my router, i discovered that the 3 VMS (Debian 10.4),
requested all the available ip address through their DHCP client.

All 3 VMS did exactly the same, probably when the DHCP offer ended, and it was
time to renew it.

This completely crashed the network, making unavailable to everyone
Fortunately, my DHCP range is only half of the network, so i was able to
recover it by using a static ip on my linux machine.

If in a bigger network, or company-network, this could have caused immense
damage.

Notes:
- The 3 VMS are hosted on VMWARE ESXI 7.0, with k3s installed on each of them.
- Both ESXI and command-line "shutdown" were not able to shutdown the VMS.
- The NAS was holding their disks through iSCSI. 
- The NAS was shutdown around 3PM of yesterday.
- Network started malfunctioning around 2AM of today.
- The Network CIDR is 192.168.1.0/24, and the DHCP range is
192.168.1.2-192.168.1.99.
- The VMS names are: K3S-Master, K3S-Node1, K3S-Node2

The attachments show the IPs requested by the VMS, their console, and the
router DHCP page.
Here is the imgur album with all the screenshots of the case:
https://imgur.com/a/XThsBdz

I dont know to which field this is related, as i am not a Kernel expert.

Thanks in advance to everyone looking in to this.

-- 
You are receiving this mail because:
You are the assignee for the bug.