2025-06-04 GnuCash IRC logs
00:20:24 *** chris has joined #gnucash
00:20:24 *** ChanServ sets mode: +v chris
00:41:42 *** chris has quit IRC
00:44:58 *** chris has joined #gnucash
00:44:58 *** ChanServ sets mode: +v chris
01:11:20 *** chris has quit IRC
01:50:16 *** fell has quit IRC
01:51:35 *** fell has joined #gnucash
01:51:35 *** ChanServ sets mode: +o fell
03:14:26 *** tmrsm has quit IRC
03:16:51 *** tmrsm has joined #gnucash
03:52:21 *** chris has joined #gnucash
03:52:21 *** ChanServ sets mode: +v chris
05:04:29 *** chris has quit IRC
05:07:45 *** chris has joined #gnucash
05:07:45 *** ChanServ sets mode: +v chris
05:58:31 *** chris has quit IRC
06:04:27 *** chris has joined #gnucash
06:04:27 *** ChanServ sets mode: +v chris
06:11:00 *** chris has quit IRC
06:16:49 *** chris has joined #gnucash
06:16:49 *** ChanServ sets mode: +v chris
06:38:49 *** chris has quit IRC
07:28:38 <Simon> I think you're going to have to implement Anubis
09:05:19 <warlord> Simon, Interesting. An hour ago it was fine. Now the load is at 130!
09:06:23 <warlord> Looks like mysqld is the bottleneck
09:06:36 <warlord> Or at least, via top, that's what's at the top of the top list
11:13:05 *** chris has joined #gnucash
11:13:05 *** ChanServ sets mode: +v chris
11:13:05 *** gncbot sets mode: +o chris
11:30:59 <warlord> I gave code more V-Ram... That seems to have helped!
11:50:54 <Simon> They'll just make more requests
11:58:43 <warlord> Perhaps...
11:58:50 <warlord> But it's helped the immediate load,
11:59:09 <warlord> Instead of load-level at 120-150 we're down to about 30
11:59:22 <warlord> I also added more vCPUs, but that won't take effect until the next reboot.
12:07:57 <warlord> Simon, alas, you were too prophetic.. They caught up pretty quickly and now it's back to load of 150
12:09:54 <fell> Is there a possibility to load-balance by IP?
12:14:19 <chris> warlord time to raid the armory
12:15:43 <warlord> fell, i'm not sure how that would work -- we only have a single server.
12:17:21 <warlord> chris, I wish the IETF had actually implemented the push-back protocol that was under discussion 20+ years ago.
12:17:40 <warlord> But that doesn't help in a DDoS..
12:17:55 <warlord> And lookie -- load is back down to 4! So the added RAM did help.
12:19:16 <Simon> At some point I'm going to have to do something about my own cgit website because the bots are relentless
12:20:02 <warlord> They are. And fail2ban is not sufficient.
13:05:37 <warlord> Whee... up to 220.... Now back to 100... 90...
13:06:10 <warlord> I think I'm going to reboot code so I can get it to take the vCPU upgrade.
13:06:42 <fell> OK for me
13:07:38 *** gncbot has joined #gnucash
13:07:53 *** warlord sets mode: +o gncbot
13:10:49 <warlord> Doubled the RAM and tripled the CPUs.
13:12:00 <warlord> Let's see if that helps?
13:12:28 <warlord> Wish I had a nice load-monitor graph.
13:20:06 *** chris has quit IRC
13:38:48 <Simon> you can't ban them because they come from so many different IPs and all pretend to be browser User-Agents
13:39:31 <Simon> all you can really do is prevent them getting the next list of URLs to defeat them
13:40:03 <Simon> aiui they're training LLMs and have no cache whatsoever which is why they keep downloading the entire Internet over and over again
13:47:45 <warlord> So Stupid.
13:50:52 <warlord> And they clearly keep pushing until the server stops responding.
13:50:56 <warlord> Load is back up to 45.
13:51:18 <warlord> It's pulling down bugzilla too.
23:10:00 *** jonakeys has quit IRC
23:10:07 *** jonakeys has joined #gnucash