Switch to DuckDuckGo Search
   January 15, 2020  
< | 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 | 31 | >

Toggle Join/Part | bottom
[00:06:11] *** AllanJude <AllanJude!~allan@allanjude.bsdbasement.com> has joined #illumos
[00:17:48] <neirac> mgerdts what's for linux-live repo? it has some relation with triton?
[00:18:46] <mgerdts> I'm working on creating a Linux compute node. If you look in the linuxcn branch, you will find a way to build something that is in the first steps of working like SmartOS.
[00:19:40] <mgerdts> As the various agents are ported, a headnode (running SmartOS for now at least) will be able to use it like it currently uses SmartOS compute nodes.
[00:19:45] <neirac> mgerdts oh, which is the url I don't seem to find it
[00:19:52] <mgerdts> The intent is to give a 100% compatible linux container solution.
[00:20:13] <mgerdts> https://github.com/joyent/linux-live/tree/linuxcn
[00:20:57] <neirac> mgerdts thanks!
[00:21:42] <mgerdts> Once we get a little further down the road, a Linux CN should be able to run lx images. We'll likely make it so that triton-docker works with it too.
[00:22:06] <neirac> mgerdts is there vmadm also in there ?
[00:22:26] <neirac> mgerdts have you checked out firecracker ?
[00:22:54] <mgerdts> There will be a different vmadm in there. Rather than the approach that SmartOS has, it will be a thin shim that calls the local cn-agent
[00:23:42] <mgerdts> Yeah, I'm aware of it and work that came before it.
[00:24:25] <jbk> so we've expanded the scope?
[00:25:07] <mgerdts> Very interesting stuff - I'd like to have a way to use something like firecracker on SmartOS.
[00:25:41] <mgerdts> Also, looking at the possibility of virtual linux CNs running in bhyve, giving an experience similar to Kata containers offer.
[00:25:52] <mgerdts> jbk - in what way?
[00:27:27] * mgerdts running away for a while
[00:27:41] <mgerdts> nothing to do with the conversation. :)
[00:27:46] <jbk> i guess what do you mean by 100% compatible?
[00:36:00] *** jellydonut <jellydonut!~quassel@s91904423.blix.com> has quit IRC (Quit: jellydonut)
[00:38:07] *** jellydonut <jellydonut!~quassel@s91904425.blix.com> has joined #illumos
[00:39:37] <neirac> mgerdts maybe bhyve could trimmed?
[00:40:20] <LeftWing> I reckon it definitely could be
[00:42:50] <neirac> mgerdts why a linux compute node? for performance and gpu related workloads?
[00:45:22] <LeftWing> Presumably for better Linux binary compatibility without the engineering overhead of working on LX
[00:56:31] *** jenelizabeth <jenelizabeth!~jenelizab@cpc155793-brmb11-2-0-cust474.1-3.cable.virginm.net> has quit IRC (Ping timeout: 268 seconds)
[01:04:28] *** jenelizabeth <jenelizabeth!~jenelizab@cpc155793-brmb11-2-0-cust474.1-3.cable.virginm.net> has joined #illumos
[01:07:18] <KungFuJesus> should I be observing zfs_write operations with a recordsize of 1M?
[01:08:57] <KungFuJesus> I wouldn't think I'd be seeing that at the VOP level, but maybe something fancy has happened with mv
[01:09:32] <jbk> are you moving across datasets?
[01:09:35] <KungFuJesus> yes
[01:09:39] *** jenelizabeth <jenelizabeth!~jenelizab@cpc155793-brmb11-2-0-cust474.1-3.cable.virginm.net> has quit IRC (Ping timeout: 268 seconds)
[01:10:11] <KungFuJesus> namely because my sequential reads seem to be terrible unless I use a max recordsize of 1M
[01:10:21] <jbk> i'd look at vfs stats
[01:10:30] <jbk> IIRC, under certain circumstances, it might use mmap
[01:10:36] <jbk> instead of read/write
[01:11:02] <jbk> i'd have to look at the code to get more specific
[01:11:02] <KungFuJesus> ahhh, I think I remember seeing some evidence of that with the proposed sendfile cp patch
[01:12:15] <KungFuJesus> I am most certainly causing some semi-intentional read amplification for these files as their intended workload is purely sequential. I don't know why zfetch wasn't able to do this with 128k records, though
[01:15:27] <KungFuJesus> jbk: sorry, meant to say 8MB zfs_write operations - not sure how that got omitted
[01:24:14] <KungFuJesus> here's the write syscall's src: https://github.com/illumos/illumos-gate/blob/4e0c5eff9af325c80994e9527b7cb8b3a1ffd1d4/usr/src/uts/common/syscall/rw.c#L211
[01:26:43] <neirac> LeftWing oh ok, that's right.
[01:30:03] <KungFuJesus> jbk: here's the source in question: https://github.com/illumos/illumos-gate/blob/master/usr/src/lib/libcmdutils/common/writefile.c
[01:48:33] <despair_> alanc: i mean, OSF upstream CDE is getting patched timely - 6b32246d06ab16fd7897dc344db69d0957f3ae08
[01:51:45] <despair_> (and some of the bugs only affect oracle CDE and not OSF CDE)
[01:54:20] <despair_> (as i look at my local copy, CVE-2020-2696 patch checked in while i was afk)
[01:54:25] * despair_ shrugs
[01:55:29] *** thebug <thebug!~thebug@unaffiliated/thebug> has quit IRC (Quit: leaving)
[01:58:19] *** despair_ <despair_!~despair@24.170.8.11> has quit IRC (Quit: Leaving)
[01:58:27] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has quit IRC (Remote host closed the connection)
[01:58:44] *** despair86 <despair86!~despair@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has joined #illumos
[02:00:53] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[02:11:06] *** jenelizabeth <jenelizabeth!~jenelizab@cpc155793-brmb11-2-0-cust474.1-3.cable.virginm.net> has joined #illumos
[02:45:43] *** ChrisJames02170 <ChrisJames02170!~py@169.48.236.23.bc.googleusercontent.com> has quit IRC (Remote host closed the connection)
[02:58:10] *** ChrisJames02170 <ChrisJames02170!~py@169.48.236.23.bc.googleusercontent.com> has joined #illumos
[03:01:16] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has quit IRC (Remote host closed the connection)
[03:30:03] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has quit IRC (Quit: jcea)
[06:34:41] *** AllanJude <AllanJude!~allan@freebsd/developer/AllanJude> has quit IRC (Remote host closed the connection)
[06:43:13] *** AllanJude <AllanJude!~allan@allanjude.bsdbasement.com> has joined #illumos
[07:23:20] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Quit: tsoome)
[07:25:53] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[07:31:37] *** leoric_ <leoric_!~leoric@2a02:2698:602b:2f73:553b:d649:a64:d64f> has joined #illumos
[07:31:47] <leoric_> Please, review https://illumos.org/rb/r/2489/
[07:39:09] *** BOKALDO <BOKALDO!~BOKALDO@81.198.23.132> has joined #illumos
[07:49:12] <tsoome> jimklimov: I got typewriter
[07:50:25] <tsoome> jimklimov: had to disable serial port, then it disappeared.
[07:54:58] *** leoric_ <leoric_!~leoric@2a02:2698:602b:2f73:553b:d649:a64:d64f> has quit IRC (Quit: Konversation terminated!)
[07:59:00] *** bahamas10 <bahamas10!~dave@cpe-72-231-182-75.nycap.res.rr.com> has quit IRC (Quit: leaving)
[08:02:56] <mnowak_> tsoome, how does one disable it?
[08:03:31] <tsoome> vm settings -> ports untick serial port
[08:04:52] <mnowak_> thanks, I'll test that
[08:06:15] <tsoome> or make sure you have serial port backend (raw file, pipe etc) connected. apparently the vbox serial buffering does not cope well with overflow.
[08:06:26] <tsoome> or just missing output
[08:10:40] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Quit: tsoome)
[08:20:45] *** bahamas10 <bahamas10!~dave@cpe-72-231-182-75.nycap.res.rr.com> has joined #illumos
[08:23:56] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has quit IRC (Quit: Leaving)
[08:24:37] *** bahamas10 <bahamas10!~dave@cpe-72-231-182-75.nycap.res.rr.com> has quit IRC (Client Quit)
[08:26:04] *** bahamas10 <bahamas10!~dave@cpe-72-231-182-75.nycap.res.rr.com> has joined #illumos
[08:26:59] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has joined #illumos
[08:29:27] <mnowak_> hmm, well I lost bold text and text colors in the loader: old (https://paste.ec/paste/HDZbJGxO#NJPq1RgAqJcNhf91giY1f+09CYtzFJZWjuso-01W/wl) v. new (https://paste.ec/paste/ql0yZOb7#gTkoEBhKnPpHDaxfIFLJLvO1hAzR4l+0U4+wsRUj7L4)
[08:30:07] <mnowak_> this is VirtualBox but my Lenovo is the same. both show 'text' for OsConsole
[08:31:20] *** wiedi <wiedi!~wiedi@ip5b4096a6.dynamic.kabel-deutschland.de> has quit IRC (Quit: ^C)
[08:32:29] *** bahamas10 <bahamas10!~dave@cpe-72-231-182-75.nycap.res.rr.com> has quit IRC (Quit: leaving)
[08:32:48] *** bahamas10 <bahamas10!~dave@cpe-72-231-182-75.nycap.res.rr.com> has joined #illumos
[08:44:01] *** zsj <zsj!~zsj@3EC95F11.catv.pool.telekom.hu> has quit IRC (Quit: leaving)
[08:45:03] <sjorge> mgerdts: have you considered virtio-p9 with booting just a Linux kernel of a dataset?
[08:45:18] <sjorge> Seem FreeBSD is looking into that once the p9 stuff lands in bhyve
[08:45:50] <sjorge> So should be pretty lightweight
[08:47:56] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[09:07:34] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has joined #illumos
[09:11:26] *** neuroserve <neuroserve!~toens@195.71.113.124> has joined #illumos
[09:18:49] *** joltman <joltman!~znc@unaffiliated/joltman> has quit IRC (Ping timeout: 265 seconds)
[09:21:33] *** wiedi <wiedi!~wiedi@185.85.220.177> has joined #illumos
[09:23:32] *** joltman <joltman!~znc@unaffiliated/joltman> has joined #illumos
[10:00:15] <jimklimov> tsoome: thanks... there is a serial port (that worked for ages) for this VM as a repairs backdoor, should be listening on a TCP port
[10:00:50] <jimklimov> and the "connect to existing pipe/socket" is not checked, so it shoudl instantiate its own anew
[10:01:20] <jimklimov> hope more and more this is a host Windows issue... its audio disappeared too, makes calls complicated :\
[10:17:10] *** man_u <man_u!~manu@manu2.gandi.net> has joined #illumos
[10:36:36] <leoric> andyf: Hi. What is the status of #11764 ?
[10:41:02] <andyf> leoric: it's kind of stalled. I got some feedback from jlevon I need to look at. There's a conversation that has to happen around what the defaults in Makefile.master should reflect, and the same for the default illumos.sh env file
[10:41:27] <andyf> and re 11522, I'll have a look today and test it on OmniOS too
[11:02:35] *** tru_tru <tru_tru!~tru@157.99.90.140> has quit IRC (Ping timeout: 246 seconds)
[11:02:52] *** leoric <leoric!~alp@pyhalov.cc.rsu.ru> has quit IRC (Remote host closed the connection)
[11:03:25] *** tru_tru <tru_tru!~tru@157.99.90.140> has joined #illumos
[11:05:20] *** leoric <leoric!~alp@pyhalov.cc.rsu.ru> has joined #illumos
[11:05:21] *** alp_ <alp_!~alp@pyhalov.cc.rsu.ru> has joined #illumos
[11:05:28] *** alp_ <alp_!~alp@pyhalov.cc.rsu.ru> has quit IRC (Remote host closed the connection)
[11:27:59] *** zsj <zsj!~zsj@3EC95F11.catv.pool.telekom.hu> has joined #illumos
[11:42:46] *** jimklimov <jimklimov!~jimklimov@ip-86-49-254-26.net.upcbroadband.cz> has quit IRC (Read error: Connection reset by peer)
[11:42:53] <gitomat> [illumos-gate] 11967 need TAA mitigation -- John Levon <john.levon at joyent dot com>
[11:43:09] <jlevon> finally!
[11:53:08] *** BOKALDO <BOKALDO!~BOKALDO@81.198.23.132> has quit IRC (Quit: Leaving)
[11:54:26] *** jimklimov <jimklimov!~jimklimov@ip-86-49-254-26.net.upcbroadband.cz> has joined #illumos
[11:58:11] <jimklimov> tsoome: thanks for the hint, without the serial port enabled, the VM is no longer in typewrite mode ;)
[11:58:26] <tsoome> :)
[12:23:04] <andyf> jlevon, nice, thanks!
[13:10:51] *** superjen96 <superjen96!~jenelizab@cpc155793-brmb11-2-0-cust474.1-3.cable.virginm.net> has joined #illumos
[13:13:00] *** jenelizabeth <jenelizabeth!~jenelizab@cpc155793-brmb11-2-0-cust474.1-3.cable.virginm.net> has quit IRC (Ping timeout: 268 seconds)
[13:26:46] *** Kurlon <Kurlon!~Kurlon@cpe-67-253-141-249.rochester.res.rr.com> has quit IRC (Ping timeout: 258 seconds)
[13:39:24] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Quit: tsoome)
[13:43:42] *** BOKALDO <BOKALDO!~BOKALDO@81.198.23.132> has joined #illumos
[13:59:26] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has joined #illumos
[14:11:41] <mgerdts> jbk 100% compatible in that it will be Linux user space running on an actual Linux kernel. The lx brand is great, but has well-known limitations that won't be addressed any time soon (i.e. namespaces, cgroups)
[14:13:39] <mgerdts> sjorge - virtio-9p is something "we" (esp. papertigers) are looking at.
[14:16:54] <mgerdts> meirac - Linux CNs (on metal or perhaps virtualized in bhyve) for multiple reasons. A key one is for less pain when the next version of some software you get via "
[14:17:34] <mgerdts> "apt update" decides to use new features that aren't implemented properly in lx on SmartOS.
[14:18:48] <mgerdts> There are some customers and potential customers that are just more comfortable with Linux. They lack SmartOS/illumos/Solaris skills and having such a dependence on SmartOS impedes sales.
[14:19:49] <mgerdts> GPU or other forms of PCI passthrough are not one of the motivations. That can already be done with bhyve on SmartOS. We need to add some plumbing to Triton to make it usable when your SmartOS CN is part of a Triton datacenter.
[14:40:26] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has quit IRC (Ping timeout: 240 seconds)
[14:49:05] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has joined #illumos
[14:51:35] <neirac> mgerdts oh ok, that understandable thanks for the explanation.
[15:02:08] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has joined #illumos
[15:04:45] *** thebug <thebug!~thebug@unaffiliated/thebug> has joined #illumos
[15:04:48] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has quit IRC (Remote host closed the connection)
[15:05:18] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has joined #illumos
[15:13:14] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has joined #illumos
[15:20:27] *** gjnoonan <gjnoonan!sid95422@gateway/web/irccloud.com/x-rkvvwcrsedktgozb> has quit IRC ()
[15:20:50] *** gjnoonan <gjnoonan!sid95422@gateway/web/irccloud.com/x-bhzzpmoklvdkabff> has joined #illumos
[15:24:40] <neuroserve> mgerdts : although I guess, that the age of Linux distributions as we know them is coming to an end
[15:29:14] *** despair86 <despair86!~despair@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has quit IRC (Remote host closed the connection)
[15:34:02] *** superjen96 <superjen96!~jenelizab@cpc155793-brmb11-2-0-cust474.1-3.cable.virginm.net> has quit IRC (Ping timeout: 240 seconds)
[15:48:39] <Smithx10> neuroserve: say it....
[15:48:43] <Smithx10> i dare you :P
[15:48:47] <Smithx10> KUBERNETES!$!$!$! hahahaha
[15:51:35] <neuroserve> Smithx10 : I wanted to hint at my feeling, that dealing with package managers, repositories and updates is something people don't want to be bothered with
[15:53:39] <Smithx10> One day all of the worlds orchestration and automation will be solved by a package manager :P
[15:54:26] *** jenelizabeth <jenelizabeth!~jenelizab@cpc155793-brmb11-2-0-cust474.1-3.cable.virginm.net> has joined #illumos
[16:07:24] <neuroserve> Smithx10 : works for me - as long as I don
[16:07:35] <neuroserve> 't have to deal with it :)
[16:17:02] <KungFuJesus> is it likely that this patch https://github.com/illumos/illumos-gate/commit/d59679dc4ee5ea26c61e7762a3f7a6f74a1f4c2c is undermining this fix https://github.com/illumos/illumos-gate/commit/882fdc88a3017c111dceac12ee7c111b7c4f1232 ?
[16:17:18] <KungFuJesus> because after my latest be, it's hung at the reset again
[16:25:10] <jbk> i saw a tweet that kubernetes is just VCS for millennials
[16:25:37] <KungFuJesus> hah
[16:27:39] <KungFuJesus> can someone reopen this issue? https://www.illumos.org/issues/11767 it seems to have been made worse for me
[16:27:59] *** Wraithh <Wraithh!ilkka@kapsi.fi> has joined #illumos
[16:37:37] <KungFuJesus> ugh, even in the old be it's hanging on "Configuring Devices" now
[16:37:45] <KungFuJesus> SAS controllers are just the worst
[16:42:03] *** neuroserve <neuroserve!~toens@195.71.113.124> has quit IRC (Ping timeout: 268 seconds)
[16:46:15] *** superjen96 <superjen96!jenelizabe@gateway/vpn/privateinternetaccess/jenelizabeth> has joined #illumos
[16:50:03] *** jenelizabeth <jenelizabeth!~jenelizab@cpc155793-brmb11-2-0-cust474.1-3.cable.virginm.net> has quit IRC (Ping timeout: 260 seconds)
[17:08:28] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[17:34:54] *** Teknix <Teknix!~pds@69.41.134.110> has quit IRC (Ping timeout: 268 seconds)
[17:37:50] *** Teknix <Teknix!~pds@69.41.134.110> has joined #illumos
[17:44:39] *** Teknix <Teknix!~pds@69.41.134.110> has quit IRC (Ping timeout: 272 seconds)
[17:46:59] *** superjen96 <superjen96!jenelizabe@gateway/vpn/privateinternetaccess/jenelizabeth> has quit IRC (Quit: Leaving...)
[17:47:15] *** jenelizabeth <jenelizabeth!jenelizabe@gateway/vpn/privateinternetaccess/jenelizabeth> has joined #illumos
[17:47:17] *** Teknix <Teknix!~pds@69.41.134.110> has joined #illumos
[17:55:18] *** chandlore____ <chandlore____!sid259138@gateway/web/irccloud.com/x-ygrfpawfbebryuvy> has quit IRC ()
[17:55:31] *** chandlore____ <chandlore____!sid259138@gateway/web/irccloud.com/x-plpfgcinvhvppvxx> has joined #illumos
[18:12:48] *** wiedi <wiedi!~wiedi@185.85.220.177> has quit IRC (Quit: ^C)
[18:15:16] *** man_u <man_u!~manu@manu2.gandi.net> has quit IRC (Quit: man_u)
[18:17:14] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[18:23:05] *** Yogurt_ <Yogurt_!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[18:23:05] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has quit IRC (Read error: Connection reset by peer)
[18:24:40] <sjorge> mgerdts one it v9p lands i was going to try to use coreboot/linux boot to boot am inimal kernel from a dataset and see if I could get a minimal thing going
[18:25:04] <sjorge> Glad to hear there is also interest in v9p from Joyent, atleast once it lands upstream it might get pulled in quick
[18:27:28] <mgerdts> For experimenting with that, it may be helpful to hack up the joyent brand to allow the devices that are present in bhyve zones. (diff /usr/lib/brand/{joyent,bhyve}/platform.xml)
[18:28:18] <mgerdts> Then you would be able to start bhyve in the zone with whatever options you need and can easily fiddle with the contents of VM without having to be in the global zone.
[18:43:20] <sjorge> not a bad idea
[18:46:32] *** isaacdavis <isaacdavis!68f41bf5@104-244-27-245.PUBLIC.monkeybrains.net> has joined #illumos
[18:47:07] *** isaacdavis <isaacdavis!68f41bf5@104-244-27-245.PUBLIC.monkeybrains.net> has quit IRC (Remote host closed the connection)
[18:47:54] *** KungFuJesus <KungFuJesus!~adamstyli@207.250.97.74> has quit IRC (Remote host closed the connection)
[18:48:30] *** KungFuJesus <KungFuJesus!~adamstyli@207.250.97.74> has joined #illumos
[18:48:36] <KungFuJesus> tsoome: how do I blacklist a kernel module from the loader?
[18:48:38] <sjorge> And if that works, try and boot a docker dataset
[18:48:48] <KungFuJesus> I believe the IPMI module is preventing me from booting, strangely
[18:49:10] <tsoome> set disable_modulename=true
[18:49:39] <KungFuJesus> ok, do you know what the ipmi module's name is? And can I do it at any of the loader prompts?
[18:50:08] <tsoome> yes, ort boot -B disable_modulename=true
[18:50:51] <tsoome> not sure about ipmi rriver
[18:51:00] <tsoome> driver*
[18:53:49] <papertigers> sjorge: maybe we will get an update on v9p on tomorrows bhyve call. I don't know how close they are to landing it
[19:03:29] <KungFuJesus> tsoome: is there a way for me to mount a be from another to remove a package?
[19:03:53] <tsoome> beadm mount bename /mnt
[19:04:19] <KungFuJesus> ok, and how do I make pkg remove something from that environment? Do I just need to chroot into it?
[19:04:33] <tsoome> pkg -R /mnt uninstall packagename
[19:05:51] <KungFuJesus> very few things are different between these BE's so it's either that or something in /etc/system it doesn't like, but I'm pretty sure the options I have in there are fairly safe
[19:11:00] *** amrfrsh <amrfrsh!~Thunderbi@134.19.189.92> has quit IRC (Quit: amrfrsh)
[19:14:22] <KungFuJesus> hmm, it had to update the boot archive and reboot and then it failed
[19:14:32] <KungFuJesus> can you think of a reason why the boot process would hang on "configuring devices"?
[19:14:51] <KungFuJesus> the previous be won't even boot for some reason, but the one prior to this one does
[19:15:52] *** AllanJude <AllanJude!~allan@freebsd/developer/AllanJude> has quit IRC (Remote host closed the connection)
[19:16:16] <tsoome> configuring devices means all drivers are loaded (except disabled ones) and driver attach method is called.
[19:16:34] <tsoome> but what may help is boot -v
[19:16:44] <KungFuJesus> this is really weird. I mean I removed the SAS controllers even, there should be nothing hanging this process
[19:16:48] <KungFuJesus> ok, I'll try that
[19:16:54] <tsoome> but you probably know that alreadyD:
[19:17:07] <KungFuJesus> I forgot that illumos zealously hides kernel messages
[19:17:16] <KungFuJesus> this of course is going to be fairly painful with the slow fb device
[19:17:49] <tsoome> bios boot or uefi?
[19:18:07] <KungFuJesus> bios
[19:19:10] <tsoome> press space on first spinner and type -t on boot: prompt, or reduce color depth — like framebuffer set XxYx8 for example
[19:20:17] <KungFuJesus> missed the first one. Can I do boot -v -t on the second?
[19:20:58] <tsoome> just enter framebuffer off and then boot -v
[19:22:01] <KungFuJesus> ahh too late, oh well
[19:22:19] <KungFuJesus> lol, dual socket system, too
[19:23:38] <jbk> are there fmd events?
[19:23:46] *** AllanJude <AllanJude!~allan@freebsd/developer/AllanJude> has joined #illumos
[19:23:47] <jbk> or were there i should say
[19:24:25] <KungFuJesus> yes, the mpt_sas controller "took too long" and had offlined two drives that I could subsequently re-offline. I've since disconnected those controllers though
[19:24:28] <KungFuJesus> so it shouldn't be hanging on that
[19:24:33] <KungFuJesus> and 2 be's ago it's able to boot
[19:24:46] <KungFuJesus> reonlined*
[19:24:49] <jbk> try boot -s and acquit all the events in the fmd log
[19:25:29] <KungFuJesus> ok - I mean fmadm faulty came up empty prior to the reboot but I suppose it's possible that something else fishy is happening
[19:26:11] <KungFuJesus> if this is what it is - that seems to be a pretty serious flaw in fmd
[19:27:14] *** igork <igork!~igork@static.146.8.9.5.clients.your-server.de> has quit IRC (Ping timeout: 240 seconds)
[19:27:16] *** wiedi <wiedi!~wiedi@ip5b4096a6.dynamic.kabel-deutschland.de> has joined #illumos
[19:27:23] *** igork <igork!~igork@static.146.8.9.5.clients.your-server.de> has joined #illumos
[19:28:40] <KungFuJesus> alright, booting -s and -v
[19:28:48] <jbk> it's not fmd itself
[19:28:54] <jbk> but it could be triggering a zfs bug
[19:29:13] <jlevon> anyone else kind of hate ASSERT(0) ? can we have DEBUG_PANIC("msg") or something
[19:29:52] <jbk> in userland?
[19:29:56] *** jkni <jkni!sid85151@gateway/web/irccloud.com/x-retfqzfleschcwli> has quit IRC (Ping timeout: 248 seconds)
[19:30:03] *** _jack_ <_jack_!sid396411@gateway/web/irccloud.com/x-tqqmjnfmacrugyea> has quit IRC (Ping timeout: 250 seconds)
[19:30:26] *** rann <rann!sid175221@gateway/web/irccloud.com/x-qitibshnrnwgqpru> has quit IRC (Ping timeout: 240 seconds)
[19:30:45] *** scoobybejesus <scoobybejesus!sid271506@gateway/web/irccloud.com/x-jterntugazixpjlz> has quit IRC (Read error: Connection reset by peer)
[19:31:10] *** _jack_ <_jack_!sid396411@gateway/web/irccloud.com/x-karcdmjamfsxmtkc> has joined #illumos
[19:31:16] <jlevon> jbk: kernel side
[19:31:20] <KungFuJesus> it's just hanging here https://i.imgur.com/3Q9830t.png
[19:31:22] *** rzezeski <rzezeski!uid151901@gateway/web/irccloud.com/x-jxvzzzdoinfknegc> has quit IRC (Ping timeout: 260 seconds)
[19:31:32] <KungFuJesus> oh wait there it goes
[19:31:41] *** chandlore____ <chandlore____!sid259138@gateway/web/irccloud.com/x-plpfgcinvhvppvxx> has quit IRC (Ping timeout: 272 seconds)
[19:31:59] *** rann <rann!sid175221@gateway/web/irccloud.com/x-lclwfdwpaldihuln> has joined #illumos
[19:32:43] <KungFuJesus> ok, have single user mode root, jbk what should I do now?
[19:33:34] <jbk> does fmadm faulty show anything?
[19:34:00] <KungFuJesus> seems fmd didn't launch from smf, I'm getting a failed to connect RPC message
[19:35:05] *** chandlore____ <chandlore____!sid259138@gateway/web/irccloud.com/x-vesreclkevybavmx> has joined #illumos
[19:35:21] <jbk> hrm.. what distro?
[19:35:32] <KungFuJesus> OpenIndiana hipster
[19:35:44] <KungFuJesus> if it's going to show anything it's going to be with fmadm faulty -a
[19:35:45] *** rzezeski <rzezeski!uid151901@gateway/web/irccloud.com/x-npfxqxyxhtzfewlp> has joined #illumos
[19:36:33] <KungFuJesus> should I svcadm enable fmd?
[19:36:40] <jbk> yeah..
[19:37:34] <KungFuJesus> so fmadm faulty shows nothing but -a does
[19:37:45] <KungFuJesus> do I just acquit these things one by one or is there an aqcuit all?
[19:37:48] <KungFuJesus> acquit*
[19:38:15] *** amrfrsh <amrfrsh!~Thunderbi@134.19.189.92> has joined #illumos
[19:38:22] <jbk> i think if you don't specify a uuid, it does all of them
[19:38:31] <jbk> (at least according to the man page)
[19:39:37] *** chandlore____ <chandlore____!sid259138@gateway/web/irccloud.com/x-vesreclkevybavmx> has quit IRC (Ping timeout: 265 seconds)
[19:39:37] <KungFuJesus> didn't seem to take that
[19:39:50] <KungFuJesus> looks like fmri or uuid is required
[19:42:19] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has quit IRC (Read error: Connection reset by peer)
[19:42:25] *** mnowak__ <mnowak__!~mnowak_@94.142.238.232> has joined #illumos
[19:42:41] <KungFuJesus> jbk: ok I acquited both FMRIs but they still show up fmadm faulty -a
[19:42:43] <KungFuJesus> is that an issue?
[19:42:52] *** chandlore____ <chandlore____!sid259138@gateway/web/irccloud.com/x-wfsckeijkrfsebmx> has joined #illumos
[19:43:11] <jbk> if it's the issue i'm thinking of that you're hitting, that should be enough
[19:43:22] <KungFuJesus> ok, going for a reboot
[19:43:36] <jbk> should be enough to let it boot the rest of the way from there
[19:43:54] <KungFuJesus> in the future this can be fixed by acquit defaults prior to reboot until the bug is finally fixed?
[19:44:09] <KungFuJesus> acquiting the faults*
[19:44:15] <KungFuJesus> sorry, little sleep last night
[19:45:47] <jbk> possibly.. i'm still trying to confirm if it's a different manifestation of an issue, or something else that is being exaggerated by it
[19:50:28] <KungFuJesus> jbk: tsoome: this seems like it may be pertinent - https://i.imgur.com/Pml6xQv.png
[19:50:51] <jbk> that seems strange
[19:50:54] <tsoome> bad bootpath variable
[19:51:17] <tsoome> see the message about unused variable which is obviously cut
[19:51:17] <KungFuJesus> tsoome: would would cause that and how can I fix it?
[19:51:41] <tsoome> see if you have /boot/transient.conf file
[19:51:43] <KungFuJesus> seems like something maybe screwed up making the boot archive?
[19:52:10] <KungFuJesus> tsoome: how should I access that? boot the 2 be's ago and mount or boot single user mode of the existing be?
[19:52:51] <tsoome> that or ls /boot from loader:D
[19:53:02] <KungFuJesus> oh, that might be better, I'll try that
[19:53:53] <tsoome> but to clear it, you need to remove the file
[19:55:15] <KungFuJesus> tsoome: I do have a transient.conf file according to the loader prompt
[19:55:19] <KungFuJesus> so, just rm it?
[19:55:38] <KungFuJesus> from single user environment, that is
[19:56:57] <tsoome> move aside perhaps, but I think it is misquoting
[19:57:11] <tsoome> and probably fixed
[19:57:22] <KungFuJesus> wow it's taking it's sweet time to get to the login prompt from single user mode
[19:57:36] <tsoome> I kind of recall something about ...
[19:58:09] <jbk> i've noticed that as well with recent(ish) builds
[19:58:20] <jbk> but haven't had a chance to track it down
[19:58:57] <tsoome> 11782 reboot -f ends up creating bad /boot/transient.conf
[19:59:16] <KungFuJesus> ok, BMC's KVM is being a bit shitty, let me walk in there
[19:59:39] *** mnowak__ <mnowak__!~mnowak_@94.142.238.232> has quit IRC (Quit: Leaving)
[19:59:45] <tsoome> update often:)
[20:01:12] <KungFuJesus> ok, let's see if it boots multi user...
[20:01:59] <KungFuJesus> I'll give it a -v just in case we can find _something_ useful
[20:02:40] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has joined #illumos
[20:05:03] <KungFuJesus> ok, looks as though it came up, but after an annoyingly long time
[20:05:20] <KungFuJesus> I guess I'll readd the SAS controllers
[20:23:30] <KungFuJesus> hmm, hanging more or less right now
[20:28:53] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has quit IRC (Remote host closed the connection)
[20:32:10] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has joined #illumos
[20:38:05] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has quit IRC (Remote host closed the connection)
[20:42:20] <KungFuJesus> well finally have the machine booted - you've basically guaranteed I'll never reboot again though, hah
[20:43:22] <KungFuJesus> jbk: I look forward to more info for the possible root cause here, the main coincidence is that fmd found "faulted" drives (which by the way look fine, they just were super busy, so I suspect it was a timeout thing on the multipath driver)
[20:43:58] *** Yogurt_ <Yogurt_!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has quit IRC (Remote host closed the connection)
[20:44:34] <KungFuJesus> tsoome: I do update often is what's funny - very often
[20:44:40] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[20:44:41] <KungFuJesus> like twice a month
[20:45:41] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has joined #illumos
[20:47:37] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has quit IRC (Remote host closed the connection)
[20:47:43] *** igork <igork!~igork@static.146.8.9.5.clients.your-server.de> has quit IRC (Remote host closed the connection)
[20:47:44] *** igork1 <igork1!~igork@static.146.8.9.5.clients.your-server.de> has joined #illumos
[20:48:17] <KungFuJesus> what had prompted me to update/reboot was mostly the update to mpt_sas about bad drives causing multiple drives to fall off of the controller. I thought "hey, that's suspect, I should pull in that update"
[20:49:22] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has joined #illumos
[20:49:36] <jbk> did it come up then?
[20:49:45] <jbk> (sorry on a call, so only partially paying attention here0
[20:50:05] <KungFuJesus> eventually - it took its sweet time, though
[20:51:00] <KungFuJesus> and I'm pretty sure it wasn't going to do that on its own, I let it attempt to boot for like 30 minutes prior to me coming on site and troubleshooting / removing sas controllers / booting single user mode / removing ipmi from be / removing bad transient.conf
[20:51:08] <jbk> this is the original bug: https://www.illumos.org/issues/4454
[20:51:19] <jbk> however most of the incidents of that seem to trigger panics
[20:51:24] <KungFuJesus> I'm still don't _really_ know what fixed things, which scares me
[20:51:51] <jbk> though we've seen around that is what appears to be zfs hangs caused by fmd trying to retire devices
[20:52:11] <jbk> but no panics.. so i don't know yet if that's just another symptom of that bug
[20:52:20] <jbk> or another one that it's just making it easier to hit
[20:52:36] <jbk> (and it's #$@$#@ hard to recreate in a VM unfortunately)
[20:53:08] <KungFuJesus> the actual faults that had occurred happened cleanly. The devices were marked faulted in the pool and I reonline'd and cleared the errors and the pool was perfectly healthy
[20:53:43] <jbk> yeah, but during boot, I think fmd will try to offline/retire the devices in an effort to prevent them from causing problems
[20:53:48] <KungFuJesus> it was only after I image-update'd and booted into a new be that everything just hung on boot
[20:53:59] <KungFuJesus> ahhhh, it's too smart for its own good
[20:54:22] <KungFuJesus> ok, well next time if it occurs (hopefully it never does) I'll acquit the errors
[20:55:50] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has quit IRC (Remote host closed the connection)
[20:56:11] <jbk> i think at least for the panics, we have a fix, and the issues encountered during testing appear to be unrelated (and just extant iSCSI bugs since it was used to test)
[20:56:19] <jbk> but i don't know about the hangs
[20:57:36] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has joined #illumos
[20:59:48] *** alanc <alanc!~alanc@inet-hqmc01-o.oracle.com> has quit IRC (Remote host closed the connection)
[21:00:15] *** alanc <alanc!~alanc@inet-hqmc01-o.oracle.com> has joined #illumos
[21:02:19] *** neirac <neirac!~neirac@pc-184-104-160-190.cm.vtr.net> has quit IRC (Quit: Leaving)
[21:07:19] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has quit IRC (Remote host closed the connection)
[21:09:11] *** KungFuJesus <KungFuJesus!~adamstyli@207.250.97.74> has quit IRC (Quit: leaving)
[21:10:55] <tsoome> if bad bootpath was provided by transient.conf, then that was most probable cause for hung system — the validation code about bootpath and deviceid could be much better
[21:18:43] <sjorge> @papertigers ugh ow, the sooner it lands the sooner I want pr 888
[21:18:53] <sjorge> so I can do a follow up to have it working for bhyve :D
[21:24:59] <danmcd> leoric: Do you have -gate push access for 11522 ?
[21:27:35] *** BOKALDO <BOKALDO!~BOKALDO@81.198.23.132> has quit IRC (Quit: Leaving)
[21:34:14] <danmcd> Actually, I don't think you do, leoric, so I'll do the copy/paste of testing info and push for you.
[21:39:26] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has joined #illumos
[21:43:20] <gitomat> [illumos-gate] 11522 illumos-gate has some runtime dependencies on Python 2.7 -- Alexander Pyhalov <apyhalov at gmail dot com>
[21:43:21] <gitomat> [illumos-gate] 12188 libfmd_adm: Wrong prerequisites for generated files -- Marcel Telka <marcel at telka dot sk>
[21:57:58] <gitomat> [illumos-gate] 12186 The n2rng man page incorrectly refers to n2cp -- Peter Tribble <peter.tribble at gmail dot com>
[22:31:55] *** jellydonut <jellydonut!~quassel@s91904425.blix.com> has quit IRC (Quit: jellydonut)
[22:35:59] *** jellydonut <jellydonut!~quassel@s91904428.blix.com> has joined #illumos
[22:57:28] *** Kurlon_ <Kurlon_!~Kurlon@bidd-pub-04.gwi.net> has joined #illumos
[23:00:26] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has quit IRC (Ping timeout: 240 seconds)
[23:01:38] *** Kurlon_ <Kurlon_!~Kurlon@bidd-pub-04.gwi.net> has quit IRC (Ping timeout: 240 seconds)
[23:29:01] *** amrmesh <amrmesh!~Thunderbi@134.19.189.92> has joined #illumos
[23:29:40] *** amrfrsh <amrfrsh!~Thunderbi@134.19.189.92> has quit IRC (Ping timeout: 268 seconds)
[23:29:40] *** amrmesh is now known as amrfrsh
[23:33:26] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has quit IRC (Quit: Leaving)
[23:38:07] *** Kurlon <Kurlon!~Kurlon@cpe-67-253-141-249.rochester.res.rr.com> has joined #illumos
[23:50:42] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has quit IRC (Quit: Textual IRC Client: www.textualapp.com)
top

   January 15, 2020  
< | 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 | 31 | >