The admins at Stanford ought to be made aware of this, because it is a sizeable security hole, and especially because they ought to have a bleeding edge installation. The bot managers look for vulnerabilities in software - this is no novelty. What is novel here is that they have targetted a big-ass trunk-level system. Maybe Keith is just the testing ground victim for an APT who need a backwater like bitfolk in their scale-up debug routine. Did Stanford use any Chinese routers? If so, are the Chinese responsible? Or have the Chinese router backdoors been discovered by another group who want the suspicion thrown onto the Chinese? -------------------------------------------- En date de : Mar 9.4.19, admins <admins(a)sheffieldhackspace.org.uk> a écrit : Objet: Re: [bitfolk] I know I should not take it personally but ... À: users(a)lists.bitfolk.com Date: Mardi 9 avril 2019, 13h36 I wish all the pesterees I have been monitoring came from one block. We had a run of being targeted by a botnet herder. The IP's were far too many, and far too globally diverse to summarize into a handy block. I did ensure it cost them a couple of bots though by forwarding on a size-able sample to the relevant abuse emails, looking them up via whois. For what good this does. (Very little). Wish there was a streamlined script/tool to do this. If everyone reported those that try it on, and ISP's did something about it, there would be a fraction at it. If I had more time and inclination (which I had neither) I would probably looked at the fact that they would have all been a consistent bot to see if I could reverse a bot and then take down the net, from what I had learned form the one. As my ssh was not a general use I could whitelist the ranges that would reasonably have access to it,  and port knock a disable to the whitelist to allow initial connections to be made from the wider net if we needed. Thereafter con-track allowed the session to continue. 80 and 443 I get, but what was on 7777, would that have been your ssh port by any chance ?? BTW it is difficult not to take it personally when it is something we have built and nurtured. Your feelings are fully understood. Noe where did I stash that minigun.... LOL Cheers Kirbs On 09/04/2019 04:44, Keith Williams wrote: No questions, just a bit of spleen venting. Having been on a little break to deepest province where internet is very poor, I came back to find my vps under a lot of attacks. Firstly once or twice a day a website was going down for upto 5 minutes a day. Sorted that. Fail2ban was not running for some reason (again sorted by reinstalling from Debian backports) Found that known spamming IPs were hitting it hard but also were hitting at virtual hosts that no longer exist - Apache then redirects to the default virtual host. All sorts of thing then happening including SSL timeouts etc.. Fail2ban, adding a daily updated set of addresses from a content spammer blacklist to the firewall and removing A and AAAA records where possible from Bind for those old domains. ( I had to leave some like weirdname.exmple.com as they are used by other systems such as honeytraps etc) all seemed to bring that very much under control. Some were looking for URLs that have not existed for a long long time. Hours of perusing debug logs and tracking IPs via Google persuaded me to reinstall something I have not used in a while. My SSH is quite safe, I use a different port, don't allow password sign on etc. So there is nothing listening on port 22.  So set up that any attempt there, the IP gets added to a naughtyboy set then is logged and dropped. Any future visits by that IP to any port, logged and dropped. Bit like F2B but this is more of a permaban.  Within seconds there were half a dozen IPs in the set. All in the same /21 CIDR block. The logs show them coming back up to twice a second each for at least 24 hours now. They go for ports 22.23.53, 80, 443 and 7777. That last one is particularly nasty. They have each done a couple of pings (blocked of course) The group of 3 IPs all are registered to Stanford University, So probably some students Keith _______________________________________________ users mailing list users(a)lists.bitfolk.com https://lists.bitfolk.com/mailman/listinfo/users _______________________________________________ users mailing list users(a)lists.bitfolk.com https://lists.bitfolk.com/mailman/listinfo/users -----La pièce jointe associée suit-----