[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: aptitude update / upgrade broke my Rapbian bullseye



On Wed, Dec 7, 2022 at 12:33 AM Steve Keller <keller.steve@gmx.de> wrote:
>
> On Nov 26, I upgraded a Raspberry 4 from buster to bullseye using the
> standard procedure of edit /etc/apt/sources.list and then apt-get
> update && apt-get dist-upgrade.  Everything went fine, it ran stable
> for some days and one annoying bug in the openbox window manager
> occured less often.
>
> On Dec 1, I did aptitude update && aptitude full-upgrade and it found
> lots of dependency problems.  It seems this was mainly caused by
> changing from gcc-8 to gcc-10 and python 3.7 to python 3.9.  I accepte
> the second suggestion to solve these problems, removing a couple of
> packages, and the whole upgrade process seemed to work smoothly.
>
> But on the following reboot, it showed lots of problems:
>
> 1. dhcpcd doesn't reliably get its config, /etc/resolv.conf remains
>    empty and the eth0 interface doesn't get an IP address.
>
> 2. Because of 1. NFS mounts fail
>
> 3. With restarting dhcpcd this gets fixed but after some time
>    (e.g. hours) dhcpcd fails again repeatedly with
>
>       dhcpcd[416]: ipv6nd_sendadvertisement: No buffer space available
>
> 4. Even when dhcpcd had success and the network is configured,
>    avahi-daemon for no apparent reason eventually changes the
>    interface address.  From daemon.log:
>
> Dec  4 20:04:50 <host> avahi-daemon[321]: Withdrawing address record for 10.0.0.8 on eth0.
> Dec  4 20:04:50 <host> connmand[328]: eth0 {del} address 10.0.0.8/24 label eth0
> Dec  4 20:04:50 <host> connmand[328]: eth0 {del} route 10.0.0.0 gw 0.0.0.0 scope 253 <LINK>
> Dec  4 20:04:50 <host> connmand[328]: eth0 {del} route 10.0.0.254 gw 10.0.0.254 scope 0 <UNIVERSE>
> Dec  4 20:04:50 <host> connmand[328]: eth0 {del} route 0.0.0.0 gw 10.0.0.254 scope 0 <UNIVERSE>
> Dec  4 20:04:50 <host> dhcpcd[416]: eth0: adding default route
> Dec  4 20:04:50 <host> dhcpcd[416]: eth0: pid 0 deleted default route
> Dec  4 20:05:34 <host> avahi-daemon[321]: Registering new address record for 169.254.129.129 on eth0.IPv4.
> Dec  4 20:05:34 <host> connmand[328]: eth0 {add} address 169.254.129.129/16 label eth0 family 2
> Dec  4 20:05:34 <host> connmand[328]: eth0 {add} route 0.0.0.0 gw 0.0.0.0 scope 253 <LINK>
>
> 5. Also connmand, which I hadn't in buster, continously logs messages
>    to daemon.log every few seconds:
>
> Dec  7 05:48:34 <host> connmand[328]: Skipping server 10.0.0.254 KoD code RATE
> Dec  7 05:48:43 <host> connmand[328]: Skipping server 10.0.0.1 KoD code RATE
> Dec  7 05:48:52 <host> connmand[328]: Skipping server 10.0.0.1 KoD code RATE
> Dec  7 05:48:57 <host> connmand[328]: Skipping server 10.0.0.254 KoD code RATE
> Dec  7 05:49:06 <host> connmand[328]: Skipping server 10.0.0.1 KoD code RATE
>
> 6. The whole network seems unreliable.
>
> 7. A couple of changes in the GUI, which I don't care about much at the moment:
>
>    The Debian logo in the top left of the panel where you get the
>    start menu for app is replaced by a green arrow pointing left.
>
>    The small icons in the start menus to launch apps are missing
>
>    The entries in the panel for audio volume, network, and keyboard
>    language are missing.
>
> 8. I think there were some other minor issue which I currently don't
>    remember.
>
> Now I wonder if there is a chance to get all these things fixed or if
> a fresh re-install would be easier, faster, and more sucessful.
>
> I think, first I should remove connmand, since I don't know what I'd
> need it for.

This smells (to me) like the SDcard is starting to go bad. I've had
several SDcards fail like this. In fact I used to encounter it
frequently because I would put a swapfile on the system.

(A swapfile was better than the compiler OOM errors. Killing the
compiler stopped my work dead in its tracks. With a swapfile I
expected the SDcard to fail in about three months).

Issue (4), avahi-daemon eventually changes the interface address,
sounds like that MAC address or IP address hardening. It's the one
where the address is randomized and rotated. See [1,2].

Does setting wifi.scan-rand-mac-address=no help with the issue?

[1] https://forums.debian.net/viewtopic.php?t=138939
[2] https://linux.debian.bugs.dist.narkive.com/k2IzRoYp/bug-879484-network-manager-should-default-to-non-random-mac-address-on-wifi

Jeff


Reply to: