10.0p12 - 10.1-RELEASE - 10.1p2 failure (lost ssh, ... etc... etc...)

Hi,

I recently tried up upgrade my server. It seemed to go fine from 10.0 to 10.1-RELEASE.
I did a subsequent freebsd-update fetch / install and then on reboot ... I can ping my machine, but no longer ssh to it, the web server isn't running etc... This is colo hosted machine, so I have to take time out to travel to the data centre, very tempted to just chuck Ubuntu or Centos on there ;) Its obviously quite painful, this isn't even a major version upgrade and just effectively patching has taken out my server.

Any ideas on where to start looking if I can attach a KvM and get root access, I guess the sushi config will remain the same, just no idea as to why the patch has hosed my machine...
Anybody else have similar issues?

Thanks,
Mike
 
Hmm I had a VPS having some upgrade problems, mostly because I didn’t read the output of freebsd-upgrade(8).

The upgrade from 10.0px-> 10.1-RELEASE required me to run: freebsd-upgrade fetch install, reboot, pkg upgrade all installed software, and then re-run freebsd-upgrade install before last reboot.

Missing this broke lot of stuff like PF (pf started just fine, but failed on loading limiting rules etc)

Doing a few freebsd-upgrade(8) rollbacks, cleaning out /var/db/freebsd-upgrade and doing the upgrades the correct way (reading the output of freebsd-upgrade(8)) fixed this for me.


Hope this might help.
 
Ok, so not as bad as I thought (I did do the upgrade at about 1am in the morning, which left me a little bit tired/annoyed)... I got to data centre armed with CentOs disk and one of the guys there suggested we hook up KVM and power-cycle the machine.
So, not sure if everything is 100% but at least my server is back in the land of the living. Think I will give it 6 months before doing another upgrade...
 
I'm by no means an authority on this, but I've had this problem on nearly every updated system. I believe it has to do with daemons binding to IPV6 link-local IP addresses which are not provided by 10.1 or vice versa or something. Anyway, it's very common to have hiccups like this in the 10.1 upgrade process from my experience.
 
I have the exact same issue. After upgrading from 10.0 to 10.1 randomly my server will be unresponsive to afp and ssh. I can still ping it. Rebooting the server will fix the issues. Reviewing the logs, nothing strange appears.
 
Back
Top