Hello, I have a really weird issue on my OpenBSD 6.8 - upgraded from 6.7 x86 box. Worked like champ on 6.7, problems started with 6.8. Could be coincidence of course. Tried googling, did not find out anything.
After week or so - I can not login with ssh. First I suspected that sshd had somehow borked itself (altough I have never seen such a thing happen) but as the server was headless I had no real way of verifying it. I even set up a cron job to restart sshd but that did not help, so I gave up and attached a monitor. Now, the box is again in the state that it does not allow login via ssh. However, trying to login from console does not work either. It allows me to enter username, but no password prompt appears. I can switch virtual consoles but I can not login from any console. That also explains the ssh behaviour: ssh -v [email protected] OpenSSH_8.2p1 Ubuntu-4ubuntu0.1, OpenSSL 1.1.1f 31 Mar 2020 debug1: Reading configuration data /etc/ssh/ssh_config debug1: /etc/ssh/ssh_config line 19: include /etc/ssh/ssh_config.d/*.conf matched no files debug1: /etc/ssh/ssh_config line 21: Applying options for * debug1: Connecting to 192.168.1.11 [192.168.1.11] port 22. debug1: Connection established. debug1: identity file /home/saarejy/.ssh/id_rsa type 0 debug1: identity file /home/saarejy/.ssh/id_rsa-cert type -1 debug1: identity file /home/saarejy/.ssh/id_dsa type -1 debug1: identity file /home/saarejy/.ssh/id_dsa-cert type -1 debug1: identity file /home/saarejy/.ssh/id_ecdsa type -1 debug1: identity file /home/saarejy/.ssh/id_ecdsa-cert type -1 debug1: identity file /home/saarejy/.ssh/id_ecdsa_sk type -1 debug1: identity file /home/saarejy/.ssh/id_ecdsa_sk-cert type -1 debug1: identity file /home/saarejy/.ssh/id_ed25519 type -1 debug1: identity file /home/saarejy/.ssh/id_ed25519-cert type -1 debug1: identity file /home/saarejy/.ssh/id_ed25519_sk type -1 debug1: identity file /home/saarejy/.ssh/id_ed25519_sk-cert type -1 debug1: identity file /home/saarejy/.ssh/id_xmss type -1 debug1: identity file /home/saarejy/.ssh/id_xmss-cert type -1 debug1: Local version string SSH-2.0-OpenSSH_8.2p1 Ubuntu-4ubuntu0.1 so ssh connects but can not login. The box itself is a older generic PC with couple of hard disks acting as smb server / rsynced backup for home network and web server for our band http://www.heavyweights.fi (only static html & images & video files) The web site is up even though I can not login to the box. However, the next thing that will happen is that the website vanishes. The site's dns is handled by dynamic dns provider - and the update daemon seems to break too - at least judging from external symptoms. Almost as if the kernel could not create more processes - but is that even possible ? I recall from my University course (which I took at -98 or so) that at least older Unixes had fixed size process table but I really have no idea how OpenBSD works in this sense. The home router should block all access to the box from outside world, except for httpd. I have not been able to verify this of (yet - now I have company vpn which I could use to give it a try) - so there should not be too much of outside attack possibilities causing this kind of problems. Daily root mail does not suggest any such activities either, at least to the extent I understad it. Now, I need some advice. 1) Is there a way to shut the box down gracefully? Ctrl-Alt-Delete does not do anything - should it ? I'm not really familiar with OpenBSD - last BSD I used/had root access was some SunOS on Sun 3/60 in -96 or so but I have run Linux from those days. There is no Stop key to be pressed with A either ;) 1b) How fragile is the default filesystem when power is cut ? There should not be much going on. 2) How to debug the situation ? I guess I could login to several virtual consoles and leave those logged in - preferrably with top running in one so that I could kill any rogue processes. I assume that top kills processes internally - not by forking /bin/kill. Is there anything else I could / should do ? I read the bug reporting guidelines. Sorry, I can not provide the info asked there at least right now. However, I don't know would that help too much either. This is just the kind of bug report I'd hate to receive. I personally receive too many of those in my day job as a C++ code monkey. Yes, it could be hardware issue. Any tips to rule that out ? Yes, it can be some stupidity on my part too :) I'm willing to debug this but I need some help on it. I'm fully aware that you most likely are not able to reproduce this. To me it is too easy to reproduce - I just need to wait. br, Jyrki
