Switch to DuckDuckGo Search
   April 15, 2019
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30

Toggle Join/Part | bottom
[00:28:29] *** clapont <clapont!~clapont@unaffiliated/clapont> has quit IRC (Ping timeout: 250 seconds)
[00:30:27] *** merzo <merzo!~merzo@87-49-132-95.pool.ukrtel.net> has quit IRC (Ping timeout: 240 seconds)
[00:43:16] <nbjoerg> grmbl. stupid in.ndpd
[00:51:04] *** btibble <btibble!~brantibbl@c-69-94-200-89.hs.gigamonster.net> has quit IRC (Ping timeout: 264 seconds)
[00:55:24] *** clapont <clapont!~clapont@unaffiliated/clapont> has joined #illumos
[00:57:52] *** merzo <merzo!~merzo@87-49-132-95.pool.ukrtel.net> has joined #illumos
[01:09:11] *** snuff-work <snuff-work!~snuff-wor@202-161-112-134.tpgi.com.au> has joined #illumos
[01:10:15] *** andy_js <andy_js!~andy@94.6.62.238> has quit IRC (Quit: andy_js)
[01:33:56] *** lewellyn <lewellyn!~lewellyn@about/windows/lewellyn> has joined #illumos
[01:33:56] *** ChanServ sets mode: +o lewellyn
[02:03:06] *** Gaston|DSP <Gaston|DSP!~quassel@80.252.185.191> has joined #illumos
[02:03:52] <richlowe> nbjoerg: complain at danmcd when he's around, probably?
[02:05:17] <nbjoerg> will try :)
[03:10:59] *** merzo <merzo!~merzo@87-49-132-95.pool.ukrtel.net> has quit IRC (Ping timeout: 250 seconds)
[03:20:00] *** ed209 <ed209!~ed209@165.225.128.67> has quit IRC (Remote host closed the connection)
[03:20:07] *** ed209 <ed209!~ed209@165.225.128.67> has joined #illumos
[04:30:01] *** rsully <rsully!~rsully@unaffiliated/rsully> has quit IRC (Quit: rsully)
[04:44:22] *** Kurlon_ <Kurlon_!~Kurlon@98.13.72.207> has quit IRC (Remote host closed the connection)
[04:45:02] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has joined #illumos
[04:50:16] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has quit IRC (Ping timeout: 252 seconds)
[05:55:19] *** baojg <baojg!~baojg@162.243.44.213> has joined #illumos
[06:07:52] *** freakazoid0223 <freakazoid0223!~matt@pool-108-52-159-210.phlapa.fios.verizon.net> has quit IRC (Remote host closed the connection)
[06:11:13] *** richlowe <richlowe!~richlowe@cpe-74-139-197-163.kya.res.rr.com> has quit IRC (Ping timeout: 246 seconds)
[06:23:10] *** richlowe <richlowe!~richlowe@2605:a000:160c:8b5b::bb7> has joined #illumos
[06:23:46] *** hemi770_ <hemi770_!~hemi666@unaffiliated/hemi770> has joined #illumos
[06:25:59] *** hemi770 <hemi770!~hemi666@unaffiliated/hemi770> has quit IRC (Ping timeout: 250 seconds)
[07:19:00] *** jimklimov <jimklimov!~jimklimov@ip-86-49-245-88.net.upcbroadband.cz> has quit IRC (Ping timeout: 246 seconds)
[08:07:04] <gitomat> [illumos-gate] 10598 loader: implement map-vdisk and unmap-vdisk commands -- Toomas Soome <tsoome at me dot com>
[08:22:53] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Quit: tsoome)
[08:32:04] *** hemi770_ is now known as hemi770
[08:46:53] <LeftWing> So, on account of the wiki being out of service for the time being, we've pressed forward to get the new docs site up and running:
[08:46:55] <LeftWing> https://illumos.org/docs/
[08:52:13] <bdha> LeftWing++
[08:58:34] *** merzo <merzo!~merzo@195.20.130.10> has joined #illumos
[09:03:12] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[09:08:18] <LeftWing> bdha++ too!
[09:19:30] *** neuroserve <neuroserve!~toens@195.71.113.124> has joined #illumos
[09:20:18] *** fanta7531 <fanta7531!~fanta7531@p200300F76BC17C0018FB5ADD849A45CA.dip0.t-ipconnect.de> has joined #illumos
[09:29:07] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #illumos
[09:33:37] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[09:33:43] *** jimklimov1 <jimklimov1!~jimklimov@31.7.243.238> has joined #illumos
[09:36:23] <bdha> Tornado warning. :|
[09:40:09] <drscream> fantastic :-) (not the tornado warning)
[09:54:36] *** lblume <lblume!~lblume@greenviolet/laoyijiehe/lblume> has joined #illumos
[10:00:40] *** _joes_ <_joes_!~Thunderbi@82-132-246-81.dab.02.net> has joined #illumos
[10:09:28] *** _joes_1 <_joes_1!~Thunderbi@82-132-246-173.dab.02.net> has joined #illumos
[10:11:19] *** _joes_ <_joes_!~Thunderbi@82-132-246-81.dab.02.net> has quit IRC (Ping timeout: 264 seconds)
[10:15:05] <jimklimov1> gone with the wind?
[10:16:12] *** _joes_1 <_joes_1!~Thunderbi@82-132-246-173.dab.02.net> has quit IRC (Quit: _joes_1)
[10:19:55] <Smithx10> Pass me over some wunix
[10:19:55] <Smithx10> https://illumos.org/docs/about/history/
[10:35:28] *** mnowak_ <mnowak_!mnowak_@nat/suse/x-ecwcgneryeddyapi> has quit IRC (Ping timeout: 246 seconds)
[10:38:58] *** man_u <man_u!~manu@manu2.gandi.net> has joined #illumos
[10:44:21] *** mnowak_ <mnowak_!mnowak_@nat/suse/x-euatyiagjuwgrygh> has joined #illumos
[10:59:06] *** mavhc <mavhc!~mavhc@cpc77319-basf12-2-0-cust433.12-3.cable.virginm.net> has quit IRC (Ping timeout: 252 seconds)
[11:01:24] *** _joes_ <_joes_!~Thunderbi@82-132-246-173.dab.02.net> has joined #illumos
[11:01:59] *** mavhc <mavhc!~mavhc@cpc77319-basf12-2-0-cust433.12-3.cable.virginm.net> has joined #illumos
[11:08:44] <tsoome> great s11.4 panic in production system… grr.
[11:13:27] *** jimklimov1 <jimklimov1!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[11:13:46] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #illumos
[11:15:59] *** denk <denk!~denis@devel.tambov.ru> has quit IRC (Remote host closed the connection)
[11:35:17] *** denk <denk!~denis@devel.tambov.ru> has joined #illumos
[11:37:33] *** _joes_ <_joes_!~Thunderbi@82-132-246-173.dab.02.net> has quit IRC (Ping timeout: 245 seconds)
[11:45:12] *** andy_js <andy_js!~andy@94.6.62.238> has joined #illumos
[11:56:07] *** psarria <psarria!~psarria@142.red-79-146-98.dynamicip.rima-tde.net> has joined #illumos
[12:00:33] *** baojg <baojg!~baojg@162.243.44.213> has quit IRC ()
[12:04:39] *** _joes_ <_joes_!~Thunderbi@82-132-246-173.dab.02.net> has joined #illumos
[12:04:43] *** BH23 <BH23!~BH23@82-132-246-173.dab.02.net> has joined #illumos
[12:16:56] *** baojg <baojg!~baojg@162.243.44.213> has joined #illumos
[12:20:00] *** ed209 <ed209!~ed209@165.225.128.67> has quit IRC (Remote host closed the connection)
[12:20:07] *** ed209 <ed209!~ed209@165.225.128.67> has joined #illumos
[12:20:45] *** gitomat <gitomat!~nodebot@165.225.148.18> has quit IRC (Remote host closed the connection)
[12:20:55] *** gitomat <gitomat!~nodebot@165.225.148.18> has joined #illumos
[12:20:56] *** _joes_1 <_joes_1!~Thunderbi@81.134.255.170> has joined #illumos
[12:22:11] *** _joes_ <_joes_!~Thunderbi@82-132-246-173.dab.02.net> has quit IRC (Ping timeout: 250 seconds)
[12:22:27] *** BH23 <BH23!~BH23@82-132-246-173.dab.02.net> has quit IRC (Ping timeout: 246 seconds)
[12:32:57] *** vila <vila!~vila@lstlambert-658-1-248-94.w80-15.abo.wanadoo.fr> has quit IRC (Ping timeout: 245 seconds)
[12:35:53] *** vila <vila!~vila@lstlambert-658-1-248-94.w80-15.abo.wanadoo.fr> has joined #illumos
[12:35:54] *** fanta7531 <fanta7531!~fanta7531@p200300F76BC17C0018FB5ADD849A45CA.dip0.t-ipconnect.de> has quit IRC (Quit: fanta7531)
[12:45:48] *** _joes_1 <_joes_1!~Thunderbi@81.134.255.170> has quit IRC (Ping timeout: 252 seconds)
[12:52:16] <tomww> im-... possible!
[12:56:00] <igork> tsoome: what is panic dump?
[12:56:23] <igork> is it on sparc or intel?
[13:52:39] *** rsmarples <rsmarples!~roy@cpc115040-bour7-2-0-cust370.15-1.cable.virginm.net> has joined #illumos
[14:13:43] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has joined #illumos
[14:14:38] *** merzo <merzo!~merzo@195.20.130.10> has quit IRC (Ping timeout: 245 seconds)
[14:16:43] *** freakazoid0223 <freakazoid0223!~matt@pool-108-52-159-210.phlapa.fios.verizon.net> has joined #illumos
[14:26:35] <rsmarples> Hi!
[14:28:25] <rsmarples> I am getting garbage RTM_NEWADDR (and friends) messages - every sockaddr past IFP seem to be corrupted
[14:28:29] <rsmarples> https://paste.dilos.org/?64c11e25ace173a2#b7quzFdj8y9qsH/3eXxIKddhkxbPiHDxDO+Nwj3UJsk=
[14:28:46] <rsmarples> this makes the messages useless as they make no sense without a valid IFA
[14:53:10] *** merzo <merzo!~merzo@195.20.130.10> has joined #illumos
[15:08:54] *** gh34 <gh34!~textual@rrcs-162-155-144-114.central.biz.rr.com> has joined #illumos
[15:16:34] *** rsully <rsully!~rsully@unaffiliated/rsully> has joined #illumos
[15:20:34] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[15:32:58] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[15:49:39] *** danmcd <danmcd!~danmcd@static-71-174-113-16.bstnma.fios.verizon.net> has joined #illumos
[16:04:40] <andy_js> Why is lofiadm so slow to attach/detach?
[16:06:44] <nbjoerg> danmcd: could you please fix in.ndpd so that it deals with later-created interfaces properly?
[16:06:50] <nbjoerg> danmcd: that's highly annoying to debug :(
[16:07:27] <nbjoerg> danmcd: e.g. configure a prefix in /etc/inet/ndpd.conf references a vnic that doesn't exist yet, create it, never observe the RAs
[16:09:56] <jlevon> andy_js: it's probably doing a reconfig and some set of devices are slow in doing so
[16:19:44] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Quit: tsoome)
[16:24:06] <andy_js> It seems to be consistently slow.
[16:24:29] <andy_js> And when you say devices, are we talking physical devices?
[16:28:36] *** khng300 <khng300!~khng300@unaffiliated/khng300> has quit IRC (Quit: ZNC 1.7.1 - https://znc.in)
[16:30:01] *** khng300 <khng300!~khng300@unaffiliated/khng300> has joined #illumos
[16:34:45] *** khng300 <khng300!~khng300@unaffiliated/khng300> has quit IRC (Client Quit)
[16:36:19] *** khng300 <khng300!~khng300@unaffiliated/khng300> has joined #illumos
[16:38:33] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Ping timeout: 268 seconds)
[16:43:17] <toasterson> alanc (IRC):
[16:44:13] <toasterson> fail... alanc (IRC) are you using the OCI solaris Specifications somewhere? I was looking to use the specs for my own Project.
[16:47:27] <toasterson> oh I just found https://github.com/oracle/railcar
[16:50:10] <andyf> andy_js - there was a bug fixed in the past few months about lofi that fixed a massive delay in it.
[16:50:19] <andyf> andy_js - how up-to-date are you?
[16:52:46] *** serapheim <serapheim!~serapheim@DELPHIX-COR.ear2.SanJose1.Level3.net> has joined #illumos
[17:09:21] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[17:16:26] *** neuroserve <neuroserve!~toens@195.71.113.124> has quit IRC (Ping timeout: 252 seconds)
[17:34:54] <andy_js> andyf: My build is pretty old I think.
[17:35:14] <andy_js> I’m running OmniOSce r151026.
[17:37:02] *** sarvet <sarvet!~sarvet@dslb-092-073-090-050.092.073.pools.vodafone-ip.de> has joined #illumos
[17:43:45] <andyf> The answer, I think, is that it flushes it all out to swap even if it's in tmpfs
[17:43:57] <andyf> but I can't find the commit for the fix just now
[17:44:09] <andyf> I'm pretty sure it came after r26 though
[17:44:30] <rmustacc> https://www.illumos.org/issues/10262
[17:44:43] <andyf> rmustacc Thanks :)
[17:46:37] <andyf> andy_js - ok, not even in r151028, but it will be in the upcoming r151030
[17:48:05] *** sarvet <sarvet!~sarvet@dslb-092-073-090-050.092.073.pools.vodafone-ip.de> has quit IRC (Remote host closed the connection)
[17:49:07] *** sarvet <sarvet!~sarvet@dslb-092-073-090-050.092.073.pools.vodafone-ip.de> has joined #illumos
[17:50:46] <andy_js> Nice. I guess this also explains the horrible tmpfs performance I’ve been seeing too.
[17:52:02] *** sarvet <sarvet!~sarvet@dslb-092-073-090-050.092.073.pools.vodafone-ip.de> has quit IRC (Remote host closed the connection)
[17:52:36] *** sarvet <sarvet!~sarvet@dslb-092-073-090-050.092.073.pools.vodafone-ip.de> has joined #illumos
[18:02:02] <andy_js> Oh it’s specific to lofi. I guess not then.
[18:02:54] *** polishdub <polishdub!~polishdub@207.86.38.254> has joined #illumos
[18:10:18] *** alanc <alanc!~alanc@129.157.69.37> has quit IRC (Remote host closed the connection)
[18:10:45] *** alanc <alanc!~alanc@129.157.69.37> has joined #illumos
[18:22:58] *** polishdub <polishdub!~polishdub@207.86.38.254> has quit IRC (Quit: leaving)
[18:38:46] *** polishdub <polishdub!~polishdub@207.86.38.254> has joined #illumos
[19:06:33] <alanc> toasterson: I don't think we've really used OCT for Solaris since the Docker port was killed
[19:06:44] <alanc> ^OCT^OCI
[19:07:02] <alanc> there were discussions of using it without docker, but they fizzled out
[19:07:56] <toasterson> ah well than I am free to do whatever. Thanks for the Specs anyway no need for me to map stuff around :)
[19:08:05] <toasterson> *then
[19:17:13] *** man_u <man_u!~manu@manu2.gandi.net> has quit IRC (Ping timeout: 252 seconds)
[19:17:56] <igork> cat someone paste : file /sbin/ifconfig
[19:18:03] <igork> from omnios and smartos
[19:25:15] *** sarvet <sarvet!~sarvet@dslb-092-073-090-050.092.073.pools.vodafone-ip.de> has quit IRC (Ping timeout: 246 seconds)
[19:35:29] <alanc> Solaris 11.4: /sbin/ifconfig: ELF 64-bit MSB dynamic lib SPARCV9 Version 1, position-independent executable, dynamically linked, not stripped, no debugging information available
[19:36:02] <igork> alanc: thanks
[19:36:11] <igork> SPARCV9 :)
[19:37:18] *** fanta7531 <fanta7531!~fanta7531@p200300F76BC17C0018FB5ADD849A45CA.dip0.t-ipconnect.de> has joined #illumos
[19:37:51] <alanc> yeah, well, that was from the window I was ssh'ed into jurassic in
[19:39:11] <alanc> x86 just has the expected differences: /sbin/ifconfig: ELF 64-bit LSB dynamic lib AMD64 Version 1 [SSE2 SSE], position-independent executable, dynamically linked, not stripped
[19:39:38] <igork> thanks
[19:49:48] <ptribble> /sbin/ifconfig: ELF 32-bit LSB executable 80386 Version 1, dynamically linked, not stripped, no debugging information available
[19:50:04] <ptribble> That's what both Tribblix and omnios have
[19:50:42] <igork> on dilos:
[19:50:42] <igork> root@p1c1s:~# file /sbin/ifconfig
[19:50:42] <igork> /sbin/ifconfig: ELF 64-bit LSB executable, x86-64, version 1 (Solaris), dynamically linked, interpreter /lib/amd64/ld.so.1, not stripped
[19:52:41] <richlowe> in general, we have not really bothered 64bitt-ing things without reasons
[19:53:03] <richlowe> and presumably the first candidates would be time time_t sensitive
[19:53:22] <igork> richlowe: try setup on your env date: 28 September 2135 year, 1pm
[19:53:48] <igork> and run mesg
[20:03:23] *** psarria <psarria!~psarria@142.red-79-146-98.dynamicip.rima-tde.net> has quit IRC (Quit: Leaving)
[20:15:43] <alanc> well, you've got 19 years left to either make things 64-bit or implement https://sourceware.org/glibc/wiki/Y2038ProofnessDesign
[20:16:14] <alanc> we had the added motivation of being able to use ADI on 64-bit SPARC stuff that you don't have
[20:23:37] <richlowe> alanc: skimming -- just like largefile, but time?
[20:23:48] <alanc> yep
[20:24:54] <richlowe> sounds perfect for rmustacc :)
[20:24:54] <lewellyn> pfft. in the next 20 years we might not even USE time anymore. we might become more enlightened beings transcending such outmoded notions! then 2038 will be a non-issue. :D
[20:25:21] <richlowe> lewellyn: totally the wrong way of thinking!
[20:25:34] * lewellyn sticks his head back in the sand
[20:25:37] <richlowe> lewellyn: if we don't fix it now, people will pay me a _lot_ in 2037 when they find that machine they forgot about and _absolutely_ need to get fixed
[20:25:50] <richlowe> it's not a bug, it's my pension plan.
[20:26:01] <lewellyn> richlowe: yes. i'm providing a reason to not fix it :D
[20:26:18] <lewellyn> even if it needs new-age-y music to work
[20:26:21] <alanc> lol - given I turn 65 in 2038, I'm expecting some retirement funds myself that way
[20:27:05] <lewellyn> heh. the cobol jobs i've talked to lately, i've been asking about things like 2038. sounds like most of the big cobol codebases are pretty futureproofed these days
[20:27:34] <lewellyn> one of the BIG names is looking at moving from microfocus to gnu cobol, which was an interesting project. but i don't think i'll get that one
[20:29:04] *** merzo <merzo!~merzo@195.20.130.10> has quit IRC (Ping timeout: 246 seconds)
[20:29:19] <lewellyn> the trendy kids can keep their rust and go. the money long term is in things which have proven to not be just the latest fad.
[20:29:37] * lewellyn gets back to this vala project >.>
[20:43:19] *** fanta7531 <fanta7531!~fanta7531@p200300F76BC17C0018FB5ADD849A45CA.dip0.t-ipconnect.de> has quit IRC (Quit: fanta7531)
[20:46:53] <rmustacc> richlowe: What am I doing now?
[21:17:34] *** khng300_ <khng300_!~khng300@unaffiliated/khng300> has joined #illumos
[21:17:39] *** khng300 <khng300!~khng300@unaffiliated/khng300> has quit IRC (Quit: ZNC 1.7.1 - https://znc.in)
[21:34:28] <Smithx10> lol LeftWing , it seem's Illumos and Solaris have a very difficult time being separate. But for some reason the BSDs don't have this same stigma. "But again, the bigger concern here is that if we do this split, then who runs the Oracle Solaris builders?" bradfitz just said on that github thread.
[21:34:46] <Smithx10> I'd imagine someone who owns the Oracle Solaris Port should. hahha
[21:35:54] <LeftWing> I mean, my answer is mostly.... Who cares?
[21:36:22] <Smithx10> Yea..... I mean I don't notice this type of interaction
[21:36:27] <LeftWing> Or at the very least, if Oracle doesn't care enough to staff it, or to make their OS open source, it doesn't feeling like we should be on the hook for that
[21:37:07] <Smithx10> With other OS's etc. Like I think BSD has 4 tags or something and I dont think anyone ever said..... But what about the <somebsd>BSD?
[21:37:18] <Smithx10> ahahha
[21:37:25] <LeftWing> I think you asked the right question in the ticket, anyway.
[21:37:41] <Smithx10> Know what I mean?
[21:37:41] <rmustacc> The split is also more obvious from an OS perspective there and has been more true for a longer time for most of the folks concerned.
[21:38:00] <LeftWing> On some level it's on us to market ourselves better as being totally separate
[21:38:34] <Smithx10> LeftWing: The other day during a call someone called it the S word..... and I was like.... OK EVERYONE. STOP. ITS NOT SOLARIS.
[21:39:03] <LeftWing> It's tough, because we are at least somewhat *source* compatible, at least for classic UNIX interfaces, and that has obviously been a huge help with software porting
[21:39:37] <sjorge> Nice work on the new wiki!
[21:40:12] <Smithx10> Yeah, I wonder if you would have made it incompatible if it would have made the split easier?
[21:40:22] <Smithx10> just some random bit tossed in there hahhaha
[21:40:48] <lewellyn> like a different uname? :P
[21:40:55] <Smithx10> SICK BURN
[21:40:58] *** gh34 <gh34!~textual@rrcs-162-155-144-114.central.biz.rr.com> has quit IRC (Quit: Textual IRC Client: www.textualapp.com)
[21:41:20] <lewellyn> eh. it's been a topic since day 1, on and off.
[21:41:30] <Smithx10> lewellyn hahaha, mistakes were made? :P
[21:41:37] <lewellyn> that depends upon your point of view
[21:41:40] <rmustacc> It's all about trade offs.
[21:41:40] <Smithx10> hahah pros and cons, PROS ! and CONS!
[21:42:07] <igork> about illumos wiki - https://illumos.org/docs/about/distro/ - i'm interested in how omnios ce comparable with Debian?
[21:44:11] <Smithx10> Igork are you down with Wunix?
[21:44:12] <Smithx10> https://illumos.org/docs/about/history/
[21:44:44] <igork> Smithx10: i don't know what is it :)
[21:45:05] <Smithx10> Evidently there is a incredible history about it
[21:45:21] <Smithx10> I heard that Netflix has a documentary coming out and everything
[21:45:31] <sjorge> Does apple still shit dtrace?
[21:45:39] <Smithx10> hahahahah
[21:45:43] <Smithx10> sjorge: ship* ?
[21:45:51] <sjorge> Yes, ship :D
[21:45:55] <Smithx10> I think they were in the processing of shitting dtrace
[21:46:04] <Smithx10> process* !
[21:46:17] <lewellyn> Smithx10: the biggest problem imo is that the common way of determining if one's compiling on solaris is by checking if __sun and __SVR4 (or their alternates) are defined. there's no good way to change that for illumos without breaking the world at large
[21:46:54] <toasterson> BTW what are people's opinion on GOOS illumos implying the solaris tag. It would be my preference?
[21:47:07] <toasterson> On a source level at least.
[21:47:10] <Smithx10> lewellyn: haha, I was just kidding. I've just noticed the social interaction is a little strange is all
[21:47:15] <igork> toasterson: go for it :) i'll wait
[21:47:39] <Smithx10> toasterson: my only concern if that requires someone to have a Solaris builder
[21:47:54] <Smithx10> my only concern is*
[21:47:55] <lewellyn> Smithx10: you brought up bsd. but each bsd is usually identified by e.g. __FreeBSD__. i think we have bsdi to thank for that legacy.
[21:47:59] <toasterson> well ok if the solaris tag dies anyway....
[21:48:16] <toasterson> then things break though...
[21:48:36] <igork> toastersonL it will break a lot in golang components
[21:48:51] <Smithx10> Perhaps from this version on illumos exists and is maintained, and for prior its solaris?
[21:49:03] <toasterson> igork I know. I brought that up myself.
[21:49:09] <igork> toasterson: but it can be start point to work on updates
[21:49:30] <Smithx10> lewellyn: Yeah, that makes it way easier to identify the differnce
[21:49:44] <Smithx10> perhaps that is what leads to the cleaner divergence in conversation / threads
[21:50:05] <Smithx10> toasterson: How would you like to see the tag play out?
[21:50:09] <lewellyn> yeah. also, the BSD forking happened back when UNIX was "relevant"
[21:50:11] <toasterson> i think I'll ask in the github discussion if we need a solaris builder when illumos tag implies solaris for compatibility reasons.
[21:50:25] <Smithx10> lewellyn: Letting the stingers fly!!!
[21:50:31] <Smithx10> :P
[21:51:06] <lewellyn> eh. relevance is largely a matter of "how much do people talk about it?" and these days people talk about linux
[21:51:15] <lewellyn> in 10 years, who knows what might displace that?
[21:51:30] <Smithx10> bezOS?
[21:51:33] <toasterson> smithx10 illumos as a tag implying solaris and they can do with their builders whetever they want. if the go people find that solaris tag should be killed they can do that with a switch statement during build
[21:51:34] <Smithx10> :P
[21:52:11] <Smithx10> yea, I think that's fine. Just as long as we can implement change when required.
[21:52:57] <toasterson> thats what illumos tag is for. so we can implement changed stuff and leave solaris be (in the dust)
[21:53:27] <toasterson> preferably while singing queen songs :P
[21:54:12] <Smithx10> But what happens if Solaris makes a change?
[21:54:21] <ptribble> but we have to match solaris by default, given that that's what pretty much every existing project in go assumes
[21:54:22] <Smithx10> Do we automatically inherit that?
[21:55:01] <toasterson> yes thats the tag implying I was talking about
[21:55:10] <toasterson> android does the same with linux
[21:55:26] <Smithx10> So.... what is the behaviour when we have a conflict?
[21:55:52] <toasterson> they'll need to do a tag for themselves aswell like solaris11+
[21:56:07] <Smithx10> yea, perhaps we can bring that up to Bradtiz on that thread
[21:56:49] <Smithx10> From that version in time, solaris is kind of the generic, and moving forward solaris is implied by illumos and the tag they choose ?
[21:57:15] <toasterson> that would be my preference
[21:57:17] <Smithx10> Does that correct? I think that addresses the existing library problem until we can go touch them when required.
[21:57:22] <Smithx10> I think that is the simplest path forward
[21:57:29] <toasterson> yep
[21:57:36] <Smithx10> LeftWing: any thoughts on that?
[21:57:44] <LeftWing> Smithx10: I'm writing a response
[21:57:59] <igork> LeftWing: what is the process for update some wiki pages ?
[21:58:03] <igork> https://illumos.org/docs/about/who/
[21:58:06] <toasterson> btw. alanc want to join that discussion from solaris side?
[21:58:07] <igork> this one for example
[21:58:24] <LeftWing> igork: Read the contributing section?
[21:58:41] <igork> 4 links
[21:58:47] <igork> what is for wiki ?
[21:59:08] <Smithx10> igork: are you gonna fix our Wunix History?!$!$!
[21:59:10] <Smithx10> How dare you!
[21:59:18] <igork> Smithx10: nope :)
[21:59:25] <Smithx10> YES, FOREVER WUNIX!
[21:59:32] <igork> it's for you ;)
[21:59:43] <LeftWing> igork: Have a look at the front page of the documentation section. It explicitly mentions the contribution stuff.
[22:00:11] <igork> https://illumos.org/docs/contributing/docs/
[22:00:25] <igork> is it correct?
[22:00:48] <igork> i just try identify if i'm on right place
[22:04:28] <rmustacc> Yes. Is there some other page that you were considering it might be?
[22:04:30] <ptribble> LeftWing: I have a handful of potential IPDs, what's the best way to get them moving as part of the IPD process?
[22:05:11] <LeftWing> ptribble: Well, first I can add you to the repository
[22:05:42] <ptribble> That would be great
[22:06:15] <LeftWing> I envisage we'll use statuses of: predraft, before you're asking for comment; draft, as you put it on on developer@ asking for feedback; published, as we're all kind of agreed on how to move forward and you're going to start working etc
[22:06:32] <LeftWing> richlowe keeps prodding me to actually finish IPD 4 and get it out on the list for instance
[22:07:07] <LeftWing> Once you're ready to seek feedback I think we'll want to nominate an advocate, which will probably be me, to kind of watch over it and get us on a path to consensus without too much consternation
[22:07:15] <LeftWing> I'll add you now though!
[22:09:23] <ptribble> The first one would be to try and get some official status on the SPARC purge, which is probably draft (and has had some discussion already)
[22:11:47] *** baojg_ <baojg_!~baojg@162.243.44.213> has joined #illumos
[22:14:05] <LeftWing> Yeah, that sounds fine too -- I know you've been discussing that one with folks. It'd be great to have a plan that covers what you've done so far, and where we're going next!
[22:14:32] *** baojg <baojg!~baojg@162.243.44.213> has quit IRC (Ping timeout: 252 seconds)
[22:15:17] <LeftWing> It'd be good to have a serious plan for how we might get to project-level build/test host(s) for SPARC too
[22:15:26] <LeftWing> I'm tinkering with Jenkins but I obviously only have x86 myself
[22:16:20] <lewellyn> ptribble: is the general consensus still basically "intersection of affordable availability and resultant unremovable cruft"?
[22:17:03] <ptribble> something like that
[22:18:27] <Smithx10> Is there any Baremetal cloud provider like Packet that have sparc boxes?
[22:19:47] <ptribble> I'm not aware of any provider (I would use them if there were)
[22:20:01] *** ed209 <ed209!~ed209@165.225.128.67> has quit IRC (Remote host closed the connection)
[22:20:07] *** ed209 <ed209!~ed209@165.225.128.67> has joined #illumos
[22:20:27] <LeftWing> I'd use a V210 at Tribble's Technicolour Wondercloud
[22:20:31] <LeftWing> You should start one :D
[22:20:48] <ptribble> Operationally, it's fairly easy to put a sparc box somewhere. I looked at the cost for my boxes and it's more than I can justify.
[22:21:34] <lewellyn> Smithx10: most places will take a sparc box and pop it into a rack for you, if nothing else
[22:22:34] <lewellyn> once i'm gainfully employed again, i plan on getting a rack at a local facility and probably ending up with at least one sparc box supported by illumos once the list of supportable machines is hammered out
[22:23:14] <lewellyn> there's a facility mere blocks from one of the houses i've been wanting. talk about convenient :D
[22:24:27] <Smithx10> ErrRrr I just lost 6 bucks to AWS.
[22:24:47] <Smithx10> Spun up a fargate cluster to test the Virtual Kubelet to see it how it worked.... and forgot.
[22:24:57] <Smithx10> ./cries a river.
[22:28:42] <Smithx10> w00t, looks we are in a good place with the build tag
[22:29:00] <Smithx10> Everyone, it's all on LeftWing's Shoulders now :P
[22:29:04] <Smithx10> ./loves open source
[22:29:05] <jbk> but didn't try to steal cable I hope :P
[22:34:34] <arekinath> could always hit up a few universities and see if any are willing to keep a sparc box in a corner for us :P
[22:44:34] <Smithx10> Maybe GCP should give us one for Illumos GO SPARC :P
[22:56:39] *** rsully <rsully!~rsully@unaffiliated/rsully> has quit IRC (Quit: rsully)
[23:10:43] *** lewellyn <lewellyn!~lewellyn@about/windows/lewellyn> has quit IRC (Remote host closed the connection)
[23:11:12] *** lewellyn <lewellyn!~lewellyn@about/windows/lewellyn> has joined #illumos
[23:11:12] *** ChanServ sets mode: +o lewellyn
[23:17:44] <toasterson> Smithx10 still playing with kubernetes?
[23:18:14] <LeftWing> Is there actually a SPARC Go thing already?
[23:18:26] <LeftWing> I could have sworn I recall Aram talking about it at some point in the past
[23:19:19] <lewellyn> i recall it being talked about but i never knew if anything came of it
[23:19:29] <toasterson> Leftwing yes there are working ones
[23:19:35] <toasterson> Linux only afaik
[23:19:56] <LeftWing> That's exciting
[23:19:56] <toasterson> People were telling they were using it.
[23:20:22] <LeftWing> Well one imagines if they have code generation for SPARC, it could be adapted to the Solaris ABI with relative ease
[23:20:37] <toasterson> But not in main repo. several forks that would need tender love and care to integrate.
[23:20:43] <igork> LeftWing: Go on sparc has issues with unwind
[23:20:54] <igork> but you can compile it as static
[23:21:09] <toasterson> unwind?
[23:21:30] <LeftWing> Go is pretty much all static, aside from the obvious need to use libc.so.1 for the system call entry points
[23:21:35] <igork> # gccgo-6 SPARC
[23:21:36] <igork> gccgo -m64 -v -static-libgo -static-libgcc -o hello hello.go -Wl,-dy -lnsl -lsocket -lrt -lsendfile
[23:21:41] <alanc> Aram had a working SPARC Solaris Go port for 1.7 or so, but then upstream rewrote the compiler and broke everything and we ended up deciding Go was stupid and useless
[23:21:59] <alanc> so we're going with gccgo instead
[23:22:03] <LeftWing> I see
[23:22:13] <LeftWing> Does anybody in the Go community actually use gccgo in anger?
[23:22:32] <LeftWing> It feels like the attention is chiefly on the reference implementation
[23:22:35] <alanc> also, our primary driver for Go was to support Docker, and once that team got laid off, much less interest...
[23:22:37] <LeftWing> But maybe that's just my percerption
[23:23:00] <alanc> toasterson: I don't know where the discussion is happening that you asked about
[23:23:20] <LeftWing> Yeah, I care mostly because there are some popular bodies of software now written in it. Of particular note, for Joyent anyway, is Prometheus.
[23:23:32] <toasterson> alanc https://github.com/golang/go/issues/20603
[23:23:41] <igork> LeftWing: you have to build golang on sparc first and can use gccgo. but with broken unwind need some hacks
[23:23:51] <alanc> also, see plenty of other people complaining that if your platform is not of interest to Google, it's nearly impossible to get good upstream support in golang for it
[23:24:31] <igork> alanc: it's true not only with google :)
[23:26:38] <toasterson> google pushes features. the rest of the community is very conservative. Which is waht I like.
[23:27:35] <LeftWing> alanc: I'm giving direct engagement a shot.
[23:28:07] <LeftWing> First I have to make this clunky rust bucket of a UNIX run on the GCE hypervisor.
[23:29:10] <alanc> https://illumos.org/docs/about/who/ is missing Microsoft DTrace, though I guess that's coming indirectly via FreeBSD
[23:29:24] <toasterson> Leftwing why? bradfitz said that is optional.
[23:29:34] <LeftWing> toasterson: Because the thing he's describing otherwise isn't super good
[23:29:51] <LeftWing> It would be much easier for everybody if we could just boot a guest. I'm 25% of the way in I reckon.
[23:30:08] <LeftWing> I got it to boot on Sunday. I'm reaching out to the GCE folks to try and get a technical issue sorted out on their end.
[23:30:28] <LeftWing> In the meantime vioif needs to support hypervisors that don't have indirect descriptors (not too bad)
[23:30:39] <LeftWing> and then we need vioscsi, which honestly also shouldn't be too bad
[23:30:52] <LeftWing> Most of the work will be done by SCSA in that case
[23:30:59] <toasterson> is vioblock less performant than vioscsi?
[23:31:23] <LeftWing> Err not clear, really. It's very simple, but the simplicity presents challenges. I can see why they went with vioscsi, to be honest.
[23:31:36] <LeftWing> e.g., hotplugging vioblk disks means PCI hotplug
[23:31:48] <LeftWing> hotplugging vioscsi disks means SCSI hotplug, which we already do well for SAS/SATA devices
[23:32:17] <LeftWing> Also the number of LUNs you can have in the vioblk world is constrained by the number of PCI devices you can usefully expose on a bus, etc
[23:32:31] <LeftWing> In the vioscsi world you get one "controller" device, with a large namespace for LUNs underneath
[23:32:51] <ptribble> It would be good to support GCE anyway; I played with vioscsi on qemu and it fell over fairly quickly
[23:32:54] <Smithx10> toasterson: Nah, I moved on to Nomad. It allows you to bring your own schema
[23:32:54] <Smithx10> https://github.com/Smithx10/nomad-driver-triton
[23:33:16] <Smithx10> I've gotta wrte some tests and clean up a bit, but its functionining
[23:33:50] <toasterson> yeah we have a backupserver on libvirt with virtioblk and I felt it was somehow sluggish
[23:33:58] <Smithx10> toasterson: virtual kubelete had a github issue open for deletepod, and I don't want to maintain that with them
[23:34:17] <Smithx10> lol, if you returned an error to delete pod...... it would still succeed. "_"
[23:34:40] <Smithx10> I'll revist that in a few months or so.... perhaps when deleting works.
[23:34:50] <toasterson> welp... awesome... I'll have to look into nomad then.
[23:35:20] <Smithx10> Yea, its a simpler interface.... I'm hoping to at some point get Nomad Agent to run in the GZ
[23:35:42] <Smithx10> There was some work done already for that.... for Vanilla SmartOS that could be quite useful
[23:35:47] <LeftWing> Nomad isn't my favourite piece of software, but I feel like it's not trying to be an all-encompassing platform like The Kube.
[23:35:55] <LeftWing> Which ... is probably for the best.
[23:35:58] <Smithx10> LeftWing: yea, definitely
[23:36:13] <Smithx10> I think it is 20% as opinionated, and achieves about 80% the same
[23:36:45] <Smithx10> I still Prefer triton, the Virtual Networking and the entire triton eco system would be hard to recreate for little gain
[23:37:11] <Smithx10> Now that Affinity Rules arent racing ... wrapping CloudAPI is pretty valuable
[23:37:28] <toasterson> Would be nice to have for OI aswell. I want to give my OI zone hosts some better management tools.
[23:37:39] <toasterson> And see what from Triton I can integrate
[23:37:43] <Smithx10> toasterson: you bring up a good point
[23:37:54] <Smithx10> Does OI have vmadm and those tools?
[23:38:00] <toasterson> nope
[23:38:02] <Smithx10> :(
[23:38:08] <LeftWing> It has zonecfg/zoneadm though
[23:38:13] <LeftWing> You could drive those from Nomad too
[23:38:22] <Smithx10> yea, I was trying to figure out where to do the abstraction
[23:38:30] <LeftWing> I think on SmartOS you should use vmadm
[23:38:35] <LeftWing> Because that's really all we promise will work
[23:38:46] <Smithx10> I looked at the vmadm code
[23:38:48] <LeftWing> On OI/OmniOS though, zonecfg/zoneadm are the first class tools
[23:39:01] <Smithx10> and was like......idk if i want to venture ino all of that
[23:39:11] <lewellyn> alanc: re: platforms vs go: there are people upset that go wants to drop support for plan 9 rather than fixing the performance issues which have creeped up in recent Go versions. my gut from the conversations i've seen says that 9 is just showing that there's a huge flaw needing fixing, so killing platform support for a canary in a coal mine seems... potentially ungood
[23:39:17] <LeftWing> Smithx10: vmadm is definitely ... big
[23:39:25] <Smithx10> hahahahahahaha yes.
[23:39:50] *** jthan <jthan!~jonathan@maxilaria.jrdemasi.com> has joined #illumos
[23:40:01] <Smithx10> And its pretty much doing all of the work that the noamd agent would do.... so maybe first thing to do is to make the library in go for vmadm
[23:40:12] <toasterson> I want to drive zoneadm and have ported the zonecfg code to go already some months ago. So I want some tools taht can support some docker like (but better) workflow and do image stuff.
[23:40:38] <Smithx10> toasterson: what is reason people don't use vmadm?
[23:40:43] <Smithx10> do you have imgadm?
[23:41:00] <toasterson> nope neither.
[23:41:20] <toasterson> mostly because they are/were in the illumos-joyent repo and not seperate
[23:41:30] <toasterson> not clear if that is still the case.
[23:42:04] <toasterson> isn't vmadm javascript?
[23:42:15] <Smithx10> you say that.... like its bad...... :P
[23:42:18] <LeftWing> vmadm and imgadm are very opiniated, and make a lot of assumptions
[23:42:26] <LeftWing> I can see why people haven't adopted them on other illumos distributions
[23:43:19] <toasterson> we have enough trouble with multiple node releases not working properly...
[23:43:49] <toasterson> and nobody likes to write json by hand :)
[23:43:58] <Smithx10> toasterson: or ..... do we wait for the new rust revolution?
[23:44:16] <toasterson> that would probably wotk.
[23:44:27] <toasterson> isn't there already a vmadmin port in rust?
[23:44:38] <Smithx10> yea, Ive never used it.
[23:44:40] *** rsmarples <rsmarples!~roy@cpc115040-bour7-2-0-cust370.15-1.cable.virginm.net> has left #illumos ("WeeChat 2.2")
[23:44:44] <Smithx10> I don't know how complete it is.
[23:45:53] <toasterson> me neither. mostly because rust wasn't available on OI back then.
[23:46:30] <toasterson> I don't know if it supports zones at all. It was for freebsd
[23:57:25] <LeftWing> I recall somebody was working on something called "zcage" recently
[23:57:29] *** merzo <merzo!~merzo@45-47-132-95.pool.ukrtel.net> has joined #illumos
[23:59:16] <andyf> https://github.com/cneira/zcage
[23:59:27] <LeftWing> That one! Yes.
top

   April 15, 2019
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30