00:07:33 anyone knows what the L column ( first one ) in gstat represents? 00:08:57 last1, length of queue? 00:12:48 queue of what ? 00:13:17 whenever I do a zpool create on SSD, for example on a 7Tb drive, it takes like 10 minutes and gstat shows activity at 100% and that L(q) going from 850 to 0 00:13:22 during that 10 minutes 00:13:35 what is it doing exactly and can it be sped up ? 00:16:42 https://imgur.com/a/8VFgMVe 00:16:43 Title: Imgur: The magic of the Internet 00:28:32 last1, queue of IO requests/commands to geom 00:29:03 > zpool create on SSD, for example on a 7Tb drive, it takes like 10 minutes 00:29:31 so it's normal ? 00:29:53 I did zpool create on HDD 6TB and on 6x6TB ZRAID2 - can't remember even 1 minute wait time 00:31:46 check hardware, look at /var/log/messages 00:32:04 for me it isn't look normal 00:32:51 try "diskinfo -ctvi /dev/da3" 00:33:06 and check "smartctl -a /dev/da3" 00:33:17 it's been doing this forever, on lots of servers 00:33:22 this is all brand new 00:33:34 intel ssd s4520 7.68tb ssd 00:33:53 likely trimming the ssd? 00:35:11 yuripv: 21:00 ah, yuri left before i caould write that i don't have any. i've left a tip session to that vm to capture the eventual output. 00:35:41 meena: yep, thanks, already saw it on the list 00:44:15 I don't think it's trimming 00:44:26 it only happens on single drives 00:44:42 if I create a mirror then it's fast 00:58:00 last1: you could actually check, dtrace -n 'fbt::da_delete_unmap:entry' -c 'zpool create .......' 00:58:18 there are also related sysctls, check if they are growin 00:58:22 +g 01:24:45 something to do with zfs for sure 01:24:52 while it's doing that, I can't even do zpool list 01:24:56 it just hangs 01:35:19 hello Everyone ---- I noticed that FreeBSD 14 was pushed to October 2023 - I wonder if anyone has insight why ? I read few lists on current - and some feedback here - people were talking about OpenSSl3 needed to be used in FreeBSD 14 - it seem that OpenSSL 3.0 was released in 2021 we are in 2023 - so I am sure I am missing something here, I know the FreeBSD leadears have done countless releases before, so I doubt that this 01:35:19 overlooked... it would be good if someone with real knowledge would place some more pertinent announcement in FreeBSD.org website :) 02:03:44 It's ready when it's ready ? 02:27:50 a lot of software doesn't support openssl 3 yet 02:27:58 most of software… 02:29:08 so it's 90% issue of upstreams 07:06:13 * dch needs the equivalent of tcpdump but for unix domain sockets 07:21:03 otis: have you had the pleasure of any elasticsearch -> opensearch migrations? 07:21:14 in the end I gave up, and dropped the DB 07:21:54 I love graylog, but the effort needed to maintain mongod & ES/OS underneath simply doesn't work out for me 07:24:33 https://openobserve.ai/ despite the world's worst name, is a single rust binary 07:24:34 Title: OpenObserve | Open Source Observability Platform for Logs, Metrics, Traces, and More – Your Ultimate Dashboard for Alerts and Insights 07:36:12 acu: despite OpenSSL 3.0 being released a lot of software wasn't made compatible with it. what forced the project's hand was not the release of OpenSSL 3.0, but the EOL announcement of OpenSSL 1.1 set for this November. integrating OpenSSL 3.0 and clang 15/16 was a lot of work 07:47:57 acu: here's the full (public facing) discussion: https://lists.freebsd.org/archives/freebsd-arch/2023-April/000353.html 07:47:58 Title: OpenSSL in the FreeBSD base system / FreeBSD 14 07:52:47 dch: apparently you can intercept it with socat, and either dump it all to std(out|err) or to a port and then use tcpdump 07:53:29 dch: no, not yet, i'm with elastic because of $reasons :-/ 07:56:37 meena: ok that is the sort of hack I was hoping to avoid I might beg my dtrace buddies for some tips 07:57:04 otis: its the 3rd time I dumped graylog history because of ES/OS garbage 07:57:32 so frustrating. but in practice, at least for me, I don't *need* the logs. I have an offline syslog history for worst case. 07:57:39 I just want fancier grep 07:57:43 for the last ~ week approcx 08:07:05 otis: well I would avoid the upgrade as long as possible, its unbelievably flakey and error-prone 08:33:56 Fuck, hps@ is no longer with us :( 08:35:04 its an absolute tragedy :-( 08:38:31 Yes, yes it is. He was a delight to talk to at various virtual conferences we both attended, and we even found a shared interest in music that led us to sing+play together occationally (in the down-time before the conferences, usually). 08:39:01 He was also incredibly helpful, always ready to give advice in his many areas of expertise; He helped me debug USB issues several times. 08:51:18 meena: eol is Sept 23 not nov 08:53:07 I had two uncles pass away while in quarantine during covid while awaiting to be deployed to Guam. Couldn't even get out to smoke a cigar. 08:57:53 cpet: read 2023-09-11 as American. tho the American that wrote the email lives in Canada 08:58:51 I go by the actual openssl.org eol and that says Sept 09:01:06 Not sure why I'm being called American but ok I can do it in navy format if you want 09:01:46 emaste wrote the email 09:05:08 Ok 10:31:25 Trying FreeBSD after not using it since 2003. Many things have changed, yet I still like the quirkiness of getting X to run, albeit I'm running it in qemu. 10:42:13 pkg seems kinda stupid: "This port is deprecated; you may wish to reconsider installing it" shows up after a pkg upgrade, but there is *no way* to know which port/package it's referring to. About 250 packages were upgraded, and there are a *lot* of pkg messages, neither of which can be associated with a package. 11:08:56 does anyone know about the state of ath11k? 11:09:27 i'm considering getting compex wle3000hx for my home router and ath11k is currently in main, but not connected to build 11:48:46 Wifi didn't get faster than 54g so good luck with the fancy dancy 11:49:01 Ltning: I'm sure there's more to it than that 14:43:45 Do we have nfs team? Or nfs is part of the network team? 14:45:08 https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=261411 https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=147881 14:45:10 Title: 261411 – feature request: extend exports(5) format to allow specifying multiple host/permission tuples in one line 14:51:42 VVD: somebody recently did a lot of work on it, but I don't remember their name. lemme check 14:53:34 VVD: Rick Macklem: https://freshbsd.org/freebsd/src/branch/main?q=nfs 14:53:37 Title: FreeBSD / src - FreshBSD 14:53:59 yeah, rick's the nfs maintainer 14:55:00 ^^^ quote a grey beard get a grey beard 14:55:10 ;-) 14:55:58 I keep forgetting to check MAINTAINERS 14:57:29 ah, yes, he is in there 14:57:47 heh, "rmacklem is best for reviews" is probably an understatement 14:57:50 he already added himself to CC for both PRs 14:59:06 thanks 17:43:42 VVD: Rick has been working on NFS since the 80s, it's pretty fair to say he's THE expert. 17:46:13 now i understand what kevans meant with understatement 18:23:53 hello dear friend i try very much run the inspircd in freebsd version 13.2 and i get error https://bsd.to/xHn0 if someone can please what i need to do for install that i will happy 18:23:54 Title: dpaste/xHn0 (Plain Text) 18:26:39 Looks to me like the source file is missing a "#include " header and therefore missing the declaration of the isascii() macro. 18:27:07 maybe this a install? 18:27:16 or something i run it from home not from ports 18:27:59 after configure i do it gmake CXXFLAGS=-I/usr/local/include LDFLAGS=-L/usr/local/lib install 18:28:58 I assume you are porting what looks like inspircd-4.0.0a21 to your system? Where would I look to find that source? 18:29:43 Note that the error is an undeclared isascii() and not that it could not open a file and not that it could not find a file. This looks like a bug in the source using isascii() but not including the in it. 18:30:33 If you are working on ports then perhaps ask about the port in #freebsd-ports where experts in porting hang out and might have answers. I know nothing. 18:30:38 from where i find it https://github.com/inspircd/inspircd 18:30:39 Title: GitHub - inspircd/inspircd: A modular C++ IRC server (ircd). 18:31:13 but all thing i run from home not from ports and everthing is ok but only it is make problem 18:32:10 i think i need something for install for this error? usr/include/c++/v1/__locale:676:16: fatal error: use of undeclared identifier ' isascii' 18:32:22 debdrup, nice! 18:32:26 I see inspircd-3.15.0 already available as a binary pkg. Someone else has already ported it to FreeBSD. It would be easier to install the binary pkg and use it. Maybe? 18:33:22 That "fatal error: use of undeclared identifier" looks to me like a bug in the upstream source code. You will need to modify the source in order to fix the bug. That is a typical task for porters when porting source code. 18:33:44 I have bw issue on 10GE - can't get more than 150-160MBytes/s even from ramdisk. Clinet and server both are FreeBSD. 18:34:08 The source probably compiles on the author's system because some header file probably includes there but it does not get included when compiling on FreeBSD because it is a different OS. 18:34:56 then where is problem in here? 18:35:35 The problem here is that the upstream source is buggy. Getting buggy code to compile is a task for anyone who is porting code from one OS to another OS. Very typical. 18:36:22 then what i need to do is possible run it from other commands? 18:36:39 meena: thanks a bunch about the link - OpenSSL 3 and the delay reasons - https://lists.freebsd.org/archives/freebsd-arch/2023-April/000353.html --- it would be useful to have an article on the freebsd.org about it...so people do not need to go and read the lists (well - the links the freebsd.org to lists requires you to subscribe with your local mail app - ) 18:36:39 Title: OpenSSL in the FreeBSD base system / FreeBSD 14 18:39:48 this a after i do update to 13.2 18:41:55 I suggest you look at what was needed for the port of the previous version inspircd-3.15.0 as the same things will likely be needed for the new version too. 18:42:16 I suggest asking about this in #freebsd-ports as I know very little about how the ports system works but the experts there will know. 18:47:58 acu: delays happen for various reasons with almost every release, and while I agree the a writeup would be invaluable, one could argue that people were too busy getting a release out to write down a retrospective 18:50:23 no body talk there 18:50:32 i need please help about my problem please 18:51:17 acu: be the change that you want to see. 19:17:08 ASHER: Did you try 'pkg install inspired'. 19:20:36 Under OpenBSD's PF, I use; pass quick inet proto icmp icmp-type $icmp_types max-pkt-rate 100/10 however, when it comes to FreeBSD&PF, it's just a "syntax error", is there any equivalent of max-pkt-rate for FreeBSD's PF? 19:23:06 yes i try it 19:44:10 tercaL, there's a max-src-conn-rate but I don't know whether it does what you want - or works with ICMP packets. 20:02:25 I tried a lot to change size of cursor but all in vein. I use freebsd 13.2 and awesomewm. X would work same but let me know. https://wiki.archlinux.org/title/Cursor_themes#X_resources 20:02:26 Title: Cursor themes - ArchWiki 20:27:56 hi all, n00b question: let's say i start with a fully pkg-based system, but need to install a few ports with custom build options, and i want to use binary pkgs for as many build deps as possible 20:27:57 so far i fetched ports with plain git, and basically used make install-missing-packages && make install. this somewhat worked but i assume that pkg upgrade will eventually replace my customized ports with regular packages? also pkg autoremove wants to uninstall the build deps. also the binary deps are not always enough so i'd like to select some of them (but not all) to be also built from ports. 20:27:59 what would be the easiest / most frictionless solution to maintain such a mostly-binary system with a few ports? the handbook/google give different options, but i'm still not clear which one will be best 20:33:48 luke8086: yes, if you upgrade with pkg you will lose those options 20:34:31 luke8086: generally, the advice is not build on the same system as you install. a good way to separate these concerns is to use poudriere 20:35:17 or just build everything from ports; not sure what the drive is to use prebuilt ones 20:35:40 poudriere also helps with some of that upgrade pain, in that you can give your repo higher prio 20:36:43 Simplicity. Some of us are bad sysadmins and have no interest or ability in becoming better ones, but still use FreeBSD for various reasons. 20:36:53 rtprio: i'm using an ancient laptop, compiling everything would took ages 20:37:19 luke8086: good reason 20:37:34 V_PauAmma_V: pkg upgrade -a is as simple as portupgade -a 20:39:08 The command line may be. The process and the pain it causes? Not at all sure they're even in the same league. But I've had that discussion before. 20:39:10 meena: i saw different people recommending poudriere and synth. are there solid reasons to use one over the other, or will they both achieve the same thing? 20:39:42 i don't know synth. porters use poudriere 20:40:20 *most porters (i interact with) 20:42:09 meena: gotcha, being more popular among devs is a solid reason to me :) 21:11:43 hello here 22:05:54 is there any alternative for pulseeffects ? i really miss it.. music sounds on a different level with it 22:11:42  this does not works `xsetroot -xcf /usr/local/share/icons/redglass/cursors/arrow 64` it only loads cursor on window edges 22:16:18 ok I got it working 22:16:50 workedon new widnows/apps 22:26:47 hernan: I never heard of it. Do you mean Easy Effects? (I never heard of that, either, but I see .) 22:26:48 Title: GitHub - wwmm/easyeffects: Limiter, compressor, convolver, equalizer and auto volume and many other plugins for PipeWire applications 22:28:17 I guess that no-one can buildworld for FreeBSD-CURRENT at the moment. True? 22:28:47 hernan, i think need use jackd by default 23:08:53 grahamperrin: easyeffects is supposed to be the "evolved" version.. however its linux only. however it uses some linux sound backend that is incompatible with freebsd. 23:11:36 grahamperrin: pulseeffects was is able to "re-decompress" a compressed audio.. it was kind of able of transforming a 96kbps sound into something better than a 320kbps 23:12:27 any youtube audio sounded a lot better with it 23:14:55 you can't restore the bits that aren't there :-) 23:22:45 RhodiumToad, +1 23:23:40 jack with 48khz and 32-bits 23:24:03 good night