Switch to DuckDuckGo Search
   May 7, 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 | 31 | >

Toggle Join/Part | bottom
[00:00:19] <igork> i tried gcc6 and gcc7 - the same issue
[00:00:38] <igork> /rws1/users/igor/dilos-illumos.zfs/usr/src/tools/proto/root_i386-nd/opt/onbld/bin/ctfconvert -i -L VERSIONCTF -X debug64/zfs_ioctl.o
[00:00:38] <igork> ctfconvert: CTF conversion failed: failed to get unsigned attribute for type: DW_DLE_ATTR_FORM_BAD (114)
[00:01:26] <igork> try to port this update: https://github.com/zfsonlinux/zfs/commit/1b939560be5c51deecf875af9dada9d094633bf7#diff-5e2ee96ed866ac9856ad948d83693d46
[00:01:38] <richlowe> dwarfdump will probably help
[00:01:41] <igork> _ this one: https://github.com/zfsonlinux/zfs/commit/b83a0e2dc1b972e3ae5c0c17b714941b4604dfd6
[00:02:37] <igork> well - ctfconvert remove object - i'll try reproduce it
[00:03:33] <richlowe> I would probably look at zfs_ioc first
[00:03:53] <igork> object file or C file?
[00:04:23] <igork> without both patches i have no build issue
[00:04:31] <richlowe> the type
[00:05:01] <igork> i can share final merged file if it can help
[00:05:36] *** Kurlon <Kurlon!~Kurlon@98.13.72.207> has joined #illumos
[00:06:24] <igork> https://bitbucket.org/dilos/dilos-illumos/commits/a3f4dc105286639bd404199a9af4f8e1ba8b230c#chg-usr/src/uts/common/fs/zfs/zfs_ioctl.c
[00:06:27] <richlowe> the object before ctfconvert is what would help
[00:06:43] <richlowe> SOURCEDEBUG=yes when you build zfs, or STRIP=:
[00:06:57] <igork> ok, waiting full rebuild
[00:07:20] <igork> with full cleanup
[00:24:53] <igork> richlowe: old cft tools are fine
[00:25:11] <igork> i have added 'export BUILD_OLD_CTF_TOOLS=' to illumos.sh
[00:26:08] <igork> with your flags object file still removing. i can try to hack build env if you are interested in broken object + new ctf tools
[00:27:02] <richlowe> oh, duh, sorry I didn't think before I typed
[00:27:25] <igork> i can do ctfdump to zfs_ioctl.o with old ctf tools
[00:27:26] <richlowe> and yes, the broken object is the only way to debug it.
[00:27:50] <igork> and try to do the same with new ctf tools
[00:27:52] <richlowe> what I _wanted_ you to do is not ctfconvert it at all, but that's absolutely not what I said :)
[00:28:06] <igork> ok
[00:28:07] <richlowe> CTFCONVERT_O=: and rebuild in intel/zfs?
[00:28:17] <igork> i know how to disable it, thanks
[00:28:20] <richlowe> maybe need STRIP=: CTFCONVERT_O=:
[00:28:21] <igork> one moment
[00:39:49] *** polishdu1 <polishdu1!~polishdub@207.86.38.254> has quit IRC (Remote host closed the connection)
[00:40:12] <igork> http://apt2.dilos.org/dilos/logs/ctf/4/
[00:40:17] <igork> https://paste.dilos.org/?e8614e8e9cb09b81#lckhjNNA4HK7hTs79vQhKbpBVB0FubmjC6HpaJdP8H8=
[00:40:44] <igork> richlowe: ctfdump not working on object, but dwarfdump produce more info
[00:41:11] <igork> o.good - it is file with old ctf tools
[00:41:33] <igork> i have added flag '-k' to new ctf tools - keep object
[00:42:01] <igork> also, tested ctfdump on others objects - it is working fine
[00:42:10] <igork> but not on zfs_ioctl.o
[00:43:43] <igork> i'll try produce full bulid with old ctf tools and test it
[00:45:02] <igork> i need to test 'TRIM' if it will work with zfs tests
[01:06:39] <richlowe> looks like we're annoyed about zfs_keys_clone
[01:07:03] <igork> with new ctf tools?
[01:07:52] <igork> if you have some ideas/or patch - ping me my email - i can test tomorrow
[01:08:22] <richlowe> Oh, no
[01:08:29] <richlowe> I'll file a bug, rmustacc will be thrilled
[01:08:39] <igork> thanks
[01:10:54] *** whalebiologist is now known as mnrmnaugh
[01:15:12] <richlowe> https://www.illumos.org/issues/10922
[01:17:02] *** andy_js <andy_js!~andy@94.14.32.34> has quit IRC (Quit: andy_js)
[01:27:57] *** MilkmanDan <MilkmanDan!~dan@wilug/expat/MilkmanDan> has quit IRC (Ping timeout: 245 seconds)
[01:29:06] *** MilkmanDan <MilkmanDan!~dan@wilug/expat/MilkmanDan> has joined #illumos
[02:11:10] *** MarcelT <MarcelT!~marcel@tortuga.telka.sk> has quit IRC (Ping timeout: 244 seconds)
[02:17:10] *** MarcelT <MarcelT!~marcel@tortuga.telka.sk> has joined #illumos
[02:19:09] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has quit IRC (Quit: jcea)
[02:57:16] *** jocthbr <jocthbr!~salci@138-122-44-143.host.cicloti.com.br> has quit IRC (Quit: No Ping reply in 180 seconds.)
[02:58:36] *** jocthbr <jocthbr!~salci@138-122-44-143.host.cicloti.com.br> has joined #illumos
[03:02:06] *** jocthbr <jocthbr!~salci@138-122-44-143.host.cicloti.com.br> has quit IRC (Remote host closed the connection)
[03:03:13] *** jocthbr <jocthbr!~salci@138-122-44-143.host.cicloti.com.br> has joined #illumos
[03:09:49] *** idodecla_ <idodecla_!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[03:20:01] *** ed209 <ed209!~ed209@165.225.128.67> has quit IRC (Remote host closed the connection)
[03:20:09] *** ed209 <ed209!~ed209@165.225.128.67> has joined #illumos
[03:36:57] *** baojg <baojg!~baojg@162.243.44.213> has joined #illumos
[04:12:17] *** idodecla_ <idodecla_!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[04:31:42] *** MilkmanDan <MilkmanDan!~dan@wilug/expat/MilkmanDan> has quit IRC (Ping timeout: 245 seconds)
[04:33:07] *** MilkmanDan <MilkmanDan!~dan@wilug/expat/MilkmanDan> has joined #illumos
[04:43:29] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[05:01:06] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[05:03:02] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[06:04:21] <LeftWing> I would be thrilled with that much analysis in most bugs to be honest :P
[06:07:30] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[07:38:58] *** danmcd <danmcd!~danmcd@static-71-174-113-16.bstnma.fios.verizon.net> has quit IRC (Ping timeout: 246 seconds)
[07:39:12] *** jeffpc <jeffpc!~jeffpc@josefsipek.net> has quit IRC (Ping timeout: 245 seconds)
[07:40:02] *** jeffpc <jeffpc!~jeffpc@josefsipek.net> has joined #illumos
[07:44:29] *** neuroserve <neuroserve!~toens@195.71.113.124> has joined #illumos
[08:10:13] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Quit: tsoome)
[08:50:38] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[08:56:02] *** alanc <alanc!~alanc@129.157.69.43> has quit IRC (Remote host closed the connection)
[08:56:29] *** alanc <alanc!~alanc@129.157.69.43> has joined #illumos
[09:25:26] <leoric> andyf: Hi. Have you finally got hyperv drivers working?
[09:43:47] *** merzo <merzo!~merzo@34-48-133-95.pool.ukrtel.net> has quit IRC (Ping timeout: 245 seconds)
[09:55:47] *** BH23 <BH23!~BH23@194.75.135.30> has joined #illumos
[10:02:22] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has joined #illumos
[10:13:24] *** man_u <man_u!~manu@manu2.gandi.net> has joined #illumos
[10:24:44] *** BH23 <BH23!~BH23@194.75.135.30> has quit IRC (Remote host closed the connection)
[10:27:29] *** _joes_ <_joes_!~Thunderbi@194.75.135.30> has joined #illumos
[10:27:31] *** BH23 <BH23!~BH23@194.75.135.30> has joined #illumos
[10:29:29] *** sjorge <sjorge!~sjorge@unaffiliated/sjorge> has quit IRC (Quit: 410 Gone)
[10:29:34] *** sjorge_be <sjorge_be!~sjorge@unaffiliated/sjorge> has joined #illumos
[10:29:58] *** sjorge_be is now known as sjorge
[10:31:03] *** _joes_ <_joes_!~Thunderbi@194.75.135.30> has quit IRC (Remote host closed the connection)
[10:35:40] *** andy_js <andy_js!~andy@94.14.32.34> has joined #illumos
[10:42:48] *** merzo <merzo!~merzo@91.225.165.251> has joined #illumos
[11:12:57] <andyf> leoric, yes. At least under hyperv on windows, as long as there are multiple virtual CPUs. One of our developers uses that for all their work.
[11:13:28] <andyf> For running under Azure, I think there is still a missing userland component, but Delphix have offered to help with thatr.
[11:15:02] <leoric> Do you (or Delphix guys) have plans to upstream it?
[11:17:07] <andyf> Yes, I think we will do it once we've fixed the single CPU thing
[11:17:27] <andyf> Delphix don't as far as I know
[11:17:43] *** merzo <merzo!~merzo@91.225.165.251> has quit IRC (Ping timeout: 246 seconds)
[11:18:24] <leoric> will be happy to help (at least with testing)
[11:27:31] <andyf> Great, thanks :)
[11:27:44] <andyf> Now that we have OmniOS r151030 out, there should be more time again
[12:20:01] *** ed209 <ed209!~ed209@165.225.128.67> has quit IRC (Remote host closed the connection)
[12:20:17] *** ed209 <ed209!~ed209@165.225.128.67> has joined #illumos
[12:42:28] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has joined #illumos
[12:47:04] *** merzo <merzo!~merzo@195.20.130.10> has joined #illumos
[13:07:55] *** ullbeking <ullbeking!sid5364@gateway/web/irccloud.com/x-gqqdmkvesfeiztzg> has left #illumos
[13:31:07] *** atrius <atrius!~atrius@2600:1700:22d1:518:d71:89a4:946d:599> has quit IRC (Ping timeout: 264 seconds)
[13:33:48] *** atrius <atrius!~atrius@2600:1700:22d1:518:dd0d:e61d:d61f:6a93> has joined #illumos
[13:44:04] *** baojg <baojg!~baojg@162.243.44.213> has quit IRC (Remote host closed the connection)
[13:49:15] *** BH23 <BH23!~BH23@194.75.135.30> has quit IRC (Ping timeout: 248 seconds)
[14:03:49] *** andy_js <andy_js!~andy@94.14.32.34> has quit IRC (Read error: Connection reset by peer)
[14:04:11] *** andy_js <andy_js!~andy@94.14.32.34> has joined #illumos
[14:05:25] <tsoome> unless I am mistaken, it seems we found bug in ZoL :D
[14:06:22] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has quit IRC (Quit: jcea)
[14:08:38] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has joined #illumos
[14:11:00] <tsoome> and hidden by compiler:)
[14:13:42] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has joined #illumos
[14:26:12] *** ngchk1 <ngchk1!~ngchk1@b2b-92-50-91-166.unitymedia.biz> has joined #illumos
[14:27:55] *** ngchk1_ <ngchk1_!~ngchk1@b2b-92-50-91-166.unitymedia.biz> has quit IRC (Ping timeout: 268 seconds)
[14:54:42] <andyf> Nice find :) Hidden by our compiler, theirs or both?
[14:55:23] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Ping timeout: 248 seconds)
[14:58:16] *** BH23 <BH23!~BH23@194.75.135.30> has joined #illumos
[15:03:13] *** rsully <rsully!~rsully@unaffiliated/rsully> has joined #illumos
[15:20:34] <igork> tsoome: what issue?
[15:20:34] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[15:25:45] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[15:26:51] *** Kurlon <Kurlon!~Kurlon@98.13.72.207> has quit IRC (Ping timeout: 248 seconds)
[15:32:46] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[15:33:17] *** ngchk1 <ngchk1!~ngchk1@b2b-92-50-91-166.unitymedia.biz> has quit IRC (Ping timeout: 268 seconds)
[15:35:52] *** ngchk1 <ngchk1!~ngchk1@b2b-92-50-91-166.unitymedia.biz> has joined #illumos
[15:47:02] *** ruid__ <ruid__!~ruid@unaffiliated/ruid> has joined #illumos
[15:47:58] *** ngchk1 <ngchk1!~ngchk1@b2b-92-50-91-166.unitymedia.biz> has quit IRC (Ping timeout: 245 seconds)
[15:49:19] *** ruid_ <ruid_!~ruid@unaffiliated/ruid> has quit IRC (Ping timeout: 246 seconds)
[15:53:54] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has quit IRC (Quit: Leaving)
[16:00:52] *** ruid__ <ruid__!~ruid@unaffiliated/ruid> has quit IRC (Ping timeout: 258 seconds)
[16:17:19] *** neuroserve <neuroserve!~toens@195.71.113.124> has quit IRC (Ping timeout: 246 seconds)
[16:19:27] *** thebug <thebug!~thebug@unaffiliated/thebug> has quit IRC (Quit: leaving)
[16:24:30] *** thebug <thebug!~thebug@unaffiliated/thebug> has joined #illumos
[16:29:37] *** ruid__ <ruid__!~ruid@unaffiliated/ruid> has joined #illumos
[16:36:35] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has joined #illumos
[16:42:40] <LeftWing> leoric: In the realm of guest drivers, I have a rough prototype Virtio SCSI driver too!
[16:42:58] <LeftWing> First I'm working on adding support for hosts that don't allow for "indirect descriptors" to the virtio framework, though.
[16:43:27] <LeftWing> 2019 is the year of good virtual machine guest support!
[16:45:03] <tsoome> so we get things ported to gate?:P
[16:45:19] <andyf> Can we support hypervisors without a PIC? :run:
[16:45:46] <LeftWing> tsoome: I'm working directly in illumos-gate with this stuff
[16:45:49] <andyf> tsoome, well I'll do HyperV.. I might need help with fixing the crosscall stuff
[16:46:09] <LeftWing> andyf: I've done a crosscall before! I'm sure we can work it out.
[16:47:31] <andyf> I'll take you up on that, thanks
[16:48:35] <andyf> LeftWing, when I get back to it, I'll write up some stuff and drop you the link.
[16:48:42] <LeftWing> Great
[17:00:39] <rmustacc> andyf: I'm happy to help with cross calls.
[17:00:41] <rmustacc> Please reach out there.
[17:00:53] <rmustacc> And yes, we can definitely support things without a PIC.
[17:01:31] <gitomat> [illumos-gate] 10891 dmake could support -C -- Robert Mustacchi <rm at joyent dot com>
[17:01:46] <andyf> I have a wip branch for running under a gen-2 hyperv machine, using the fixed 10MHz timer that they provide via an MSR
[17:02:02] <andyf> At least it works in terms of tenmicrosec() and APIC calibration
[17:02:20] <andyf> but we probably need some kind of pluggable/flexible framework
[17:02:38] <rmustacc> Makes sense.
[17:03:03] <LeftWing> It's exciting that they've just jettisoned the PIT emulation altogether :\
[17:03:05] <rmustacc> Do you have an idea of how you want to build that?
[17:03:26] <rmustacc> Is that something you'd like help with?
[17:03:44] <andyf> Yes and yes
[17:04:00] <andyf> It ties into LeftWing's GCE work too I think
[17:04:01] <LeftWing> Recall there's also freq_tsc() that includes assembly that uses the PIT for calibration as well
[17:04:12] <andyf> LeftWing, Yes, that's the last bit I have to look at
[17:04:13] <LeftWing> Well it looks like GCE are actually going to fix their hypervisor
[17:04:36] <rmustacc> What's the most useful way to help?
[17:04:39] <andyf> the hand-crafted assembly there got me to stop for a bit (other priorities anyway)
[17:04:39] <LeftWing> See also: https://issuetracker.google.com/issues/130531009
[17:04:54] *** m4rley <m4rley!~m4rley@207.148.96.120> has quit IRC (Quit: leaving)
[17:05:09] <LeftWing> So I suspect a framework approach would leave the assembly that exists there alone and just provide a compatible freq_tsc() using a different mechanism
[17:05:11] *** m4rley <m4rley!~m4rley@207.148.96.120> has joined #illumos
[17:05:17] <andyf> rmustacc, I should probably start an IPD around plugabble early-boot-timers
[17:05:29] <LeftWing> That's a great idea! There are like five moving parts here.
[17:05:36] <andyf> and the initial help would be guiding me in terms of making a sane architecture
[17:05:53] <rmustacc> OK, well, just let me know how we can best help.
[17:06:14] <andyf> As an extension, some hypervisors like hyperv provide a different (better?) TSC too
[17:06:46] <andyf> rmustacc, will do, thanks.
[17:06:49] <rmustacc> OK.
[17:12:17] <andyf> LeftWing - nice response from Google!
[17:18:50] <rmustacc> richlowe: I'll take a look, thanks.
[17:19:41] *** polishdub <polishdub!~polishdub@207.86.38.254> has joined #illumos
[17:26:02] *** ruid__ <ruid__!~ruid@unaffiliated/ruid> has quit IRC (Ping timeout: 250 seconds)
[17:37:20] *** BH23 <BH23!~BH23@194.75.135.30> has quit IRC (Ping timeout: 255 seconds)
[17:37:46] *** ruid__ <ruid__!~ruid@unaffiliated/ruid> has joined #illumos
[17:38:00] *** ruid__ <ruid__!~ruid@unaffiliated/ruid> has quit IRC (Remote host closed the connection)
[17:53:49] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has joined #illumos
[17:55:39] *** merzo <merzo!~merzo@195.20.130.10> has quit IRC (Ping timeout: 248 seconds)
[18:16:34] *** ngchk1 <ngchk1!~ngchk1@b2b-92-50-91-166.unitymedia.biz> has joined #illumos
[18:21:18] *** AllanJude <AllanJude!~allan@freebsd/developer/AllanJude> has quit IRC (Ping timeout: 252 seconds)
[18:22:24] *** man_u <man_u!~manu@manu2.gandi.net> has quit IRC (Quit: man_u)
[18:23:03] *** man_u <man_u!~manu@manu2.gandi.net> has joined #illumos
[18:34:57] *** ngchk1 <ngchk1!~ngchk1@b2b-92-50-91-166.unitymedia.biz> has quit IRC (Ping timeout: 258 seconds)
[18:36:48] *** AllanJude <AllanJude!~allan@allanjude.bsdbasement.com> has joined #illumos
[18:36:48] *** AllanJude <AllanJude!~allan@allanjude.bsdbasement.com> has quit IRC (Changing host)
[18:36:48] *** AllanJude <AllanJude!~allan@freebsd/developer/AllanJude> has joined #illumos
[18:37:06] *** man_u <man_u!~manu@manu2.gandi.net> has quit IRC (Ping timeout: 250 seconds)
[18:49:48] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has quit IRC (Remote host closed the connection)
[18:50:00] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has joined #illumos
[19:04:16] *** AllanJude <AllanJude!~allan@freebsd/developer/AllanJude> has quit IRC (Ping timeout: 246 seconds)
[19:07:58] *** AllanJude <AllanJude!~allan@freebsd/developer/AllanJude> has joined #illumos
[19:17:13] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has quit IRC (Remote host closed the connection)
[19:30:12] <LeftWing> andyf: Yes, I was very pleased!
[19:45:37] *** ngchk1 <ngchk1!~ngchk1@b2b-92-50-91-166.unitymedia.biz> has joined #illumos
[20:14:47] <Smithx10> lol, LeftWing is 1450 MTU a bug? :P
[20:15:01] <LeftWing> Smithx10: Hmm?
[20:15:05] <Smithx10> GCP *
[20:15:10] <LeftWing> They use 1460
[20:15:16] <Smithx10> oOOoo my bad !
[20:15:26] <Smithx10> those 10 extra, keep that from being a bug !
[20:16:16] <LeftWing> I presume it is to leave room for the 50 byte VXLAN header in a regular Ethernet frame
[20:16:31] <Smithx10> Yea. ya gotta pay the toll troll.
[20:16:39] <Smithx10> troll toll*
[20:55:31] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has joined #illumos
[21:38:32] *** wiedi <wiedi!~wiedi@ip5b4096a6.dynamic.kabel-deutschland.de> has quit IRC (Ping timeout: 255 seconds)
[21:43:31] *** wiedi <wiedi!~wiedi@ip5b4096a6.dynamic.kabel-deutschland.de> has joined #illumos
[22:06:39] <richlowe> rmustacc: jlevon already closed it duplicate, because I suck
[22:07:03] <jlevon> the smap one? rm mentioned it was already filed
[22:07:49] <richlowe> the really definitely always and forever empty struct we want to declare anyway one
[22:08:04] <richlowe> "array of struct"
[22:08:18] <rmustacc> Yeah, we should just fix it up. I need to clear some other stuff first, tbh.
[22:08:25] <rmustacc> But I was just going to treat it like [0]
[22:08:43] <richlowe> Seems like it's Jerry's openzfs that'll notice it first
[22:08:55] <richlowe> so that probably gives you some control of when you have to care
[22:09:12] <rmustacc> Just got some other stuff going on atm I need to take care of.
[22:09:27] *** neirac <neirac!~cneira@190.162.109.53> has joined #illumos
[22:09:27] <rmustacc> There are a few other things that I should deal with at the same time, I expect.
[22:09:30] <richlowe> jlevon: also, I know we're all being kind to gwr whenever he asks the "never have another 32bit kernel" question
[22:09:34] <richlowe> but like, no. There never will be
[22:09:53] <richlowe> surely we're at the point we can just like, say that?
[22:11:19] <neirac> Smithx10 I just saw your comment regarding nomad, what I understood is that we need to wait for a merge before someone tries to create a nomad plugin
[22:11:41] <jlevon> richlowe: probably? But it wasn't relevant so *shrug*
[22:12:28] <richlowe> jlevon: No, but it's relative to the conversation that's cropping up increasingly about just flushing 32bit binaries of our own to solve the time issue
[22:12:35] <Smithx10> neirac: I just figured I'd add you since you worked on the zcage stuff
[22:13:02] <Smithx10> But yeah, adding a Nomad Plugin that spawns Zones etc from the GZ is waiting on the illumos build tag
[22:13:19] <jlevon> richlowe: yes and no... I was never a fan of making 32-bit userspace impossible, not just unnecessary
[22:13:43] <Smithx10> so we can make changes to the gopsutil library they use for "fingerprinting", which just gets use metrics back
[22:13:54] <Smithx10> so you know if the node is full or not etc.
[22:14:48] <Smithx10> i wrote a small plugin for Triton, but we should probably add that to omnios / smartos etc
[22:15:37] <Smithx10> On SmartOS we'd definitely wind up using vmadm and imgadm, but on other distros maybe how you did zcage is the correct abstraction?
[22:16:43] <richlowe> jlevon: I'm not seeing reasons we'd ever need a 32bit ls again?
[22:17:14] <jlevon> richlowe: that's a different matter from disabling ALL 32-bit applications.
[22:17:26] <jlevon> we use 64-bit when we can
[22:17:37] <neirac> Smithx10, I was looking yesterday at nomad plugins to create one for illumos, I would end up calling zoneadm/zonecfg so zcage is not a dependency. I have not started any coding yet, just watching how nomad worked
[22:18:41] <Smithx10> Yea, I mean't the libraries that zcage calls otu to*
[22:18:49] <Smithx10> my apologies. I dont mean wrapping zcage *
[22:19:09] <Smithx10> I'd be interested in helping you
[22:19:27] <Smithx10> I did https://github.com/Smithx10/nomad-driver-triton
[22:19:40] <Smithx10> which needs some tests and a bit more work with hashicorp
[22:19:47] <neirac> Smithx10 no problem :), that would be awesome because I just was learning how to use nomad
[22:19:51] <Smithx10> being a "proxy" requires a few methods they havent implemented
[22:20:00] *** ed209 <ed209!~ed209@165.225.128.67> has quit IRC (Remote host closed the connection)
[22:20:08] *** ed209 <ed209!~ed209@165.225.128.67> has joined #illumos
[22:20:08] <Smithx10> right now we just use it for stateless workloads
[22:20:17] <richlowe> jlevon: Right, I'm not in favour of disabling all 32bit apps! I didn't mean to suggest that
[22:20:26] <jlevon> ok
[22:20:37] <richlowe> jlevon: though I'm sort of in favour of saying "If you need safe time, you should be 64bit" to the degree we can. LeftWing was selling me on it
[22:20:48] <Smithx10> I think it would definitely help boost the adoption of Illumos :)
[22:21:13] <rmustacc> richlowe: So I shouldn't actually look at time64_t?
[22:21:17] <rmustacc> And all that?
[22:21:25] <richlowe> rmustacc: instead of that
[22:21:39] <jlevon> I'm totally fine with requiring 64-bit.
[22:22:04] <richlowe> rmustacc: Though when I was talking to LeftWing obviously the things that are tied to the ABI of their victim like libelf and such came up
[22:22:08] <richlowe> rmustacc: so you'll probably have to do it anyway :)
[22:22:24] <rmustacc> Haha
[22:22:35] <rmustacc> So I'll have to grumble at both ends then?
[22:24:26] <neirac> Smithx10 as I see in the code you are right it would be easier to use zcage libraries, the problem is that I should translate them from node to golang
[22:26:21] <LeftWing> richlowe: You did mention that there are some things like libproc and libelf that probably _need_ to be 32-bit unless they get fully rearchitected
[22:26:46] <LeftWing> Which to be honest made me think that maybe we _should_ do largetime(5) :P
[22:26:55] <LeftWing> (ltcompile(5)?)
[22:28:15] <richlowe> so we're hoping rmustacc makes the big time?
[22:28:26] <neirac> Smithx10 you built nomad from source right? I don't see it in pkgsrc
[22:28:28] <LeftWing> richlowe: haha
[22:28:31] <LeftWing> bigtime(5)
[22:28:34] <LeftWing> Best
[22:28:50] <LeftWing> neirac: Does one not simply "go get" it?
[22:30:31] <toasterson> neirac smithx10 there is also https://git.wegmueller.it/illumos/go-zone if you need something to interact with zones on illumos and https://git.wegmueller.it/opencloud/opencloud if you need things like OCI writer/reader and image layers on zfs. Still somewhat alpha but will get more stable soon.
[22:30:40] <neirac> @LeftWing I don't know, there has been a long time I don't do anything in golang.I'll read the basics
[22:33:01] <neirac> toasterson awesome!, thanks a lot.
[22:33:32] <toasterson> I am currently working on a docker run like tool. it (almost) works
[22:33:48] <toasterson> at least it can interact with docker hub :)
[22:34:54] <neirac> toasterson, that I see. You have almost all the pieces together if not all
[22:35:29] <toasterson> all together but still buggy. imagename parsing needs a bit of work.
[22:40:09] *** jubal <jubal!~jubal@226-5-237-24.gci.net> has joined #illumos
[22:51:55] *** merzo <merzo!~merzo@0-39-133-95.pool.ukrtel.net> has joined #illumos
[22:58:08] *** clapont <clapont!~clapont@unaffiliated/clapont> has quit IRC (Ping timeout: 252 seconds)
[23:00:08] *** clapont <clapont!~clapont@unaffiliated/clapont> has joined #illumos
[23:00:12] *** sarvet <sarvet!~sarvet@dslb-188-101-015-082.188.101.pools.vodafone-ip.de> has joined #illumos
[23:07:20] <neirac> Smithx10 I think toasterson tool would be perfect for the nomad plugin if it's like docker
[23:21:19] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has quit IRC (Ping timeout: 258 seconds)
[23:31:57] *** rsully <rsully!~rsully@unaffiliated/rsully> has quit IRC (Quit: rsully)
[23:33:44] *** porkotron <porkotron!laski@punainen.org> has quit IRC (Ping timeout: 252 seconds)
[23:43:08] <Smithx10> toasterson: I think mgerdts might be interested that
[23:43:30] *** leoric_ <leoric_!~leoric@2a02:2698:6022:924a:c106:2250:9e7d:b139> has joined #illumos
[23:43:31] *** leoric_ <leoric_!~leoric@2a02:2698:6022:924a:c106:2250:9e7d:b139> has quit IRC (Client Quit)
[23:43:42] <Smithx10> In that, think he was investigating a k8s interface type thingie
[23:44:28] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has quit IRC (Quit: Leaving)
[23:44:29] <Smithx10> neirac: yeah, but on SmartOS vmadm does a lot
[23:45:35] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has quit IRC (Quit: Textual IRC Client: www.textualapp.com)
top

   May 7, 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 | 31 | >