Solved Unable to update FreeBSD

[criley]

This has been making me crazy for the last couple of weeks. I have a server that seems slow to do port upgrades however does get them done (download time seems slow), however when I try to do a OS update via freebsd-update(8) I get the following:

-> freebsd-update --debug upgrade -r 12.0-RELEASE
Looking up update.FreeBSD.org mirrors... 3 mirrors found.
Fetching metadata signature for 11.2-RELEASE from update1.freebsd.org...
latest.ssl                                    100% of  512  B 6260 kBps 00m00s
done.
Fetching metadata index...
7782b5018d447dd196426034f86e38406b1c6fed54458e100% of  225  B 2610 kBps 00m00s
done.
Fetching 2 metadata files...
/usr/libexec/phttpget update1.freebsd.org 11.2-RELEASE/amd64/m/338abe7724254c70a75df7dd846f6d6145bcf498bfe40b9220f662e312e16b81.gz 11.2-RELEASE/amd64/m/c96c5248dcbe65e0103e79436bff79270b1b0189945453acc8802ecdb7fef124.gz
phttpget: Connection failure
gunzip: (stdin): unexpected end of file
metadata is corrupt.

I have even gone so far as to set HTTP_TIMEOUT to 7200 in my environment yet I am still running into this issue.

Suggestions?

Thanks in advance.

 

[VladiBG]

Is this server behind a NAT? Can you try to fetch some file using wget to check if there's no network related issue / interruption.

 

[criley]

I did some more research and got it working. We are sitting behind a PFSense Firewall on a Proxmox server and seems that the hardware checksum offloading on the firewall was giving us grief. I am happily updating to 12.1 now.

 

[ldd]

I did some more research and got it working. We are sitting behind a PFSense Firewall on a Proxmox server and seems that the hardware checksum offloading on the firewall was giving us grief. I am happily updating to 12.1 now.

thanks for sharing your solution, I was banging my head against a wall on this for a while. guests were running as bhyve VMs in my case, disabling rx checksum offload on the host interfaces appears to have fixed it.

 

[chief]

I've ran into something similar before where a router on the hop had its mtu set too low, random traffic would seemingly get stuck.

 

[ldd]

thanks for sharing your solution, I was banging my head against a wall on this for a while. guests were running as bhyve VMs in my case, disabling rx checksum offload on the host interfaces appears to have fixed it.

I actually spoke to soon, had another upgrade fail after changing the offload settings. At this point I'm inclined to agree with chief that it is something wrong upstream, as I've been unable to replicate it in other lab environments on very similar hardware and configurations.