Dynamic DNS has been reconfigured to work with the IPv4 only setup. The A record for rtr.himbeerserver.de is up to date again. There still is no server infrastructure. I have received the ASN and configured both IRR and RPKI, so the network can be announced as soon as I gain access to new infrastructure.
A series of software fixes has successfully been deployed to the router out-of-band. The network is back up, albeit IPv4 only instead of the expected DS-Lite configuration. This allows some internal services to be brought back online, however the main server is still failing to boot. The dynamic DNS client has a bug preventing it from resolving hostnames without IPv6 connectivity, so all A and AAAA records under the himbeerserver.de zone, particularly rtr.himbeerserver.de, are incorrect. The ASN and new infrastructure should be ready soon.
The network provider transition was unsuccessful. The router is unable to establish a PPPoE session. Most attempts simply time out. Some attempts succeed without authentication, but time out shortly afterwards because neither IPCP nor IPv6CP opens. Some other attempts authenticate using PAP (with unknown status) and immediately result in a client-side crash due to it failing to deserialize LCP option 17. Since remote packet capturing is not possible under these conditions, I am currently unable to inspect the exact traffic. A software fix is being looked into, but is going to be complicated to install due to the network not being accessible remotely. Replacement infrastructure has not been provisioned yet. As a result, the entirety of the infrastructure is offline.
The main server at alptraum.himbeerserver.de has failed and it not booting up anymore. I am unable to investigate the issue in person until at least Apr 18, 2025. No data has been lost. New infrastructure is being provisioned, but is going to take several days to be ready. As a result, the following services are unavailable:
Due to an upcoming change of network providers, downtime was already expected. In order to be able to move to different infrastructure for the transition period, I attempted to download the most recent server data. The download process suddenly slowed down to 128 kbit/s permanently. New SSH connections printed out the MOTD but did not provide a shell. An existing unprivileged shell was still responsive, but on-disk executables took several minutes to load, if at all, making remote reboots impossible. I instructed a family member to reset the server, but it failed to boot back up after several attempts. Since the network was still operational, we were able to attach the SSD of the server to a different machine and download the missing data directly. The cause of the boot failures is currently unknown since there are no more debugging options that do not involve physical access. Public-facing services are going to be moved to separate infrastructure over the course of the coming weeks. Some internal services will remain inaccessible for the foreseeable future until an on-site server can be restored.
Due to the change of network providers and required router updates, a full network outage cannot be ruled out. It is also possible for inbound traffic to be filtered. As the main server has failed already, this does not affect external users. As part of the move, the network is going to go IPv6-only. I have received my own IPv6 assignment and am currently in the process of requesting an ASN. Once this is complete, all services are going to be moved to the new netblock with two IPv4-capable edge routers, making everything accessible to IPv4-only users for the first time in six years. However, this process is going to take some time to complete. Until then, the network (including the VPN endpoint) is going to be completely unavailable to IPv4-only users, including myself.