Hi, I have a Cloud VPS project (statanalyser, which is used for https://meta.wikimedia.org/wiki/Global_statistics), and am requesting an extension of about two-three weeks for T401804 (was told to file a Phabricator task). This is because I am working on hopefully migrating the project to Toolforge (and got T433184 approved for the reason), but I am not sure whether that will suffice. If it turns out that Toolforge won't work, I intend to follow the OS migration route.
Description
Description
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Open | None | T401804 Cloud VPS Debian Bullseye deprecation | |||
| Open | None | T434103 Shutdown timeline extension requests for maintained projects | |||
| Open | None | T433570 Debian Bullseye deprecation extension for statanalyser |
Event Timeline
Comment Actions
Hello @Leaderboard -- I see that this VM did not survive the in-place upgrade and is now off the network. If you need assistance with data rescue or VM recovery please let me know... I don't immediately know how to fix it but I can access the console at least.
Comment Actions
In-place upgrade @Andrew? Sorry, why was the VM upgraded without asking me? Is the data safe?
(Plus I thought that in-place upgrades weren't allowed in the first place)
Comment Actions
Hello @Leaderboard!
I haven't done anything on that VM other than investigate after it started reporting as down this morning.
The VM was built Jan. 12, 2023, 4:05 p.m. from a Bullseye base image, but if I now ask it what OS it is running it reports Trixie with a fairly old kernel:
root@statanalyser:/var/log/apt# lsb_release -a No LSB modules are available. Distributor ID: Debian Description: Debian GNU/Linux 13 (trixie) Release: 13 Codename: trixie root@statanalyser:/var/log/apt# uname -a Linux statanalyser 6.12.12-cloud-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.12.12-1 (2025-02-02) x86_64 GNU/Linux
So... at some point in the past, possibly as long as year ago, someone (you?) seems to have done an in-place upgrade from Bullseye to Trixie (or possibly even from Bullseye to Bookworm and then from Bookworm to Trixie, I have no idea.)
I don't know if it's from that upgrade, but right now it has some confusion about nic names, which seems very likely to be some sort of OS-version disagreement. It's running the 'networking' service but not 'systemd-networkd' which is the normal network agent for Trixie.
The networking service wants you to have a network interface named 'ifb0' which is a mystery to me. I have, however, run 'dhclient' to force dhcp renewal and that seems to have repaired the network for the moment.
In terms of the deprecation mentioned on T401804... you are in theory now running Trixie and so not subject to shutdown. That said, that host is a bit of a hybrid monster at this point so a rebuild is probably in order :)
Comment Actions
OK thanks for the context; I did not make any changes recently so was not sure what happened. I thought I was on Bullseye all this time.