Switch to DuckDuckGo Search
   November 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
[00:01:10] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has joined #illumos
[00:10:08] *** andy_js <andy_js!~andy@94.11.180.149> has quit IRC (Quit: andy_js)
[00:13:27] <Reinhilde> LeftWing: this may be germane to the dtrace i sent you a while ago
[00:13:31] <Reinhilde> unlikely though
[00:33:25] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has quit IRC (Read error: Connection reset by peer)
[00:33:46] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has joined #illumos
[00:47:53] *** jellydonut <jellydonut!~quassel@s91904424.blix.com> has quit IRC (Quit: jellydonut)
[00:50:15] *** baojg <baojg!~baojg@2604:a880:0:1010::16:1> has quit IRC (Remote host closed the connection)
[00:51:11] *** jellydonut <jellydonut!~quassel@s91904424.blix.com> has joined #illumos
[00:58:24] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has quit IRC (Remote host closed the connection)
[01:09:22] *** arnoldoree <arnoldoree!~arnoldore@ranoldoree.plus.com> has quit IRC (Quit: Leaving)
[01:12:20] *** mutin-sa <mutin-sa!~s-mutin@85.234.114.134> has quit IRC (Read error: Connection reset by peer)
[01:12:49] *** mutin-sa <mutin-sa!~s-mutin@85.234.114.134> has joined #illumos
[01:14:14] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has quit IRC (Remote host closed the connection)
[01:14:38] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has joined #illumos
[01:28:15] *** Kruppt <Kruppt!~Kruppt@50.111.55.255> has quit IRC (Quit: Leaving)
[01:55:57] *** Kurlon <Kurlon!~Kurlon@167-192.gwi.net> has joined #illumos
[01:56:34] *** Kurlon <Kurlon!~Kurlon@167-192.gwi.net> has quit IRC (Remote host closed the connection)
[01:57:09] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has joined #illumos
[02:02:39] *** Qatz <Qatz!~db@2601:187:8400:5::d9d> has quit IRC (Read error: Connection reset by peer)
[02:03:00] *** Qatz <Qatz!~db@2601:187:8400:5::d9d> has joined #illumos
[02:11:30] *** chandlore____ <chandlore____!sid259138@gateway/web/irccloud.com/x-jeqnqjtakgwalsrj> has quit IRC (Ping timeout: 252 seconds)
[02:11:30] *** jkni <jkni!sid85151@gateway/web/irccloud.com/x-plsrzrnipoapxzcg> has quit IRC (Ping timeout: 252 seconds)
[02:12:31] *** rzezeski <rzezeski!uid151901@gateway/web/irccloud.com/x-jbjjlxskonuimolp> has quit IRC (Ping timeout: 252 seconds)
[02:12:37] *** jkni <jkni!sid85151@gateway/web/irccloud.com/x-itbdvwyurmfjjwdi> has joined #illumos
[02:13:20] *** chandlore____ <chandlore____!sid259138@gateway/web/irccloud.com/x-ygrfpawfbebryuvy> has joined #illumos
[02:13:22] *** rzezeski <rzezeski!uid151901@gateway/web/irccloud.com/x-xxeikvpanxavulaz> has joined #illumos
[02:14:50] *** ed209 <ed209!~ed209@165.225.128.67> has quit IRC (Ping timeout: 240 seconds)
[02:20:07] *** ed209 <ed209!~ed209@165.225.128.67> has joined #illumos
[02:37:48] *** kev009 <kev009!~kev009@ip72-222-200-117.ph.ph.cox.net> has quit IRC (Remote host closed the connection)
[02:42:09] *** kev009 <kev009!~kev009@ip72-222-200-117.ph.ph.cox.net> has joined #illumos
[02:50:25] *** kev009 <kev009!~kev009@ip72-222-200-117.ph.ph.cox.net> has quit IRC (Quit: Konversation terminated!)
[02:51:13] *** kev009 <kev009!~kev009@ip72-222-200-117.ph.ph.cox.net> has joined #illumos
[02:53:10] *** kev009 <kev009!~kev009@ip72-222-200-117.ph.ph.cox.net> has quit IRC (Client Quit)
[02:53:25] *** Kurlon_ <Kurlon_!~Kurlon@167-192.gwi.net> has joined #illumos
[02:54:18] *** kev009 <kev009!~kev009@ip72-222-200-117.ph.ph.cox.net> has joined #illumos
[02:56:43] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has quit IRC (Ping timeout: 250 seconds)
[02:58:20] *** Kurlon_ <Kurlon_!~Kurlon@167-192.gwi.net> has quit IRC (Ping timeout: 268 seconds)
[03:15:25] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has quit IRC (Quit: jcea)
[03:18:22] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[03:53:35] *** Kurlon <Kurlon!~Kurlon@cpe-67-253-141-249.rochester.res.rr.com> has joined #illumos
[03:58:45] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[04:04:13] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[04:27:17] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has quit IRC (Ping timeout: 276 seconds)
[04:37:35] *** despair86 <despair86!~despair86@24.170.8.11> has joined #illumos
[05:10:27] <richlowe> LeftWing: total old man moment. OSX doesn't have gprof!
[05:13:19] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[05:13:41] <LeftWing> Do we?
[05:13:51] <richlowe> Dude, of course.
[05:13:54] <richlowe> pretty sure we have _prof_
[05:13:59] <richlowe> for those for whom gprof is too new
[05:15:13] <LeftWing> Ha
[05:15:33] <LeftWing> I confess that I do not believe I have ever used either
[05:20:37] <jbk> i used one of them
[05:21:00] <jbk> (only to verify that the demangling hooks that were already present worked w/ the new demangler)
[05:23:37] *** despair86 <despair86!~despair86@24.170.8.11> has quit IRC (Ping timeout: 250 seconds)
[05:30:21] *** despair86 <despair86!~despair86@24.170.8.11> has joined #illumos
[05:34:31] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[05:41:26] *** baojg <baojg!~baojg@2604:a880:0:1010::16:1> has joined #illumos
[05:44:14] *** sjorge <sjorge!~sjorge@unaffiliated/sjorge> has quit IRC (Ping timeout: 246 seconds)
[05:46:34] *** sjorge <sjorge!~sjorge@unaffiliated/sjorge> has joined #illumos
[05:50:05] *** kev009 <kev009!~kev009@ip72-222-200-117.ph.ph.cox.net> has quit IRC (Remote host closed the connection)
[05:55:38] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[05:57:08] *** kev009 <kev009!~kev009@ip72-222-200-117.ph.ph.cox.net> has joined #illumos
[06:00:14] *** lblume <lblume!~lblume@greenviolet/laoyijiehe/lblume> has quit IRC (Ping timeout: 276 seconds)
[06:04:36] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[06:08:12] *** despair86 <despair86!~despair86@24.170.8.11> has quit IRC (Remote host closed the connection)
[06:12:31] *** despair86 <despair86!~despair86@24.170.8.11> has joined #illumos
[06:24:37] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Quit: Textual IRC Client: www.textualapp.com)
[07:18:04] *** clapont <clapont!~clapont@unaffiliated/clapont> has joined #illumos
[07:29:28] *** clapont <clapont!~clapont@unaffiliated/clapont> has quit IRC (Remote host closed the connection)
[07:30:12] *** joltman_afk <joltman_afk!znc@gateway/vpn/privateinternetaccess/joltman> has joined #illumos
[07:33:02] *** Agnar_ <Agnar_!~olbohlen@eenfach.de> has joined #illumos
[07:33:15] *** mnowak__ <mnowak__!~mnowak_@94.142.238.232> has joined #illumos
[07:33:22] *** pmooney1 <pmooney1!~pmooney@67-4-175-230.mpls.qwest.net> has joined #illumos
[07:34:35] *** nbaksalyar_ <nbaksalyar_!~nbaksalya@188.166.7.50> has joined #illumos
[07:35:11] *** despair86 <despair86!~despair86@24.170.8.11> has quit IRC (Ping timeout: 246 seconds)
[07:35:18] *** dlg_ <dlg_!~dlg@toy.eait.uq.edu.au> has joined #illumos
[07:37:30] *** igork <igork!~igork@178.162.50.254> has quit IRC (*.net *.split)
[07:37:30] *** CaptainTobin <CaptainTobin!~tobin@c-68-38-10-41.hsd1.in.comcast.net> has quit IRC (*.net *.split)
[07:38:39] *** jwit <jwit!~jwit@gw.introvert.ca> has quit IRC (*.net *.split)
[07:38:39] *** dopplergange <dopplergange!~dop@45.12.223.70> has quit IRC (*.net *.split)
[07:38:39] *** leoric <leoric!~alp@pyhalov.cc.rsu.ru> has quit IRC (*.net *.split)
[07:38:39] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has quit IRC (*.net *.split)
[07:38:39] *** jessfraz_ <jessfraz_!~jessfraz@unaffiliated/jessfraz> has quit IRC (*.net *.split)
[07:38:39] *** yomisei <yomisei!~void@ip4d16bc15.dynamic.kabel-deutschland.de> has quit IRC (*.net *.split)
[07:38:39] *** jemershaw <jemershaw!~jemershaw@c-68-83-252-28.hsd1.pa.comcast.net> has quit IRC (*.net *.split)
[07:38:39] *** joltman <joltman!znc@gateway/vpn/privateinternetaccess/joltman> has quit IRC (*.net *.split)
[07:38:39] *** Agnar <Agnar!~olbohlen@eenfach.de> has quit IRC (*.net *.split)
[07:38:39] *** bahamas10 <bahamas10!~dave@cpe-72-231-187-25.nycap.res.rr.com> has quit IRC (*.net *.split)
[07:38:39] *** Mazon_ <Mazon_!~Mazon@37.205.127.168> has quit IRC (*.net *.split)
[07:38:39] *** dlg <dlg!~dlg@toy.eait.uq.edu.au> has quit IRC (*.net *.split)
[07:38:39] *** bahamat <bahamat!~bahamat@72.2.113.233> has quit IRC (*.net *.split)
[07:38:39] *** schily <schily!~joerg@fokus171239.fokus.fraunhofer.de> has quit IRC (*.net *.split)
[07:38:39] *** goberle <goberle!~goberle@mid.ygg.tf> has quit IRC (*.net *.split)
[07:38:39] *** pmooney <pmooney!~pmooney@67-4-175-230.mpls.qwest.net> has quit IRC (*.net *.split)
[07:38:39] *** edef <edef!edef@NixOS/user/edef> has quit IRC (*.net *.split)
[07:38:39] *** anthis <anthis!~anthis@essential.coach> has quit IRC (*.net *.split)
[07:38:39] *** Tempt <Tempt!~avenger@unaffiliated/tempt> has quit IRC (*.net *.split)
[07:38:39] *** rmustacc <rmustacc!~daemar@zelgadis.fingolfin.org> has quit IRC (*.net *.split)
[07:38:39] *** perlgod <perlgod!~cullum@puffy.c0ffee.net> has quit IRC (*.net *.split)
[07:38:39] *** nbaksalyar <nbaksalyar!~nbaksalya@188.166.7.50> has quit IRC (*.net *.split)
[07:38:39] *** joltman_afk is now known as joltman
[07:38:39] *** matt_c <matt_c!~mcroydon@162.243.108.144> has quit IRC (*.net *.split)
[07:38:39] *** _mjg <_mjg!mjg@fook.org> has quit IRC (*.net *.split)
[07:38:39] *** mpana <mpana!~mpana@193.226.149.49> has quit IRC (*.net *.split)
[07:38:39] *** despair86 <despair86!~despair86@24.170.8.11> has joined #illumos
[07:39:17] *** pwinder <pwinder!~pwinder@86.2.210.254> has joined #illumos
[07:42:45] *** matt_c <matt_c!~mcroydon@162.243.108.144> has joined #illumos
[07:42:45] *** _mjg <_mjg!mjg@fook.org> has joined #illumos
[07:42:45] *** mpana <mpana!~mpana@193.226.149.49> has joined #illumos
[07:43:03] *** edef <edef!edef@NixOS/user/edef> has joined #illumos
[07:44:27] *** yomisei <yomisei!~void@ip4d16bc15.dynamic.kabel-deutschland.de> has joined #illumos
[07:48:52] *** igork <igork!~igork@178.162.50.254> has joined #illumos
[07:48:52] *** CaptainTobin <CaptainTobin!~tobin@c-68-38-10-41.hsd1.in.comcast.net> has joined #illumos
[07:48:52] *** jwit <jwit!~jwit@gw.introvert.ca> has joined #illumos
[07:48:52] *** dopplergange <dopplergange!~dop@45.12.223.70> has joined #illumos
[07:48:52] *** leoric <leoric!~alp@pyhalov.cc.rsu.ru> has joined #illumos
[07:48:52] *** jessfraz_ <jessfraz_!~jessfraz@unaffiliated/jessfraz> has joined #illumos
[07:48:52] *** jemershaw <jemershaw!~jemershaw@c-68-83-252-28.hsd1.pa.comcast.net> has joined #illumos
[07:48:52] *** bahamas10 <bahamas10!~dave@cpe-72-231-187-25.nycap.res.rr.com> has joined #illumos
[07:48:52] *** Mazon_ <Mazon_!~Mazon@37.205.127.168> has joined #illumos
[07:48:52] *** bahamat <bahamat!~bahamat@72.2.113.233> has joined #illumos
[07:48:52] *** schily <schily!~joerg@fokus171239.fokus.fraunhofer.de> has joined #illumos
[07:48:52] *** goberle <goberle!~goberle@mid.ygg.tf> has joined #illumos
[07:48:52] *** anthis <anthis!~anthis@essential.coach> has joined #illumos
[07:48:52] *** Tempt <Tempt!~avenger@unaffiliated/tempt> has joined #illumos
[07:48:52] *** rmustacc <rmustacc!~daemar@zelgadis.fingolfin.org> has joined #illumos
[07:48:52] *** perlgod <perlgod!~cullum@puffy.c0ffee.net> has joined #illumos
[07:48:52] *** orwell.freenode.net sets mode: +oo Tempt rmustacc
[08:00:09] *** glasspelican <glasspelican!~quassel@2607:5300:201:3100::325> has quit IRC (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.)
[08:01:10] *** glasspelican <glasspelican!~quassel@2607:5300:201:3100::325> has joined #illumos
[08:08:18] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Quit: tsoome)
[08:15:18] *** igork <igork!~igork@178.162.50.254> has quit IRC (*.net *.split)
[08:15:18] *** CaptainTobin <CaptainTobin!~tobin@c-68-38-10-41.hsd1.in.comcast.net> has quit IRC (*.net *.split)
[08:15:19] *** jwit <jwit!~jwit@gw.introvert.ca> has quit IRC (*.net *.split)
[08:15:19] *** dopplergange <dopplergange!~dop@45.12.223.70> has quit IRC (*.net *.split)
[08:15:19] *** leoric <leoric!~alp@pyhalov.cc.rsu.ru> has quit IRC (*.net *.split)
[08:15:20] *** jessfraz_ <jessfraz_!~jessfraz@unaffiliated/jessfraz> has quit IRC (*.net *.split)
[08:15:20] *** jemershaw <jemershaw!~jemershaw@c-68-83-252-28.hsd1.pa.comcast.net> has quit IRC (*.net *.split)
[08:15:20] *** bahamas10 <bahamas10!~dave@cpe-72-231-187-25.nycap.res.rr.com> has quit IRC (*.net *.split)
[08:15:21] *** Mazon_ <Mazon_!~Mazon@37.205.127.168> has quit IRC (*.net *.split)
[08:15:21] *** bahamat <bahamat!~bahamat@72.2.113.233> has quit IRC (*.net *.split)
[08:15:21] *** schily <schily!~joerg@fokus171239.fokus.fraunhofer.de> has quit IRC (*.net *.split)
[08:15:22] *** goberle <goberle!~goberle@mid.ygg.tf> has quit IRC (*.net *.split)
[08:15:22] *** anthis <anthis!~anthis@essential.coach> has quit IRC (*.net *.split)
[08:15:23] *** Tempt <Tempt!~avenger@unaffiliated/tempt> has quit IRC (*.net *.split)
[08:15:23] *** rmustacc <rmustacc!~daemar@zelgadis.fingolfin.org> has quit IRC (*.net *.split)
[08:15:23] *** perlgod <perlgod!~cullum@puffy.c0ffee.net> has quit IRC (*.net *.split)
[08:16:31] <poige> how to get large page usage metrics, system-global ones, not just a process’?
[08:16:56] *** igork <igork!~igork@178.162.50.254> has joined #illumos
[08:16:56] *** CaptainTobin <CaptainTobin!~tobin@c-68-38-10-41.hsd1.in.comcast.net> has joined #illumos
[08:16:56] *** jwit <jwit!~jwit@gw.introvert.ca> has joined #illumos
[08:16:56] *** dopplergange <dopplergange!~dop@45.12.223.70> has joined #illumos
[08:16:56] *** leoric <leoric!~alp@pyhalov.cc.rsu.ru> has joined #illumos
[08:16:56] *** jessfraz_ <jessfraz_!~jessfraz@unaffiliated/jessfraz> has joined #illumos
[08:16:56] *** jemershaw <jemershaw!~jemershaw@c-68-83-252-28.hsd1.pa.comcast.net> has joined #illumos
[08:16:56] *** bahamas10 <bahamas10!~dave@cpe-72-231-187-25.nycap.res.rr.com> has joined #illumos
[08:16:56] *** Mazon_ <Mazon_!~Mazon@37.205.127.168> has joined #illumos
[08:16:56] *** bahamat <bahamat!~bahamat@72.2.113.233> has joined #illumos
[08:16:56] *** schily <schily!~joerg@fokus171239.fokus.fraunhofer.de> has joined #illumos
[08:16:56] *** goberle <goberle!~goberle@mid.ygg.tf> has joined #illumos
[08:16:56] *** anthis <anthis!~anthis@essential.coach> has joined #illumos
[08:16:56] *** Tempt <Tempt!~avenger@unaffiliated/tempt> has joined #illumos
[08:16:56] *** rmustacc <rmustacc!~daemar@zelgadis.fingolfin.org> has joined #illumos
[08:16:56] *** perlgod <perlgod!~cullum@puffy.c0ffee.net> has joined #illumos
[08:16:56] *** orwell.freenode.net sets mode: +oo Tempt rmustacc
[08:17:41] <poige> seems to be not covered subject anywhere on WEB
[08:22:51] *** amrfrsh <amrfrsh!~Thunderbi@185.212.171.68> has quit IRC (Quit: amrfrsh)
[08:38:32] *** alanc <alanc!~alanc@inet-hqmc02-o.oracle.com> has quit IRC (Remote host closed the connection)
[08:43:19] *** alanc <alanc!~alanc@inet-hqmc02-o.oracle.com> has joined #illumos
[08:49:47] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[08:51:42] *** pwinder <pwinder!~pwinder@86.2.210.254> has quit IRC (Quit: This computer has gone to sleep)
[09:15:36] *** despair86 <despair86!~despair86@24.170.8.11> has quit IRC (Remote host closed the connection)
[09:25:04] *** amrfrsh <amrfrsh!~Thunderbi@131.234.44.76> has joined #illumos
[09:35:54] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has joined #illumos
[09:51:12] *** lblume <lblume!~lblume@greenviolet/laoyijiehe/lblume> has joined #illumos
[10:07:14] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has quit IRC (Ping timeout: 276 seconds)
[10:08:34] *** arnoldoree <arnoldoree!~arnoldore@ranoldoree.plus.com> has joined #illumos
[10:13:01] *** andy_js <andy_js!~andy@94.11.180.149> has joined #illumos
[10:13:54] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has joined #illumos
[10:24:39] *** man_u <man_u!~manu@manu2.gandi.net> has joined #illumos
[10:29:07] *** amrfrsh <amrfrsh!~Thunderbi@131.234.44.76> has quit IRC (Ping timeout: 250 seconds)
[11:13:08] *** BH23 <BH23!~BH23@193.117.206.132> has joined #illumos
[11:17:44] *** amrfrsh <amrfrsh!~Thunderbi@131.234.44.76> has joined #illumos
[11:20:00] *** ed209 <ed209!~ed209@165.225.128.67> has quit IRC (Remote host closed the connection)
[11:20:06] *** ed209 <ed209!~ed209@165.225.128.67> has joined #illumos
[12:23:35] *** amrfrsh <amrfrsh!~Thunderbi@131.234.44.76> has quit IRC (Quit: amrfrsh)
[12:35:26] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has quit IRC (Ping timeout: 276 seconds)
[12:37:45] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has joined #illumos
[12:52:03] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has quit IRC (Ping timeout: 252 seconds)
[12:54:51] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has joined #illumos
[13:17:15] *** sjorge <sjorge!~sjorge@unaffiliated/sjorge> has quit IRC (Remote host closed the connection)
[13:20:56] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has quit IRC (Ping timeout: 240 seconds)
[13:21:28] *** sjorge <sjorge!~sjorge@unaffiliated/sjorge> has joined #illumos
[13:24:12] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has joined #illumos
[13:31:29] *** sjorge <sjorge!~sjorge@unaffiliated/sjorge> has quit IRC (Ping timeout: 246 seconds)
[13:32:03] *** insomnia <insomnia!~insomnia@shadowcat/actuallyamemberof/lollipopguild.insomnia> has quit IRC (Ping timeout: 265 seconds)
[13:37:11] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has quit IRC (Ping timeout: 276 seconds)
[13:38:08] *** sjorge <sjorge!~sjorge@unaffiliated/sjorge> has joined #illumos
[13:43:19] *** psarria <psarria!~psarria@26.red-79-146-96.dynamicip.rima-tde.net> has quit IRC (Remote host closed the connection)
[13:50:48] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has joined #illumos
[13:56:12] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has quit IRC (Remote host closed the connection)
[14:04:34] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has joined #illumos
[14:15:17] *** despair86 <despair86!~despair86@24.170.8.11> has joined #illumos
[14:21:34] *** amrfrsh <amrfrsh!~Thunderbi@131.234.44.76> has joined #illumos
[14:25:38] *** despair86 <despair86!~despair86@24.170.8.11> has quit IRC (Ping timeout: 240 seconds)
[14:36:51] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has joined #illumos
[15:02:44] <gitomat> [illumos-gate] 11939 libgen: this statement may fall through -- Toomas Soome <tsoome at me dot com>
[15:03:36] *** amrfrsh <amrfrsh!~Thunderbi@131.234.44.76> has quit IRC (Quit: amrfrsh)
[15:06:25] <gitomat> [illumos-gate] 11940 loader: multiboot2_loadfile read error should include file name and newline -- Toomas Soome <tsoome at me dot com>
[15:09:29] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has quit IRC (Ping timeout: 276 seconds)
[15:09:35] *** insomnia <insomnia!~insomnia@shadowcat/actuallyamemberof/lollipopguild.insomnia> has joined #illumos
[15:10:04] *** pmooney1 <pmooney1!~pmooney@67-4-175-230.mpls.qwest.net> has quit IRC (Quit: WeeChat 2.6)
[15:10:19] *** pmooney <pmooney!~pmooney@67-4-175-230.mpls.qwest.net> has joined #illumos
[15:11:38] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has joined #illumos
[15:12:00] <gitomat> [illumos-gate] 11949 make -C caches wrong directory contents -- Andy Fiddaman <omnios at citrus-it dot co.uk>
[15:26:08] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Ping timeout: 245 seconds)
[15:27:55] *** pwinder <pwinder!~pwinder@86.2.210.254> has joined #illumos
[15:32:41] *** wl_ <wl_!~wl_@2605:6000:1b0c:6060::87c> has quit IRC (Quit: Leaving)
[15:38:05] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has quit IRC (Ping timeout: 276 seconds)
[15:48:08] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has joined #illumos
[15:51:20] <KungFuJesus> eh, where's the man page for open()?
[15:51:29] <KungFuJesus> man 3 open won't give me it
[15:51:29] <andyf> man open.2 ?
[15:51:46] <KungFuJesus> man 2 open won't either
[15:52:00] <andyf> use either 'man -s2 open' or 'man open.2'
[15:52:00] <KungFuJesus> ah, open.2 is needed why?
[15:52:23] <KungFuJesus> I mean I'm sure it's the same as every posix, but I want to make sure before I use it on here
[15:53:01] <andyf> Why? probably 'man 2 open' is a Linux thing...
[15:53:10] <KungFuJesus> ahhh, didn't realize that
[15:53:33] <KungFuJesus> do you know if sendfile works with O_LARGEFILE on here?
[15:54:54] <KungFuJesus> sendfile API seems to take off_t's and not off64_t's
[15:55:44] <KungFuJesus> oh, there's a sendfile64. I'll be damned
[15:55:56] <andyf> :) I was just reading the 'lf64' man page
[15:56:15] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has joined #illumos
[15:56:41] <KungFuJesus> seems to be missing a man page, though
[15:57:22] <KungFuJesus> I'm going to be traversing 30-60GB into a file
[15:57:24] <andyf> It looks like there aren't man pages (or links) for any of the transitional interfaces
[15:57:35] <andyf> but 'lf64' lists them
[15:57:49] <andyf> and it has to be a 32-bit binary that does the work?
[15:58:26] *** man_u <man_u!~manu@manu2.gandi.net> has quit IRC (Ping timeout: 240 seconds)
[16:00:47] <KungFuJesus> nope, 64 bit is just fine
[16:01:40] <KungFuJesus> I was kind of hoping there'd be a gather list seque API in POSIX for reads into buffers, but I'm guessing that's mainly what mmap is for?
[16:02:16] <KungFuJesus> or have I found a gaping hole for where there should be a recvfile? heh
[16:03:19] *** Ericson2314 <Ericson2314!ericson231@gateway/shell/matrix.org/x-cunqxwzrefgupvne> has joined #illumos
[16:03:19] *** psydroid <psydroid!psydroidma@gateway/shell/matrix.org/x-jbspqufsfnxuevjd> has joined #illumos
[16:03:19] *** GrahamPerrin[m] <GrahamPerrin[m]!grahamperr@gateway/shell/matrix.org/x-aqlhjgxujmtfvukz> has joined #illumos
[16:03:19] *** jhot[m] <jhot[m]!jhotmatrix@gateway/shell/matrix.org/x-yzpnelniutovphjn> has joined #illumos
[16:04:16] <KungFuJesus> andyf: Just read this from the Linux man page: Consequently, Linux 2.4 added sendfile64(), with a wider type for the offset argument. The glibc sendfile() wrapper function transpar‐ently deals with the kernel differences
[16:04:36] <KungFuJesus> does Illumos do the same and off_t will be a 64 bit type if I compile as a 64 bit binary?
[16:05:21] <jbk> yes, 64-bit binaries will use 64-bit types for I/O
[16:05:41] <jbk> it's only 32-bit binaries that have to do some special stuff at build time to get 64-bit types for I/O
[16:06:03] <jbk> (either define some macros to redefine the types to 64-bit versions, or explicitly use the 64-bit versions)
[16:06:10] <KungFuJesus> oh ok, that clarifies it a bit. Off_t is typedef'd to size_t or ssize_t I'm guessing?
[16:06:45] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has quit IRC (Ping timeout: 252 seconds)
[16:07:16] <jbk> it uses 'long' but more or less
[16:08:20] <jbk> https://github.com/illumos/illumos-gate/blob/master/usr/src/uts/common/sys/types.h#L132-L157
[16:09:06] <KungFuJesus> Cool. This is a sort of hack solution but it may be acceptable. I'm still not entirely sure if I'm looking at a MATLAB bug, NFS perf bug, or something else. Colleague claims that for a 120GB file, 4 million scattered reads (at ~4k fixed interval offsets) into the file are taking like a half hour
[16:10:49] <KungFuJesus> he claims < 1 second for the 60GB file, but it's only 62,500 reads in that circumstance
[16:11:43] <KungFuJesus> for that he's probably fitting into ARC better, and it's about an order of magnitude fewer reads
[16:13:09] <KungFuJesus> I'm going to try to see what the performance is like carving these values out into a flat table
[16:13:44] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has joined #illumos
[16:14:51] *** pwinder <pwinder!~pwinder@86.2.210.254> has quit IRC (Quit: Leaving)
[16:16:00] <KungFuJesus> open() ignores the mode flag?
[16:24:53] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has quit IRC (Ping timeout: 276 seconds)
[16:35:11] *** despair86 <despair86!~despair86@24.170.8.11> has joined #illumos
[16:45:15] *** despair86 <despair86!~despair86@24.170.8.11> has quit IRC (Ping timeout: 250 seconds)
[16:47:07] <KungFuJesus> heh, sendfilev has scatter IO mechanisms but no gather. The output is assumed to be a file rather than buffers
[16:47:24] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[16:47:47] <rmustacc> open() only uses the mode if O_CREAT is specified.
[16:48:22] *** despair86 <despair86!~despair86@24.170.8.11> has joined #illumos
[16:50:24] <KungFuJesus> Yeah. I assume I can specify NULL there and it will use the umask/inheritance of the file system's ACL mode?
[16:51:25] <rmustacc> Are you passing O_CREAT?
[16:52:17] <KungFuJesus> yes
[16:52:53] <KungFuJesus> I want it to follow the parent directory's ACL, which has :fd set
[16:52:54] <rmustacc> NULL is a pointer, not an integer (generally).
[16:53:01] <rmustacc> But you don't want to pass zero per se.
[16:53:03] <KungFuJesus> well yes, though, #define NULL 0 is typical
[16:53:23] <rmustacc> Passing zero is a bad idea still.
[16:53:37] <rmustacc> The system will take a bitwise-and of the umask and the mode.
[16:53:55] <rmustacc> As far as ACLs, I'm not sure about the inheritence.
[16:53:57] <KungFuJesus> how do I make it do typical umask behavior, such as when using touch? The man page is very vague about this
[16:55:44] <rmustacc> What do you mean by typical umask behavior?
[16:56:02] <rmustacc> So, touch just passes 0666 (octal) as its argument to the mode.
[16:56:02] <KungFuJesus> touch filename, if it doesn't exist, is created with the default umask
[16:56:12] <rmustacc> It isn't created with the default umask per se.
[16:56:23] <rmustacc> If you have executable there, it doesn't actually get added.
[16:56:25] <rmustacc> IIRC.
[16:56:55] <rmustacc> My default umask is 0022. If I touch a file, it's not executable.
[16:57:05] <KungFuJesus> oh, the umask is something restrictive that's AND'd into what you pass to open, I'm guess?
[16:57:10] <KungFuJesus> guessing*
[16:57:10] <rmustacc> Yes.
[16:57:15] <rmustacc> That's exactly what it says.
[16:57:17] <KungFuJesus> ok, I see
[16:57:23] <rmustacc> The algorithm is basically mode & ~umask
[16:57:39] *** chrisBF <chrisBF!519d0501@host81-157-5-1.range81-157.btcentralplus.com> has joined #illumos
[16:57:40] <KungFuJesus> sorry, been while since I've used most of these lower level calls. I typically use fopen to get the buffering behavior
[16:57:44] <rmustacc> No worries.
[16:57:51] <rmustacc> No need to apologize.
[16:58:10] <rmustacc> It's not a straightforward thing. Especially since you record the inverse of what you want.
[17:01:48] <KungFuJesus> am I crazy for wanting gather lists into userspace buffers? I feel like maybe the kernel could leverage that a bit better than the sledgehammer of mmap
[17:02:39] <rmustacc> You mean you want to read the contents of a file descriptor and break it into multiple buffers?
[17:02:49] <KungFuJesus> I'm ok with serializing out to a new file for now, but that would have been a nice solution to this problem as well. Though, I somehow doubt this will be appreciably faster than the Matlab fread function that's churning on this
[17:02:53] <rmustacc> And when you are writing, you want to give the kernel multiple disjoint buffers that it uses linearly?
[17:03:08] *** az0r <az0r!~sh42@fsf/member/zeroSignal> has quit IRC (Read error: Connection reset by peer)
[17:03:09] <KungFuJesus> the same buffer. I want disparate offsets into the file on the reads
[17:03:17] *** az0r <az0r!~sh42@178.251.69.48> has joined #illumos
[17:03:20] <KungFuJesus> linear write into the userspace buffer
[17:03:43] <rmustacc> OK. I see. So you want the opposite mode of the preadv/pwritev calls.
[17:03:49] <KungFuJesus> precisely
[17:04:16] <rmustacc> I don't know of a mode of it existing today.
[17:04:26] <KungFuJesus> it'd been so long since I'd use these syscalls, I thought for a moment that preadv _did_ accomplish this for me. Then I noticed it's a linear read from the file
[17:05:20] <rmustacc> All the read/write APIs today that I can think of are linear.
[17:05:34] <rmustacc> In terms of communicating to/from the underlying object.
[17:07:15] <rmustacc> Naively, the kernel would basically have to treat it like multiple disjoint calls internally.
[17:07:58] <rmustacc> The kernel I/O APIs basically are all contiguous read/write from the underlying object to/from multiple buffers.
[17:07:59] *** az0r <az0r!~sh42@178.251.69.48> has quit IRC (Ping timeout: 252 seconds)
[17:08:37] <rmustacc> So, I think your best bet at the moment is to just use multiple calls to something like pread/pwrite.
[17:12:02] *** despair86 <despair86!~despair86@24.170.8.11> has quit IRC (Ping timeout: 240 seconds)
[17:18:43] <KungFuJesus> rmustacc: I'm using sendfilev to carve everything out to a new file as an ahead of time preprocessing step
[17:19:17] <KungFuJesus> I think that's about the best I'm going to do. The gaps between reads are too large for mmap to not be extremely wasteful when things are paged in
[17:19:29] *** az0r <az0r!~sh42@178.251.69.48> has joined #illumos
[17:19:48] <rmustacc> OK, fair enough.
[17:21:38] *** nde_ <nde_!~nde@96-66-145-126-static.hfc.comcastbusiness.net> has joined #illumos
[17:25:39] <Smithx10> Is there anything similar to DRBD for illumos?
[17:26:25] <rmustacc> As in the replicated block device?
[17:27:09] <Smithx10> yea
[17:27:27] <rmustacc> I don't know of anything first party.
[17:27:50] <Smithx10> Bhyve it is :)
[17:28:13] <Smithx10> Luckily, our use case is very low risk simple.... and doesn't really need any performance
[17:29:21] *** man_u <man_u!~manu@manu2.gandi.net> has joined #illumos
[17:31:46] *** psarria <psarria!~psarria@26.red-79-146-96.dynamicip.rima-tde.net> has joined #illumos
[17:31:52] *** merzo <merzo!~merzo@185.39.197.205> has joined #illumos
[17:35:32] *** despair86 <despair86!~despair86@24.170.8.11> has joined #illumos
[17:44:26] *** despair86 <despair86!~despair86@24.170.8.11> has quit IRC (Ping timeout: 240 seconds)
[17:44:34] *** amrfrsh <amrfrsh!~Thunderbi@134.19.189.92> has joined #illumos
[17:51:50] <tsoome> I have the ‘11580 dtrace: cast between incompatible function types’ in rti queue and it was about running dtrace tests with it. is it enough to run something like /opt/SUNWdtrt/bin/dtest ?
[17:51:57] *** despair86 <despair86!~despair86@24.170.8.11> has joined #illumos
[17:52:38] <_mjg> tsoome: i would just compare .o files
[17:52:56] <rmustacc> Yes, you can run dtest like that.
[17:53:03] <_mjg> tsoome: should the change be fine it probably has the same checksum which takes care of it
[17:53:40] <tsoome> aye:)
[17:57:14] *** BH23 <BH23!~BH23@193.117.206.132> has quit IRC (Ping timeout: 240 seconds)
[17:59:05] <gitomat> [illumos-gate] 11729 msgsys: cast between incompatible function types -- Toomas Soome <tsoome at me dot com>
[18:01:33] *** insomnia <insomnia!~insomnia@shadowcat/actuallyamemberof/lollipopguild.insomnia> has quit IRC (Ping timeout: 245 seconds)
[18:02:02] *** despair86 <despair86!~despair86@24.170.8.11> has quit IRC (Ping timeout: 240 seconds)
[18:05:05] *** despair86 <despair86!~despair86@24.170.8.11> has joined #illumos
[18:06:05] <gitomat> [illumos-gate] 11801 i86xpv: cast between incompatible function types -- Toomas Soome <tsoome at me dot com>
[18:14:46] *** insomnia <insomnia!~insomnia@shadowcat/actuallyamemberof/lollipopguild.insomnia> has joined #illumos
[18:20:58] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[18:30:00] <KungFuJesus> rmustacc: hmm, sendfile is erroring out with "invalid argument" after it consumes so many sendfilevec64 entries
[18:30:05] <KungFuJesus> sendfile bug?
[18:31:13] <KungFuJesus> according to the bytes written output argument, it's getting to about the 36557th vec before it errors out
[18:31:51] *** yomisei <yomisei!~void@ip4d16bc15.dynamic.kabel-deutschland.de> has quit IRC (Ping timeout: 250 seconds)
[18:33:47] <KungFuJesus> is there a limit to off_t sizes with the 64 bit sendfile API?
[18:34:56] *** despair86 <despair86!~despair86@24.170.8.11> has quit IRC (Ping timeout: 246 seconds)
[18:35:42] *** man_u <man_u!~manu@manu2.gandi.net> has quit IRC (Quit: man_u)
[18:38:09] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has joined #illumos
[18:40:10] <KungFuJesus> hah, nope, my bad. Bug in my code
[18:44:32] *** yomisei <yomisei!~void@ip4d16bc15.dynamic.kabel-deutschland.de> has joined #illumos
[18:46:53] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has quit IRC (Remote host closed the connection)
[18:47:38] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[18:47:55] <Reinhilde> why wouldn't dhcpagent be ready for ipadm to communicate with it
[18:48:16] *** despair86 <despair86!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has quit IRC (Ping timeout: 240 seconds)
[18:51:27] *** despair86 <despair86!~despair86@24.170.8.11> has joined #illumos
[18:51:47] *** nde_ <nde_!~nde@96-66-145-126-static.hfc.comcastbusiness.net> has quit IRC (Ping timeout: 250 seconds)
[19:22:09] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has quit IRC (Remote host closed the connection)
[19:23:51] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[19:26:37] <wonko> can you run iptables in an lx zone?
[19:28:03] <rmustacc> I don't believe so.
[19:29:31] <tsoome> ok, so, how can I update my patch in gerrit?:)
[19:30:17] <wonko> Ok, not a *huge* deal. I can always setup nginx if I need to. Thanks!
[19:30:35] <Reinhilde> why would you want to
[19:30:50] <rmustacc> wonko: Yeah, either that or you can use the native tooling.
[19:30:56] <wonko> the UniFi Controller doesn't want to listen on 443, just 8443
[19:31:01] <wonko> and there isn't a way to change it
[19:31:22] <wonko> so the "easy" way would just be to add pre-routing nat in iptables to forward 443 to 8443
[19:31:38] *** despair86 <despair86!~despair86@24.170.8.11> has quit IRC (Ping timeout: 240 seconds)
[19:31:59] <wonko> If it was going to be that easy I would have done it with it being on its own IP now
[19:32:05] <wonko> but meh, not really a big deal.
[19:32:06] <rmustacc> tsoome: Assuming you have the updated/squashed bits, you can then go back and you should be able just to push it again, as it has the change-id, IIRC. LeftWing, jlevon is that right for the new era?
[19:36:33] <tsoome> yep, that was simple actually:)
[19:36:39] <rmustacc> OK, great.
[19:36:59] <rmustacc> The only other thing (and we should write all this up) is if you reply to comments, like rb, you have to basically say you're ready to post them.
[19:49:39] <tsoome> some things are still weird there
[19:51:42] *** despair86 <despair86!~despair86@24.170.8.11> has joined #illumos
[20:01:52] *** Kurlon <Kurlon!~Kurlon@cpe-67-253-141-249.rochester.res.rr.com> has quit IRC (Remote host closed the connection)
[20:02:26] <tsoome> I can no add image to https://www.illumos.org/issues/11938 :(
[20:03:43] <rmustacc> It's not letting you attach something or something else?
[20:03:53] <KungFuJesus> question: does posix_fadvise actually get used in Illumos?
[20:04:05] <tsoome> was trying to attach image yep
[20:04:11] <LeftWing> tsoome: What happens?
[20:04:31] <rmustacc> KungFuJesus: I believe fadvise does get honored.
[20:04:39] <tsoome> I have screenshot from person who had issue solved
[20:04:49] <KungFuJesus> I'm not really sure that it's going to save me much, but I guess it's worth trying here
[20:05:04] <LeftWing> tsoome: Yes, I mean, what happens when you try to attach it?
[20:05:06] <KungFuJesus> basically my data is scattered at fixed, 32k offsets in this file
[20:05:07] <LeftWing> And how large is it?
[20:05:14] <KungFuJesus> with only 8 byte reads at each
[20:06:03] <tsoome> LeftWing: https://paste.ec/paste/oHZPi6YV#3Cmvv-GGLoidnarA2oG65uF9/8Wda8Mge7VUi/L+bmj
[20:06:08] <rmustacc> I don't think the fadvise will help that much at that point.
[20:06:17] <tsoome> then I press submit and nothing will appear
[20:06:23] <KungFuJesus> rmustacc: for some reason zpool iostat shows a shitload of bandwidth for this sequential but somewhat scattered read. I'm not entirely sure if that's due to having to parse indirect blocks for this huge file or if prefetching is just doing this due to the sequential search pattern
[20:06:35] <rmustacc> I have no idea.
[20:06:58] <LeftWing> tsoome: What is the size of the file?
[20:07:03] <tsoome> LeftWing: file is 2.3MB
[20:07:06] <LeftWing> huh.
[20:07:08] <LeftWing> OK
[20:07:18] <tsoome> too large?
[20:07:30] <LeftWing> It should be OK, but I'll check!
[20:07:41] <KungFuJesus> I'm getting 30MB of data from this 120GB monster, but disk bandwidth is heavy, between 30MB/sec and 190MB/sec
[20:08:04] <rmustacc> Well, you probably have read amplification.
[20:08:11] <KungFuJesus> rmustacc: do you think a larger record size might help with this?
[20:08:29] <rmustacc> Um, I mean if you have a 128k record size, you'll read 128k even if you only use 8 bytes out of every 32k.
[20:08:30] <Reinhilde> rmustacc: does ipf allow the sort of nat wonko wants
[20:08:51] <rmustacc> Reinhilde: I beieve ipnat (a part of ipf) can.
[20:09:06] <Reinhilde> goodo
[20:09:17] <KungFuJesus> ah, so it's more like a smaller record size might help
[20:09:26] <LeftWing> tsoome: Can you try now?
[20:09:32] <wonko> but can it intercept an exclusive IP?
[20:10:05] <rmustacc> wonko: What do you mean by exclusive ip?
[20:10:07] <tsoome> much better
[20:10:12] <wonko> It's a zone with an exclusive IP
[20:10:25] <rmustacc> Ah, an exclusive-IP netstack.
[20:10:29] <tsoome> ou, but why it is heads up?
[20:10:34] <LeftWing> tsoome: Great! I had added the "client_max_body_size" increase for Reviewboard, but not for Redmine. Apologies!
[20:10:36] <tsoome> err, down...
[20:10:59] <rmustacc> wonko: Yes, every netstack has its own ipf state. So it can operate on it.
[20:11:14] <rmustacc> There are two rulesets, one for the GZ and one for the NGZ to manipulate for a given netstack.
[20:11:21] <tsoome> LeftWing: np, thanks:D
[20:11:21] <rmustacc> So, yes, it can.
[20:11:22] <wonko> oh, that's slick as hell
[20:12:12] <rmustacc> So you could use the ipnat '-G zonename' argument to do this from the GZ for NGZ.
[20:12:15] <rmustacc> *for the NGZ
[20:12:18] <tsoome> if you have time, please check over https://illumos.org/rb/r/2443/ :)
[20:12:37] <wonko> rmustacc: I'll give that a play. That sounds awesome.
[20:12:48] <rmustacc> nginx may still be faster, haha.
[20:12:58] <wonko> I wonder if I can use that to trick another zone to go out a different interface on the host (natted to the host's IP)
[20:13:19] <rmustacc> wonko: So only if it's in that netstack.
[20:13:26] <wonko> rmustacc: I'm not going to bother setting up nginx, because again, meh. Not a huge deal. I use the UniFi controller interface once every never.
[20:13:28] <rmustacc> Otherwise you need to set up a more traditional NAT, etc.
[20:13:44] <rmustacc> Yeah, I understand that use case.
[20:13:48] <wonko> ok, so that probably won't work then. Hmm.
[20:14:13] <rmustacc> If you want to rewrite an address from zone 1 to an address in zone 2, you need to NAT somewhere.
[20:14:45] <rmustacc> Or some L7 lb or similar.
[20:15:03] <wonko> I'm just trying to re-route torrent traffic direct to the host with the transmission zone to keep it off the firewalls. Not that it really matters, it's not like they aren't powerful enough. :)
[20:15:05] <rmustacc> tsoome: Will take a look this afternoon.
[20:15:36] <wonko> now, where was I? :-D
[20:15:52] <tsoome> thanks. it is actually affecting MBP with UEFI 1.10, UEFI 2.3 seems to be fine.
[20:15:53] <wonko> stupid getting sidetracked.
[20:16:08] <rmustacc> Ah, yeah. They have their own UEFI world there.
[20:17:32] <KungFuJesus> rmustacc: hah, seemed to cut the time in half. Hard to know if that was due to something being warmer in ARC or not, though. I probably should have tried to flush that out
[20:17:42] <KungFuJesus> can't hurt, I guess
[20:19:13] <rmustacc> I don't know. But I guess you're doing less I/O that you don't need.
[20:19:27] <rmustacc> At least in theory.
[20:19:53] <KungFuJesus> yeah, I used the FADV_WILLNEED for the starting offsets of each vector for sendfilev
[20:20:07] <rmustacc> That will likely kick off the pre-faulter in the background.
[20:20:50] <KungFuJesus> I fully understand this is a dumb way to do IO, we didn't write the code that writes these files :-/
[20:20:50] <rmustacc> Wait, you're doing posix_fadvise()?
[20:20:53] <KungFuJesus> yes
[20:21:07] <rmustacc> I have good news for you about your advice. ;)
[20:21:16] <KungFuJesus> ?
[20:21:21] <rmustacc> It's completely ignored.
[20:21:29] <KungFuJesus> haha, pure placebo
[20:21:35] <rmustacc> I just double checked the libc bit.
[20:21:40] <KungFuJesus> alright, which fadvise is not ignored?
[20:21:58] <rmustacc> Invalid ones.
[20:21:58] <copec> the one from the wise elderly father
[20:22:06] <rmustacc> There's no kernel hook up.
[20:22:55] <rmustacc> madvise() is wired up. But for fadvise, we take your advice and then ignore it.
[20:23:09] <KungFuJesus> hah, greaaat
[20:23:21] <rmustacc> So probably you have better arc caching or similar, sorry.
[20:23:23] <KungFuJesus> I guess I could mmap, I feel like the inflationary factor there would be just as bad
[20:24:20] <KungFuJesus> mmap + writev, does that sound like a thing that would do something better?
[20:24:32] <rmustacc> I don't know.
[20:24:49] <rmustacc> I haven't really thought too much about how I would deal with your problem.
[20:25:27] <KungFuJesus> I was hoping maybe I could seed the prefetch behavior with a little bit of wisdom for this load :-/
[20:25:58] <rmustacc> I assume you're currently I/O bottlenecked?
[20:26:04] <KungFuJesus> yes
[20:27:01] <KungFuJesus> Takes about ~30 minutes to grab these values. Once I serialize them to a separate index file for reading later, the parsing is miliseconds. I guess that's a fair enough trade off, I was just hoping I could come up with something that writes these index files a little bit faster
[20:27:24] <rmustacc> I might experiment with different record sizes?
[20:27:35] <rmustacc> Dunno. Or possibly compression.
[20:27:46] <rmustacc> Basically different approaches to try and do less actual I/O.
[20:27:47] <KungFuJesus> fs is already lz4 compressed
[20:28:21] <rmustacc> How many outstanding I/Os do you have at any time?
[20:28:21] <KungFuJesus> thing is, the file itself does have bulky, contiguous regions that could benefit from the prefetching of adjacent blocks
[20:28:24] <KungFuJesus> just not for this bit
[20:28:37] <rmustacc> Is this single-threaded doing I/O?
[20:28:42] <KungFuJesus> yes
[20:28:51] <jbk> maybe look into the aio interfaces?
[20:28:54] <rmustacc> So, parallel issue of I/O _could_ help.
[20:29:02] <jbk> (they actually work on illumos)
[20:29:14] <KungFuJesus> the aio interface actually works and doesn't suck?
[20:29:30] <rmustacc> I've never used them. I would just use multiple threads personally.
[20:29:40] <KungFuJesus> yeah, it'd be a bit cleaner I suppose
[20:29:40] <LeftWing> Pretty sure that's what our aio thing does under the hood
[20:29:43] <jbk> that's pretty much what it does :)
[20:29:46] <LeftWing> (Manage a thread pool)
[20:30:16] <jbk> in fact, there was an amusing bug I hit at an old job in s10 due to that (long since fixed obviously)
[20:30:36] <jbk> (it was ignoring the return value from thr_create and always assumed it succeeded)
[20:31:05] <jbk> .. which generally works until you exhaust segkpm
[20:31:09] <KungFuJesus> of course, sendfile expects linear writes, so I'm going to have to just aggregate the reads into a buffer in userspace
[20:31:37] <jbk> there's also the even more esoteric listio interfaces :)
[20:31:42] <KungFuJesus> I doubt it was buying me a whole lot anyway, given the size of these reads
[20:33:38] *** despair86 <despair86!~despair86@24.170.8.11> has quit IRC (Ping timeout: 240 seconds)
[21:19:59] *** chrisBF <chrisBF!519d0501@host81-157-5-1.range81-157.btcentralplus.com> has quit IRC (Remote host closed the connection)
[21:20:00] *** ed209 <ed209!~ed209@165.225.128.67> has quit IRC (Remote host closed the connection)
[21:20:07] *** ed209 <ed209!~ed209@165.225.128.67> has joined #illumos
[21:39:57] *** merzo <merzo!~merzo@185.39.197.205> has quit IRC (Ping timeout: 240 seconds)
[21:49:20] <KungFuJesus> rmustacc: that nvme patch ready for integration yet? I keep getting excited and then a new nit pick is found about a comment :-/
[21:50:28] <rmustacc> I'm not the author of it.
[21:50:55] <rmustacc> It sounds like you're frustrated by my review comments and trying to understand it.
[21:51:25] <KungFuJesus> hah no, I'm just antsy for it
[21:51:56] <KungFuJesus> lower latency in the nvme path is something I've wanted for a while now
[21:52:43] <rmustacc> At the end of the day, when the only issues are typos in comments, that's a good sign.
[21:52:55] <rmustacc> Especially given that this unveiled a couple of race conditions that exist in the existing driver.
[21:53:13] <KungFuJesus> e.g. https://www.illumos.org/issues/9291
[21:53:52] <rmustacc> Others have noted major differences between the ZFS and the NVMe layers, fwiw.
[21:53:55] <KungFuJesus> the blkdev code seems have improved since then, rerunning that, the mode of distribution definitely shifted in the graph
[21:53:59] *** kerberizer <kerberizer!~luchesar@wikipedia/Iliev> has quit IRC (Ping timeout: 252 seconds)
[21:54:19] <rmustacc> And that going directly to the device they saw numbers close to what they expected.
[21:54:49] <rmustacc> But it does seem like Paul's change here should help.
[21:55:07] <KungFuJesus> that dtrace was measuring the direct IO, though, so ZFS wasn't exactly the culprit there
[21:55:32] <KungFuJesus> used the io provider with a device path filter
[21:56:04] <rmustacc> I was just trying to make a general comment on others observations.
[21:56:16] <rmustacc> This isn't something I've looked at, at all.
[21:56:24] <rmustacc> Just making sure they actually work has always been my first order.
[21:57:34] <KungFuJesus> yes, bug free should take precedence over optimization. Though from the looks of it, this optimization found a bug
[21:57:55] <rmustacc> Well, looking over the code I just asked the question of how is this safe.
[21:58:10] <rmustacc> And the answer is certain admin operations aren't.
[21:59:23] <KungFuJesus> yeah I imagine nvmeadm can do some pretty nasty stuff while the device is in use
[21:59:39] <gitomat> [illumos-gate] 11879 Intel ixgbe X553 Support -- Robert Mustacchi <rm at fingolfin dot org>
[21:59:52] <rmustacc> Most of that should be pretty reasonable, actually.
[22:08:43] <KungFuJesus> did the trim support ever get resolved for illumos? I have an rpool here at work on flash that would probably be good to trim but last I heard there was some circumstance that could render the pool unbootable
[22:19:28] <rzezeski> https://www.illumos.org/issues/11714
[22:22:29] <KungFuJesus> well that doesn't make me any less nervous about attempting to trim :-/
[22:29:47] <jbk> there was one zfs issue that was found and fixed.. i suppose if you compared what you have against blacklists on other OSes.. that might reduce the risk somewhat.. though still better to try on a test box first
[23:12:18] *** despair86 <despair86!~despair@2600:1700:8f61:163f:36e8:94ff:fede:7ce2> has joined #illumos
[23:12:50] <despair86> https://snowlight.net/files/bug/core.tar.xz - crash dump in XHCI it appears like
[23:18:15] *** arnoldoree <arnoldoree!~arnoldore@ranoldoree.plus.com> has quit IRC (Quit: Leaving)
[23:31:52] <LeftWing> huh, "panic message: XHCI runtime reset required"
[23:32:26] <rmustacc> Usually that's indicative of another issue in the xhci driver or broader stack.
[23:32:59] <LeftWing> There are actually two dumps in here, too; the other seems to be an issue with a NIC?
[23:33:02] <rmustacc> I saw this in the context of 11953, so I was going to pull it down and try to better understand how this was all happening.
[23:33:06] <despair86> interesting
[23:33:38] <despair86> pcieb dies before ZFS is available so you can't really get a core dump at that point
[23:34:08] <rmustacc> Yes, I was going to look over the ereports and try and confirm which ones were the ones that suggested we needed fatal behavior in the system.
[23:34:16] <rmustacc> And understand what was being generated and why.
[23:36:41] <rmustacc> It's not clear to me how this is all linked to enabling the bandwith monitoring though.
[23:36:46] <rmustacc> And why only this chipset is acting up.
[23:37:49] *** xzilla <xzilla!~robert@pool-71-166-61-131.bltmmd.fios.verizon.net> has joined #illumos
[23:37:55] *** _despair86 <_despair86!~despair@2605:6000:1515:45b:b62e:99ff:fea1:3bad> has joined #illumos
[23:40:18] <rmustacc> despair86: Out of curiosity is there anything in that BIOS about PCIe 'ECRC?
[23:40:43] *** despair86 <despair86!~despair@2600:1700:8f61:163f:36e8:94ff:fede:7ce2> has quit IRC (Ping timeout: 245 seconds)
[23:42:20] <_despair86> nope. i even have the manual up on the screen
[23:42:35] <_despair86> does allow you to cap PCIe speeds tho (1.1, 2.0 3.0 4.0 limits)
[23:42:53] <rmustacc> OK. No worries. I didn't expect it to. Just was curious.
[23:43:11] <rmustacc> This is a PCIe 3.x chipset though right?
[23:43:26] <rmustacc> (Just to make sure I have it right)
[23:45:20] <_despair86> PCIe 4.0
[23:45:25] <_despair86> https://cdn.mos.cms.futurecdn.net/LkEwz9RDqgnaGckH5xWKVZ-650-80.png
[23:47:58] *** _despair86 <_despair86!~despair@2605:6000:1515:45b:b62e:99ff:fea1:3bad> has quit IRC (Remote host closed the connection)
[23:48:18] *** despair86 <despair86!~despair@2605:6000:1515:45b:b62e:99ff:fea1:3bad> has joined #illumos
[23:48:40] <rmustacc> Is it easy for you to build a kernel with a diff?
[23:49:24] <despair86> should be now that i'm in. i'll have to check my build tools at least.
[23:56:06] <zsj> in my case, the asus prime b450m-k only claims pcie 3.0 support, and that's what my radeon rx 570 also claims to be using
[23:56:34] <rmustacc> I believe 4.x support is opitonal in the B450 chipset.
[23:59:49] <rmustacc> I'll summarize this on the ticket, but in all cases we're getting an ECRC errors.
[23:59:57] <rmustacc> They are non-fatal errors, but uncorrectable.
top

   November 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 | >