00:21:41 where can i suggest a package request? 00:33:54 bugs.freebsd.org 06:51:59 Has anyone used pkgbasify to move from 14.3 to 15.0? I moved to 14.4 but wanted to see about moving to 15.0. 07:00:08 well, https://wiki.freebsd.org/WantedPorts is more appropriate if it is just a request without any work to create the port 07:11:15 I maintain an unofficial code-server port fwiw, in case anyone wants it - just updated to v4.111.0 https://github.com/0x1eef/myports/tree/main/freebsd/www/code-server 09:10:37 is there a latex package in packages? doing a pkg search latex doesn't seem to show one 09:14:43 pkg which tells me that the latex executable was installed by tex-formats I just installed texlive-full as that is less hassle then thinking about what parts of texlive I could need 11:05:31 anyone has any good tips on getting ipv6 working on a freebsd guest (proxmox) on "#"(/")#=( OVH ? :P 11:07:24 normal ifconfig__ipv6="inet6 accept_rtadv" doesn't work? 11:10:29 wait, they give you one ipv6 address that changes sometimes instead of SLAAC? 11:10:47 the OVH routing is a mess.. 11:11:14 basically the gateway is outside of the allocated /64 11:12:54 nimaje: https://pastebin.com/C7jZkMHb 11:13:34 works from time to time, but mostly NOT :/ 11:20:36 no idea, looks correct to me if the addresses and prefix len is correct, ask them why they don't do SLAAC 11:37:32 ek / ivy : look what I found https://cgit.freebsd.org/ports/tree/security/openvpn/Makefile#n84 11:38:15 .ifdef (LOG_OPENVPN) <== by default, in the OpenVPN code, this is set to LOG_DAEMON 11:39:13 Now, for poudriere, I could just define LOG_OPENVPN=LOG_LOCAL6 (for example, as shown on line 132 (same file). 11:41:15 I had this thought earlier today: why not modify openvpn to do this...? 11:44:28 hm, but why is that not configurable via the config and only via compile time options? I expect most users to install a package via their package manager 11:52:36 https://services.unixathome.org/poudriere/data/150amd64-default-primary/2026-03-15_11h46m11s/logs/openvpn-2.6.19.log 11:53:16 Success. OpenVPN is now logging via facility local6 - no more duplicate log enties. 11:54:09 nimaje: Yes, the way it is configured, it must be set at built time. Perhaps it is a security consderation. 12:19:51 https://dan.langille.org/2026/03/15/hacking-openvpn-to-use-syslog-with-something-other-than-facility-daemon/ 12:37:50 That logging knob has been there since at least 2009: https://cgit.freebsd.org/ports/commit/security/openvpn/Makefile?id=b222799b625902dec465377894fb3265c7ec0a49 12:42:02 dvl: Very nice find! I am surprised it's a compile-time option. But, perhaps you're right in the security consideration (although, I'm hard-pressed to think of a good reason behind it.) 12:42:27 ek: Don't allow an intruder to change the logging. 12:42:41 I suppose so. 12:43:13 I also don't see why this couldn't easily be a port config option either. Surprised that's not a thing, actually. Super easy patch. 12:49:13 which facilities would you then expose as an option? at least LOG_SECURITY LOG_DAEMON and LOG_LOCAL<0-7> 12:51:57 nimaje: Good question. I suppose that could be entirely up for debate. Or, as dvl did, just specify whatever you'd like in make.conf. 12:53:41 dvl: Did you end up finding a reasonable solution for the log rotation? Something that doesn't require disconnects each time? 13:30:01 nimaje: I uses LOG_LOCAL6 (see bog post above). 13:30:21 ek: Yes, specified it in a poudriere make.conf file (see blog post). 13:31:15 ek: Yes, the log rotation... is not in the blog post. 13:32:57 ek: Log rotation added to that blog post. 13:33:24 FYI, I emailed the author of the patch to thank them. So useful and just what was needed. 13:33:28 cccccbkvnbghuflftvutkcnkcnjiudegfhfkkcltgneg 13:34:13 dvl: the question was: if exposing LOG_OPENVPN as an option instead as ek suggested, which facilities should be exposed there then? and mostly to show that those 9 you obviously would want to expose are likely the reason it was made into a variable instead of a big list of mutally exclusive options 14:09:22 ek: I follow, I would allow any facility usually valid. Let the user decide at run time. 14:09:47 ek: Sorry, I follow now. I rushed through your original question; I was about to go out, and now I'm back. 17:04:41 dvl: Very nice! Blog post is most certainly helpful (as always). Thank you! 17:23:17 <_shard> :b 5 17:38:24 ialokin, If the router gateway is not in the subnet then you must add a host route to the network device for it. That's a typical OVH configuration. Something like ifconfig_vtnet0_ipv6="inet6 2606:4700::6812:46a/128" ipv6_static_routes="lan" ipv6_route_lan="-host 2604:2dc0:202:300::1 -interface vtnet0" ipv6_defaultrouter="2604:2dc0:202:300::1" 20:07:34 those who use poudriere, can you build multiple variants of the same package? how are they identified to be installed after? 20:24:50 you mean with diffrent options? not in the same repo 20:28:32 that sounds like it would get complicted rathre quickly 20:52:48 with sets you can easily have multiple repos, but what is the same between them would still be build multiple times 20:56:24 Ahem, hello guys 21:00:48 join #freebsd-pulse 21:03:06 ek: thank you. 21:03:25 ek: Next deploy that logging solution to each client. 21:04:47 jjil 21:05:31 First time IRC user here, this is so cool yet so disorienting 21:17:10 rtprio: I use sets as well, for example I have desktop and servers as individual sets. 23:08:05 what i'm finding a drag is the icinga2 requires postgres, but that hoses the other places i use postgres 23:08:37 except for the one system that needs icinga2 with postgres. 23:15:18 rtprio: Yeah. Same. :( 23:16:29 Although, I have been looking for an icinga2 replacement recently. It's pretty crappy how little FBSD is supported by full-fledged monitoring. I also use nagios but I wish it had some sort of historical data. Graphs or something. 23:18:21 I do kinda like Zabbix but it is one HUNGRY beast. 23:18:33 And, again, FBSD support isn't all that fantastic. 23:31:20 i find icinga2 works well enough; adding a few of my own plugins is rather effortless 23:32:04 the 'default' nagios checks are.. yeah, don't support freebsd very well 23:32:09 I never warmed up to Ichinga2 because it is so Javascript heavy. I prefer (love and hate) Nagios over Ichinga2. But all of the cool kids these days are using Prometheus and Grafina which I find to be quite a different thing. 23:32:48 i like that i can manage it with puppet 23:33:01 Can't you manage all of them with puppet? 23:33:23 there was something that turned me off to zabbix 23:34:01 i don't know, we used icinga at an old job and i liked it well ehough 23:35:04 rwp: I like Prometheus with Grafana, but also VERY resource hungry (in my experience). Especially if the DB isn't set to dump data. More than 2 weeks of historical data would take like 2 minutes to load via Grafana. Drove me nuts. 23:35:53 Not to mention the lack of FBSD templates available. Some work, some don't, but there's only like 5 (3 or which are for FreeNAS/TrueNAS which also only partially work.) 23:36:21 I've tried Munin, no dice there. Monit kind of works, but not what I'm looking for. 23:36:22 iirc i found grafana really frustrating to add graphs / do useful inquriries 23:36:30 I need to check out Netdata, apparently. 23:36:43 netdata seemed to consume quite a few resources 23:36:46 rtprio: Yes. For sure. It was a ton of work for very little return. 23:37:01 Yeah. It's next on my list to poke around on. 23:38:00 ek, I use Monit everywhere I need to monitor and take action upon conditions. It could be better but it is not terrible. 23:38:03 but thanks to dvl i have a warning in icinga because of pkg audit 23:38:18 The old pisg/php4nagios/nagiosgraph thing used to work pretty great but it hasn't been maintained in so long PHP8 breaks it. 23:38:24 I use Munin for resource trends but if you have a lot of systems then cactus is better. 23:39:12 oh, i remember cacti... i liked it at the time but i do not enjoy keeping php scripts running anymore 23:39:59 Yep. Cacti was alright. Simple little RRD graphs. Worked okay. 23:40:00 It's just the scaling thing. Too many graphs all on one machine makes Munin bog down. 23:40:14 rwp: That's what I was seeing. 23:40:40 Same with Prometheus/Grafana as well as Zabbix. 23:40:47 Newer Munin installations use an on-demand graph creation to avoid generating graphs all of the time that are not being viewed. It's a good direction to go. 23:41:04 That would be nice. 23:41:20 rtprio: there was a discussion about muting certain pkg-audit alerts. I haven't implemented it yet. There was talk on bsd.network about it. 23:41:29 I have not used Zabbix and know nothing about it. 23:41:33 I think it had more to do with the amount of data that being graphed. Gotta find a good mid point. 23:42:17 If I kept 3+ months of data, it would take tons of time to load stuff. 23:43:16 yes, i'm observing it would be nice to only keep the alert active if there's a package to upgrade to