Switch to DuckDuckGo Search
   November 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 | >

NOTICE: This channel is no longer actively logged.

Toggle Join/Part | bottom
[00:45:28] *** Kruppt <Kruppt!~Kruppt@> has quit IRC (Quit: Leaving)
[01:30:38] *** Qatz <Qatz!~db@2601:187:8400:5::42d> has quit IRC (Read error: Connection reset by peer)
[01:30:43] *** Qatz- <Qatz-!~db@2601:187:8400:5::42d> has joined #illumos
[01:32:16] *** Qatz- is now known as Qatz
[01:53:25] <gitomat> [illumos-gate] 11922 ipmi_open looks for wrong return value -- Jason King <jason.king at joyent dot com>
[02:15:23] *** mnrmnaugh <mnrmnaugh!~mnrmnaugh@unaffiliated/mnrmnaugh> has quit IRC (Ping timeout: 276 seconds)
[02:18:42] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has quit IRC (Remote host closed the connection)
[02:20:00] *** ed209 <ed209!~ed209@> has quit IRC (Remote host closed the connection)
[02:20:07] *** ed209 <ed209!~ed209@> has joined #illumos
[02:21:41] *** mnrmnaugh <mnrmnaugh!~mnrmnaugh@unaffiliated/mnrmnaugh> has joined #illumos
[03:08:58] *** _Tenchi_ <_Tenchi_!~phil@> has quit IRC (Quit: Textual IRC Client: www.textualapp.com)
[03:27:13] *** alanc <alanc!~alanc@inet-hqmc01-o.oracle.com> has quit IRC (Remote host closed the connection)
[03:27:39] *** alanc <alanc!~alanc@inet-hqmc01-o.oracle.com> has joined #illumos
[03:34:29] *** _Tenchi_ <_Tenchi_!~phil@d-207-255-80-203.paw.cpe.atlanticbb.net> has joined #illumos
[04:14:59] *** jessfraz_ <jessfraz_!~jessfraz@unaffiliated/jessfraz> has quit IRC (Remote host closed the connection)
[04:15:56] *** jessfraz_ <jessfraz_!~jessfraz@unaffiliated/jessfraz> has joined #illumos
[04:28:14] *** sbehrens_ <sbehrens_!~berry@diome.rzone.de> has joined #illumos
[04:30:40] *** sbehrens <sbehrens!~berry@diome.rzone.de> has quit IRC (Ping timeout: 244 seconds)
[05:14:56] *** hemi770 <hemi770!~hemi666@unaffiliated/hemi770> has quit IRC (Ping timeout: 265 seconds)
[05:16:35] *** hemi770 <hemi770!~hemi666@unaffiliated/hemi770> has joined #illumos
[07:10:39] *** mnowak_irssi <mnowak_irssi!~newman@> has joined #illumos
[07:10:44] *** mnowak_irssi <mnowak_irssi!~newman@> has quit IRC (Client Quit)
[07:20:55] <jbk> does anyone know offhand if it's possible to get a notification _in userland_ when a USB device is inserted? would it be a sysevent?
[07:23:45] <LeftWing> jbk: I'm not sure. I think there are at least events about device node creation?
[07:24:00] <LeftWing> One would imagine that some kind of generic hot plug device event would be good, if we don't have one already
[07:24:22] <LeftWing> But yeah I'd expect a sysevent of some kind presumably
[07:27:36] <gitomat> [illumos-gate] 11792 ibtl: cast between incompatible function types -- Toomas Soome <tsoome at me dot com>
[07:28:09] <LeftWing> richlowe, tsoome: You're both very quick off the mark :P
[07:28:24] <tsoome> :P
[07:33:37] <tsoome> LeftWing: what is the state of gerrit now?
[07:34:16] <tsoome> I myself have not managed to set up single review up there:D
[07:34:45] <LeftWing> tsoome: I just updated the master ref there, so you may be able to push one now?
[07:35:08] <LeftWing> It'll be easier once we're pushing through it, which I should be able to finish up this week. The password resets were one of the last few steps.
[07:36:34] <tsoome> is it still true I can not have custom branch?
[07:37:27] <LeftWing> In what sense? It is in theory possible to grant different permissions to some branch other than master, if we decide we want to do thta
[07:39:57] <tsoome> with reviewboard I can post whatever patch - rbt post —parent commit
[07:41:23] <LeftWing> With Gerrit it's more that you push an actual commit. As long as the parent commit is already up there in Gerrit, it'll let you push whatever commit you have outstanding on your local branch
[07:41:27] <tsoome> if I must have patch on top of master, it means the git branches are basically useless
[07:42:02] <LeftWing> If you're getting a series of commits reviewed, you can push them all up
[07:42:25] <LeftWing> e.g., if you have a branch with three commits, I believe you can push all three up and each one will get a review of its own created
[07:42:51] <jbk> older versions worked that way
[07:44:03] <jbk> with the commit-id bit, it might actually make doing that a bit simpler than in older versions
[07:44:23] <LeftWing> Yes, it does, because it's able to match them all up by itself (as long as the Change-Id doesn't change)
[07:44:42] <jbk> where you had to push the right commit to the right branch
[07:44:47] <jbk> after updating and squashing
[07:45:22] <LeftWing> You still need to squash down to the change you're trying to get integrated
[07:46:20] <jbk> yeah, but if you have multiple changes on a branch, no more push commit1:refs/changes/123; push commit2:refs/changes/456; etc.
[07:46:45] <LeftWing> Right you can just "git push branch:refs/for/master" I think
[07:46:59] <LeftWing> Where "branch" is the branch name
[07:51:29] * LeftWing heads off to get some sleep &
[07:57:50] <tsoome> yee
[07:57:58] <tsoome> got first one up:)
[08:00:43] <tsoome> ok and I stil lcan have my local branches, just the pushes are for master.
[08:09:14] *** pwinder <pwinder!~pwinder@> has joined #illumos
[08:11:07] *** mnowak_ <mnowak_!~mnowak_@> has quit IRC (Quit: Leaving)
[08:13:11] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Quit: tsoome)
[08:26:12] *** snuff-work <snuff-work!~snuff-wor@202-161-112-134.tpgi.com.au> has quit IRC (Quit: Leaving.)
[08:26:14] *** mnowak_ <mnowak_!~mnowak_@> has joined #illumos
[08:36:53] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has joined #illumos
[08:52:36] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[08:58:32] <toasterson> tsoome (IRC): are you on openindiana-discussion mailing list? it looks like that there is an issue with loader on two of our images text and gui when used iver iDRAC virtual CD wich I would need help debugging some time in the next few days.
[08:59:13] <tsoome> no i am not
[09:01:15] <toasterson> tsoome (IRC): cool I'll CC you then. So we can find out why this happens with the images. I'll sum up the thread for you this evening. I hope our image creation script is not faulty
[09:01:17] <tsoome> bios boot or uefi?
[09:01:22] <toasterson> bios
[09:02:03] <toasterson> omnios work and oi minimal iso works but not oi text iso or oi gui iso
[09:02:31] <tsoome> does not work how?
[09:02:52] * toasterson sent a long message: < https://matrix.wegmueller.it/_matrix/media/v1/download/matrix.wegmueller.it/xoPbkHmaYcYcHsGJgAIGHNzL >
[09:04:14] <tsoome> zf_read: fill error is error from gunzip
[09:04:16] <toasterson> also another one
[09:04:20] * toasterson sent a long message: < https://matrix.wegmueller.it/_matrix/media/v1/download/matrix.wegmueller.it/LpgBdTCgIlskrIBFRoHmYcnB >
[09:07:41] <tsoome> second one is interesting in different way - we apparently get error while reading the kernel, but no physical IO error is reported.
[09:09:44] *** MarcelT <MarcelT!~marcel@tortuga.telka.sk> has joined #illumos
[09:15:05] <gitomat> [illumos-gate] 3334 zonestat missing man page -- Andy Fiddaman <omnios at citrus-it dot co.uk>
[09:16:00] <toasterson> tsoome (IRC): hrmm transparent link compression or something like that?
[09:16:59] <tsoome> since boot archive is huge, we gzip it and name as boot_archive. boot loader is reading it and uncompressing on the go.
[09:18:10] <tsoome> but since you have those 2 kinds of errors, it means we get IO issues, yet the interesting part is that those errors are not reported from biosdisk.c
[09:20:48] <toasterson> tsoome (IRC): probably hidden from bios of the server itself?
[09:21:12] <toasterson> is the freebsd boot_archive smaller? because that one loads. and omnios too
[09:38:41] *** ricksha-512 <ricksha-512!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has quit IRC (Ping timeout: 276 seconds)
[09:38:44] *** ricksha-512 <ricksha-512!~despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has joined #illumos
[09:42:41] *** kerberizer <kerberizer!~luchesar@wikipedia/Iliev> has quit IRC (Ping timeout: 246 seconds)
[09:48:07] *** andy_js <andy_js!~andy@97e29e78.skybroadband.com> has joined #illumos
[09:48:42] <tsoome> freebsd does not use boot archive. they have mix of compiled in modules and loadable modules
[09:50:27] <toasterson> tsoome (IRC): any trick to make loader spew more information about the errors? From there we could check what iDrac does weird
[09:51:02] <tsoome> checking that.
[09:57:36] *** kerberizer <kerberizer!~luchesar@wikipedia/Iliev> has joined #illumos
[10:10:01] *** man_u <man_u!~manu@manu2.gandi.net> has joined #illumos
[10:26:15] <leoric> Does someone else want to look at http://buildzone.oi-build.r61.net/webrev-11934/ ?
[10:36:44] *** arnoldoree <arnoldoree!~arnoldore@ranoldoree.plus.com> has joined #illumos
[10:44:39] <tsoome> leoric: LGTM
[10:45:11] <leoric> tvx
[10:45:13] <leoric> thx
[11:10:24] <tsoome> toasterson: it seems like your system is returning reads with one of following codes: result == 0x20 || result == 0x31 || result == 0x80 - we do not print error messages on those.
[11:11:04] <tsoome> 0x20 is controller failure, 31 and 80 are no media in drive
[11:16:26] *** jimklimov <jimklimov!~jimklimov@37-48-51-144.nat.epc.tmcz.cz> has joined #illumos
[11:19:19] <tsoome> toasterson: I’d do 2 things first - check the firmware update and try uefi boot. If we really are getting read errors, only thing we can do is to reset the controller and try again - the biosdisk.c does 3 attempts there (except for those 3 errors).
[11:20:00] *** ed209 <ed209!~ed209@> has quit IRC (Remote host closed the connection)
[11:20:07] *** ed209 <ed209!~ed209@> has joined #illumos
[11:23:42] <tsoome> toasterson: what we can do is to repeat on all errors, it may help but there is no way to tell for sure…
[11:25:56] *** pwinder <pwinder!~pwinder@> has quit IRC (Quit: Leaving)
[11:27:11] <toasterson> tsoome (IRC): ah nice thanks. It's not mine two people have reported these errors with their idrac. I'll let them know and cc you
[11:27:14] *** pwinder <pwinder!~pwinder@> has joined #illumos
[11:31:37] *** jimklimov <jimklimov!~jimklimov@37-48-51-144.nat.epc.tmcz.cz> has quit IRC (Read error: Connection reset by peer)
[12:30:56] *** neirac <neirac!~cneira@pc-184-104-160-190.cm.vtr.net> has joined #illumos
[12:32:40] <leoric> New OI snapshot is ready http://docs.openindiana.org/release-notes/2019.10-release-notes/
[12:34:01] <gitomat> [illumos-gate] 11924 infinite loop in mdb ::load -- Mark Brooks <mark.brooks at joyent dot com>
[12:37:00] <gitomat> [illumos-gate] 11929 mac_minor_hold() gets id_alloc_nosleep() wrong -- John Levon <john.levon at joyent dot com>
[13:10:59] <toasterson> leoric (IRC): 👍️
[13:22:37] *** Kurlon <Kurlon!~Kurlon@cpe-67-253-141-249.rochester.res.rr.com> has quit IRC (Ping timeout: 240 seconds)
[13:40:45] <AmyMalik> tsoome: do you think it'd be germane to have the system print "no media in drive" for 31h and 80h
[13:40:52] *** AmyMalik is now known as Reinhilde
[13:47:39] <tsoome> we print it in lsdev -v
[13:48:23] <tsoome> but bios implementations are so messed up you never know what you are going to see next
[13:50:02] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[13:55:37] <jlevon> tsoome: hey, I filed https://www.illumos.org/issues/11941 about my weird root pool just so it's not forgotten, hope that's OK?
[13:56:16] <tsoome> yes it is (have not forgotten anyhow;)
[13:58:48] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[13:59:30] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[14:01:23] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[14:02:47] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[14:03:47] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[14:04:40] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[14:05:35] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[14:06:05] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[14:07:12] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[14:07:32] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[14:08:42] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[14:10:00] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[14:11:06] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[14:12:20] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[14:14:11] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[14:16:03] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[14:23:16] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[14:23:33] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has joined #illumos
[14:23:49] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[14:33:18] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has joined #illumos
[14:33:34] *** neirac <neirac!~cneira@pc-184-104-160-190.cm.vtr.net> has quit IRC (Quit: Leaving)
[14:41:17] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Ping timeout: 240 seconds)
[14:54:20] *** heroux <heroux!sandroco@gateway/shell/insomnia247/x-uowyrinzwjvslrsi> has quit IRC (Ping timeout: 268 seconds)
[15:06:03] *** heroux <heroux!sandroco@gateway/shell/insomnia247/x-hipvyozqxgshedcz> has joined #illumos
[15:13:34] *** emfipp <emfipp!~emfipp@unaffiliated/mjkr> has joined #illumos
[15:14:18] <emfipp> does illumos have a port or a plan thereof to powerpc?
[15:14:45] <Reinhilde> no
[15:14:48] <Reinhilde> not that i know of
[15:16:29] <toasterson> emfipp (IRC): not currently
[15:16:38] <toasterson> no hardware
[15:16:47] <Reinhilde> i would throw a party if someone had even an early port
[15:16:52] <Reinhilde> to, say, ARM
[15:17:05] <toasterson> that does exist though
[15:17:09] <jperkin> we have a very early port to ARM
[15:17:13] <emfipp> ah, arm boards are in abundance though
[15:17:19] <emfipp> so are power9 boards
[15:17:33] <Reinhilde> jperkin: oh?
[15:17:44] <toasterson> not in our private households :)
[15:17:54] <emfipp> a while ago tyan was offering dirt-cheap power8 boards in EU for like 300-400 euros
[15:18:34] <jperkin> Reinhilde: like, really early, https://illumos.topicbox.com/groups/arm if you are interested
[15:19:17] <Reinhilde> jperkin: is it late enough that somebody has made an (I know it's not going to be an ISO on ARM but I'll say ISO anyway) ISO of it?
[15:19:38] <jperkin> nowhere near that stage yet
[15:19:41] <Reinhilde> dang
[15:19:52] <Kurlon> Actually, ISO booting is supported on bigger UEFI ARM platforms.
[15:20:36] <Reinhilde> Kurlon: you mean CD/DVD/Blu Ray, right?
[15:21:05] <Kurlon> Or actual ISO file
[15:21:22] <Reinhilde> that's intense
[15:24:35] <Reinhilde> jperkin: I can only see one problem with this branch of the project.
[15:24:41] <Reinhilde> I invite you to guess what that is.
[15:25:08] <Kurlon> There's been a BIG push to get the arm platform to the point where you no longer need SoC/Board specific distros/boot media/images to foster growth in the server arena.
[15:25:23] <Reinhilde> Kurlon: That's not the problem.
[15:25:25] <Reinhilde> That's the easy part.
[15:26:12] <Kurlon> HAH, you say that... have you implemented it? Coming from a Linux ARM distro, WHAT A CLUSTER! :D
[15:26:33] <Reinhilde> Kurlon: But it's the easy part because it's actually possible.
[15:26:49] <emfipp> in the server arena however, I can only think of system76/gigabyte offering servers with uefi booting
[15:26:58] <Kurlon> Now, my experience is with embedded SoCs, not the nice big server targeted stuff so I expect there to be more shenanigans in that space.
[15:27:00] <emfipp> lenovo's long promised an arm server from ampere
[15:27:07] <emfipp> however, the thing's not forthcoming
[15:27:29] <emfipp> arm servers, I mean
[15:27:53] <Reinhilde> guys, it's the elephant in the room, it's that tarball of crud you have to download from joyent
[15:29:18] <emfipp> thus, in the near future most arm server boards still need bsps to wrk
[15:29:47] <Reinhilde> somebody throw me a name of an ARM workstation
[15:29:59] <emfipp> solid run's honeycomb
[15:30:08] <toasterson> pinetop
[15:30:13] <Woodstock> Reinhilde: iyonix :)
[15:30:20] <Reinhilde> holy hell that's a lot of names
[15:30:23] <emfipp> nvidia's xavier dev box
[15:30:31] <jperkin> Woodstock: woah, blast from the past
[15:30:35] <Woodstock> :)
[15:30:35] <emfipp> none of them runs on uefi booting..
[15:31:05] <Reinhilde> what, by the way, is the difference between a WS and a desktop? I feel like I'm not up on some terminology
[15:31:20] <jperkin> I still have dnard shark somewhere (arm+openfirmware)
[15:31:40] <Reinhilde> jperkin: wasn't OF what the sun^Woracle boxes ran on?
[15:31:42] <Kurlon> The line I draw between embedded vs desktop/ws/server is if it has ram slots
[15:32:12] <Reinhilde> Kurlon: but what's the line between WS and desktop?
[15:32:30] <Kurlon> Really isn't one
[15:33:06] <Reinhilde> so the line is $1000
[15:33:06] <Kurlon> In the ye olden times, a WS was a desktop built on server level kit, so SMP, SCSI, etc. But well now, the line is more like a faint memory.
[15:33:07] <Reinhilde> ok
[15:33:23] <Agnar> Reinhilde: OpenBootPROM (OBP) is similiar to OF
[15:33:29] <Reinhilde> Kurlon: "the the olden"
[15:33:29] <Agnar> but not the same
[15:33:45] <Reinhilde> Agnar: and what's obprom
[15:33:56] <Reinhilde> so many words, I'm drowning in words all over again
[15:34:37] <Kurlon> So, UEFI is 'coming' for the HoneyComb, nice looking piece of kit.
[15:36:30] <Kurlon> Oooh, here's a pricy big boi: https://store.avantek.co.uk/ampere-emag-64bit-arm-workstation.html
[15:37:14] <Kurlon> Cavium ThunderX boxes are still out there too.
[15:38:25] <Reinhilde> all I can afford is a sour aggregate fruit 3.14
[15:38:45] <Reinhilde> and the fact that I called it that probably tells you everything you need to know about what I think about it.
[15:40:30] <emfipp> actually
[15:40:39] <emfipp> socionext's got its own arm dev box as well
[15:40:47] <emfipp> and oxalis on 96boards
[15:41:13] <emfipp> the specs though are underwhelming
[15:41:21] <Reinhilde> emfipp: why is a dev board calling itself sorrel
[15:42:00] <emfipp> solidrun's got a dev hanging around in #fedora-arm, btw
[15:42:16] <emfipp> sorrel? which one?
[15:44:10] <emfipp> I also recall some phoronix benchmarks a while back
[15:44:41] <emfipp> where raptor's blackbird (or was it talos?) soundly trump those arm server boards
[15:45:32] <emfipp> that said, SVE is nice, but then very few boards offer it
[15:45:54] * toasterson lookusup raspberry on wikipedia.
[15:45:57] <emfipp> the only arm system I heard actually offering the thing is fujitsu's post-k
[15:46:09] <toasterson> yep it could be described as aggregate fruit
[15:46:49] <emfipp> and this one: https://store.avantek.co.uk/avantek-thunderx2-arm-workstation-thunderx2station.html
[15:46:52] <Reinhilde> emfipp: oxalis = sorrel
[15:46:55] <emfipp> ^ the pricing is crazy
[15:47:00] <Reinhilde> also oww my kidneys
[15:47:46] <emfipp> these are gonna break many a bank
[15:49:03] <toasterson> raports power9 system isn't any better
[15:49:08] <Kurlon> Reinhilde, you're preaching to the choir on Pi... Ebon Upton is Broadcom's salesman of the century for how he moves that crap.
[15:49:11] <toasterson> *raptor's
[15:51:08] <Reinhilde> Kurlon: are there any similarly priced alternatives
[15:51:22] <emfipp> and I don't think avantech lists actual processor specs
[15:51:31] <emfipp> other than demanding outrageous prices
[15:51:42] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[15:52:01] <emfipp> my guess is that they are building it from some gigabyte boards
[15:52:11] <Kurlon> Lots, Hardkernel's ODroid line being the most direct comparison. There are also lots and lots and lots of fruit named Allwinner based boards out there as well.
[15:52:18] <emfipp> the pricing vaguely matches
[15:53:02] <emfipp> I wouldn't trust an allwinner board to run server apps...
[15:53:20] <Kurlon> I wouldn't trust one to tell me the time of day... :D
[15:53:30] <emfipp> and odroid's better off as IoT
[15:53:33] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has joined #illumos
[15:54:02] <emfipp> you should be able to find a lot more from either linuxgizmos.com or cnx-software.com
[15:54:12] <psarria> hi, how can i know in which nic is connected a cable ? i've tried 'dlled' without results in this supermicro motherboard
[15:54:42] <Kurlon> emfipp bear in mind I was offering comparisons to Pi products in this case, not anything server/workstation related.
[15:57:43] <Kurlon> And, full disclosure, I'm doing a 204 package rebuild on my 486's gentoo install so... clearly I'm not a good judge of hardware. :D
[15:57:55] <Reinhilde> Kurlon: you have a 485
[15:57:57] <Reinhilde> 486*
[15:58:09] <Kurlon> A few, old Soekris boxes.
[15:58:23] <Kurlon> My 386 unfortunately died a few years ago and I've not bothered recreating it.
[15:59:37] <Kurlon> They're odd CPUs in these 4801s, classed as 486s but they'll actually handle 686 instructions.
[16:00:35] *** KungFuJesus <KungFuJesus!~adamstyli@> has joined #illumos
[16:00:44] <KungFuJesus> Is zpool trim safe yet on Ollumos?
[16:00:50] <KungFuJesus> Illumos?
[16:01:07] <KungFuJesus> sorry, typo there, clearly I know how to spell illumos, I'm in this channel :)
[16:03:02] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[16:04:19] <Reinhilde> Kurlon: Is it a cyrix?
[16:04:20] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[16:04:28] <Kurlon> NatSemi Geode
[16:06:04] <Reinhilde> so it is a cyrix, kinda
[16:06:06] <Reinhilde> https://en.wikipedia.org/wiki/Geode_(processor)
[16:06:30] <Kurlon> No, the SC1100 / NatSemi version predates the MediaGX line.
[16:07:20] <Kurlon> Scratch that, GX1 so it is MediaGX based.
[16:09:13] <jimklimov> Speaking of ARM, we have SoftIron (Overdrive 3000 server) for linux builds of our product which is quite decent in speed, espesially with an SSD SATA plugged in
[16:09:41] <jimklimov> they are AFAIK decently priced, and had a smaller brother 1000-series for cheaper developer workstations
[16:10:05] <jimklimov> I think the boot is u-Boot based, but really didn't see it for so long after we powered it on and it just ticks, so not sure =)
[16:11:30] <jimklimov> in fact, while it is an aarch64 and our product is armv7l, there was quite a hassle to make it part of the build farm - instruction sets are not quite compatible... in the end, the armv7l farm workers are qemu KVMs on it, claiming to use as much of real CPU abilities as it can
[16:11:44] <jimklimov> but... whatever, it works fast and without known faults :)
[16:12:12] <Kurlon> The Overdrive 3000 is listed as UEFI as of 2015
[16:12:58] <jimklimov> so much the better then? :-)
[16:13:23] <jimklimov> might have different low-level firmwares distributed, though, I guess
[16:14:31] <Kurlon> Yeah, I think most UEFI boards ALSO have a UBoot firmware / mode as well.
[16:17:54] <tsoome> oracles x86 servers definitely do have.
[16:21:17] <Kurlon> Hah, I go searching ebay for 'arm server' and I get a 386 mobo with a $2500 asking price... W T F?!
[16:22:10] <Kurlon> There is a Gigabyte ThunderX 1U on eBay as well.
[16:24:13] <tsoome> arm server — is this like server for an arm and leg?
[16:24:19] <tsoome> :P
[16:24:52] <Agnar> arm is a german word and means "poor"
[16:24:58] <Agnar> there you go.
[16:26:17] *** heroux <heroux!sandroco@gateway/shell/insomnia247/x-hipvyozqxgshedcz> has quit IRC (Ping timeout: 240 seconds)
[16:26:21] <Reinhilde> what
[16:26:29] <tsoome> :D
[16:31:10] <Kurlon> So, the plan is to get a ThunderX server, get Gentoo lit on it, build QEMU -O42 for maximum fast and then boot Illumos AMD64 inside QEMU and claim first to have Illumos on ARM?
[16:33:19] <gitomat> [illumos-gate] 11880 changing encryption key on dataset with unencrypted children triggers VERIFY -- Tom Caputi <tcaputi at datto dot com>
[16:33:20] <jimklimov> well, we could also go the Sun-PCi way :)
[16:33:21] *** heroux <heroux!sandroco@gateway/shell/insomnia247/x-nlqtximtujxgepxk> has joined #illumos
[16:34:06] <jbk> there's work somewhere on an ARM port of illumos
[16:34:28] <jbk> i don't know how far it got
[16:35:03] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[16:35:54] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[16:37:56] <Kurlon> jimklimov: I'm having flashbacks of DOS cards for Apple products!
[16:40:54] <tsoome> Kurlon https://en.wikipedia.org/wiki/SunPCi
[16:41:09] <Kurlon> Hah, already there. :D
[16:45:31] <Reinhilde> Kurlon: I believe there's only up to -O2
[16:46:11] <Kurlon> There is -O3 at least for GCC, possibly -O4... anything above -O2 is considered unsafe generally.
[16:56:45] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[16:57:42] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[16:58:20] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[16:58:52] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[16:59:14] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[17:00:15] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[17:00:40] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[17:01:20] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[17:02:04] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[17:03:30] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[17:04:05] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[17:05:55] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[17:06:22] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[17:06:40] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[17:07:09] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[17:08:26] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[17:11:51] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Read error: Connection reset by peer)
[17:12:29] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[17:23:14] *** clapont <clapont!~clapont@unaffiliated/clapont> has quit IRC (Ping timeout: 240 seconds)
[17:24:44] *** clapont <clapont!~clapont@unaffiliated/clapont> has joined #illumos
[17:44:26] <Reinhilde> interesting thing: VIA's https://en.wikipedia.org/wiki/Alternate_Instruction_Set ... i wonder if anyone's ever written an entire OS to run on this
[17:51:46] <copec> There are these: https://www.asacomputers.com/Cavium-ThunderX.html
[18:04:27] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has quit IRC (Quit: Leaving)
[18:05:14] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has quit IRC (Ping timeout: 240 seconds)
[18:21:23] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[18:25:17] <KungFuJesus> any RISC-V plans for Illumos?
[18:27:49] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has joined #illumos
[18:29:20] <Reinhilde> lowercase i, KungFuJesus, and I don't know, but you can implement it if you have a riscv machine and are experienced with it
[18:31:23] <KungFuJesus> I do not, but I figured I'd ask while we're on the subject of hot new ISAs
[18:31:44] <KungFuJesus> new being less than 30 years old, I guess
[18:32:57] <toasterson> KungFuJesus (IRC): somebodz from sifive offered to port but we never heard from him again unfortunately
[18:33:51] *** man_u <man_u!~manu@manu2.gandi.net> has quit IRC (Ping timeout: 252 seconds)
[18:34:39] <KungFuJesus> some of that ISA is a moving target, though probably not the bits for porting a kernel
[19:24:29] *** pwinder <pwinder!~pwinder@> has quit IRC (Quit: This computer has gone to sleep)
[19:28:16] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Quit: Leaving.)
[19:28:27] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has quit IRC (Ping timeout: 250 seconds)
[19:28:48] *** pwinder <pwinder!~pwinder@> has joined #illumos
[19:33:06] *** pwinder <pwinder!~pwinder@> has quit IRC (Client Quit)
[19:42:20] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has joined #illumos
[19:47:51] *** pwinder <pwinder!~pwinder@> has joined #illumos
[19:54:04] *** pwinder <pwinder!~pwinder@> has quit IRC (Quit: This computer has gone to sleep)
[20:02:21] *** mnowak_ <mnowak_!~mnowak_@> has quit IRC (Quit: Leaving)
[20:02:32] *** mnowak_ <mnowak_!~mnowak_@> has joined #illumos
[20:04:43] <copec> After studying some ISAs I like SPARC. I wish someone would take OpenSPARC from where it left off, and do a run with a modern fab of it
[20:06:33] <andyf> Does anyone happen to have a baseline for what parts of the ZFS testsuite are currently expected to fail?
[20:06:43] <Reinhilde> If someone makes a sun4v computer, donate it to Jeff Sisek
[20:06:44] <andyf> I'm doing a full run and, while it's mostly [PASS], there is the odd [FAIL]
[20:06:52] <Reinhilde> well, not sun4v but sun4v compatible
[20:06:59] <LeftWing> andyf: Jerry and Kody probably have a concrete idea?
[20:07:08] <LeftWing> kkantor is in here, at least
[20:07:16] <jbk> which ones?
[20:07:25] <jbk> i've also been running the tests recently
[20:07:45] <jbk> IIRC, there's one test that requires fio
[20:07:48] <jbk> and will fail without it
[20:08:00] <jbk> the trim tests will fail since we still have it disabled
[20:08:31] <andyf> It's still running (and it's only the slog ones I'm really interested in), but here are the failures so far
[20:08:33] <andyf> https://paste.ec/paste/7U8LrZCZ#J1NfYcdvXWT+J0CMj0iD280Ezd1dp+W7kr1v2+Tektg
[20:09:19] <andyf> As an aside, do you know if TRIM should be safe enough now?
[20:10:11] <jbk> i'd look at the first failure -- i'm guessing it probably got messed up (there's a few tests that seem flaky), and can cascade
[20:11:24] <jbk> one test creates a zpool on top of some files to do some stuff, and IIRC, it can be a bit touchy about cleaning them up
[20:11:45] <andyf> I've seen a few of them come and go in the output from `zpool status`
[20:16:02] <andyf> Yeah, I'm getting some cascading failures as you thought
[20:16:14] <andyf> same for the two new tests I've added.. they run fine by themselves but not in sequence
[20:20:42] <andyf> some of them seem to be waiting for the pool to become degraded, which does not always happen in time
[20:22:27] <LeftWing> andyf: I think the issue with TRIM is that we don't know if it actually works with rando SSDs
[20:22:51] <LeftWing> I seem to recall that Linux, for instance, has a TRIM blacklist in their sd equivalent
[20:23:03] <LeftWing> We probably need to get one of those
[20:24:19] <igork> andyf: ZFS tests https://paste.dilos.org/?e66c00c442a772ce#v8gMWwtrjvHfs1fDmfP619x2U0AheUd/NnzKxywDTm4=
[20:24:34] <igork> but i have many updates from ZoL
[20:24:46] <igork> and additional tests
[20:25:50] <igork> it is log from jenkins job
[20:27:27] <igork> it is additional: https://paste.dilos.org/?fa648ecba7209921#URa7zuJogWDkYTeRaXdbiT9ICsw19JdeZqPZy1k6Pqw=
[20:27:33] <igork> from zloop
[20:29:32] <andyf> igork - I need a baseline from stock illumos really, otherwise it is not a useful comparison.
[20:29:59] <igork> ok, just for info
[20:34:06] <andyf> If anyone does feel able to review https://illumos.org/rb/r/2445/ , I'd appreciate it.
[20:47:43] <richlowe> LeftWing, rmustacc: Did you look at the debug thing I sent? (if I sent it, and didn't dream that)
[20:52:17] <rmustacc> richlowe: Haven't had a chance to yet, been a bit absorbed in some other things.
[20:52:28] <rmustacc> Will try to get to it before EOW.
[20:52:34] <LeftWing> richlowe: Not to the extent necessary to help yet sorry
[20:53:22] <jbk> andyf: i was hoping to use it (trim) with VMs, then i realized we need to support non-legacy virtio first :)
[20:53:36] <andyf> Heh - yes, I want that too
[20:53:43] <jbk> that at least seems like (once supported) it should be fairly safe
[20:53:51] <andyf> for vdev-backed vioblk devices that is
[20:53:58] <andyf> to save me from the dd if=/dev/zero dance
[20:54:04] <rmustacc> Then you need to find a hypervisor that actually offers non-legacy. ;)
[20:54:18] <rmustacc> Most prominent ones are all 'legacy'.
[20:54:40] <rmustacc> Feels like vioscsi should help there, since you'll be able to issue unmap.
[20:54:46] <rmustacc> Rather than retrofitting onto the block device.
[20:55:00] <rmustacc> Similar with nvme emulation.
[20:55:17] <andyf> nvme seems to work reasonably in bhyve fwiw
[20:56:23] <rmustacc> You'd need a bit of blkdev wiring and trim equivalent for nvme, but that and vioscsi are probably better paths.
[21:20:00] *** ed209 <ed209!~ed209@> has quit IRC (Remote host closed the connection)
[21:20:07] *** ed209 <ed209!~ed209@> has joined #illumos
[21:24:00] *** headw01 <headw01!~headw@> has joined #illumos
[21:29:27] *** MarcelT <MarcelT!~marcel@tortuga.telka.sk> has quit IRC (Ping timeout: 264 seconds)
[21:59:57] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has quit IRC (Remote host closed the connection)
[22:00:30] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[22:08:27] <andyf> Go Joyent Automation!
[22:11:32] <LeftWing> Yeah when I looked at the Virtio Block situation I was surprised the extent to which there was all this effort on the standards document getting to 1.1, and in practice hypervisors were ... not really doing any of that
[22:12:46] *** jimklimov <jimklimov!~jimklimov@ip-86-49-243-118.net.upcbroadband.cz> has joined #illumos
[22:12:58] <rmustacc> LeftWing: Wasn't GCP all legacy and the qemu defaults?
[22:13:48] <LeftWing> I thought it was going to be something else but yeah I'm pretty sure it was the legacy thing in the ned
[22:13:50] <LeftWing> *end
[22:16:13] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has quit IRC (Ping timeout: 265 seconds)
[22:17:22] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has quit IRC (Remote host closed the connection)
[22:17:47] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[22:25:48] <tsoome> so the zol does not really follow the cstyle any more?
[22:26:19] <andyf> Oh? That is a shame.
[22:27:10] <tsoome> code like: if (!(flag & DNODE_DRY_RUN)) {
[22:27:17] <tsoome> in your patch:)
[22:27:49] <andyf> oh.. pbchk was clean
[22:28:15] <andyf> but I see..
[22:28:48] <tsoome> I think some gcc will be upset about it
[22:29:10] <andyf> We do have that pattern in a lot of places anyway I think, and there was some discussion on here a few weeks ago
[22:29:38] <jlevon> cstyle says we can't do that?
[22:30:05] <andyf> well, we say only use conditions on booleans..
[22:30:14] <andyf> but I quote:
[22:30:16] <andyf> [2019-09-17T21:11:08+0100] <LeftWing> richlowe: If it's a single bit flag in a status field and you're treating it as a boolean, I'd say the boolean thing applies
[22:30:29] <tsoome> bitwise & is ok with gcc, but combining with logical op like ! sounds like trouble:)
[22:30:37] <jlevon> well
[22:30:43] <jlevon> if (!flag & MY_BIT)
[22:30:45] <jlevon> is trouble :)
[22:30:56] <jlevon> I got gcc to warn about that once...
[22:31:08] <LeftWing> Yeah you need to (!(flag & MY_BIT)) certainly
[22:31:15] <jlevon> if (!()) is unobjectionable imo
[22:31:21] <LeftWing> Right
[22:32:12] <andyf> mind you, there are a couple of bits of illumos cstyle that I don't understand, but I'm lucky that it mostly matches the style I have always used
[22:32:38] <LeftWing> Some of it is history, some of it is always going to be subjective, and some of it isn't frankly strict enough (and we should tighten)
[22:32:59] <LeftWing> There are a bunch of places I think we should demand braces, in particular
[22:33:33] <tsoome> when they will start to use GNU rules for { }, I’ll bite someone:=)
[22:33:45] <LeftWing> lol
[22:33:49] <LeftWing> The weird half-way thing?
[22:33:57] <tsoome> yes
[22:34:03] <LeftWing> Surely nobody seriously thinks GNU C style is the way to go
[22:34:07] <LeftWing> It was always odd
[22:34:10] <tsoome> it is absolutely pain to read
[22:34:22] <andyf> GNU C is opening brace on its own line, isn't it?
[22:34:29] <andyf> (standard, that is)
[22:34:37] <LeftWing> Yes, and half way between the indentation above and the one below
[22:34:44] <tsoome> cstyle
[22:34:56] <jlevon> we don't talk about gnu style thank you
[22:35:02] <andyf> oh yes, I found an example...
[22:35:05] <tsoome> yea
[22:35:50] <tsoome> anyhow, I was walking through the rb an have to decide if I should note those or not:D
[22:36:13] <LeftWing> So there was a lot of talk at the ZFS summit
[22:36:31] <LeftWing> And I think we're going to have to get on a trajectory to be able to ship the OpenZFS code more directly
[22:37:22] <tsoome> makes sense.
[22:37:35] <LeftWing> So if there are style issues up there (and there are, I was looking at some of the code) we'll want to push to improve the cstyle checks they're using basically :P
[22:37:43] <LeftWing> In theory at least they're still using cstyle
[22:37:59] <andyf> tsoome - let's not diverge from ZoL for just style reasons.. we should try and keep close
[22:38:15] <tsoome> yea I figured the same.
[22:38:32] <andyf> LeftWing - that's great news - yes, we really do need to follow FreeBSD's lead if possible here
[22:38:44] <LeftWing> Yeah, I mean, it's ... there will be challenges
[22:38:50] <andyf> and then some..
[22:39:36] <tsoome> tbh, the only way to set things on track is reviews on zol PR's
[22:39:41] <andyf> but every time I look at ZoL, I see features and fixes we don't have
[22:40:28] <LeftWing> I think we'll want to: [1] get the openzfs repo (after renaming from zol) to build on illumos out-of-gate
[22:40:44] <LeftWing> and then [2] look at copying releases wholesale into illumos-gate as they come out
[22:41:56] <andyf> Even looking at the tests, there are a lot of linux assumptions in there that we'll need to work through
[22:42:04] <andyf> but I see they started to add `if is_linux` etc.
[22:42:34] <LeftWing> Yeah I think waiting for the FreeBSD folks to finish their round of restructuring the code is best
[22:42:37] <andyf> although only to toggle between linux and freebsd so far, so I assume they came up from the freebsd folks
[22:42:58] <LeftWing> Yes
[22:43:00] <tsoome> for sure.
[22:43:20] <LeftWing> I had a shot at compiling spa.c the other day haha
[22:54:46] *** snuff-work <snuff-work!~snuff-wor@202-161-112-134.tpgi.com.au> has joined #illumos
[22:55:25] <andyf> Well, I'm interested in helping with that when the time comes, any way that is useful.
[22:58:13] *** jcea1 <jcea1!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has joined #illumos
[22:58:24] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has quit IRC (Read error: Connection reset by peer)
[22:58:24] *** jcea1 is now known as jcea
[23:14:08] <richlowe> well hi there, Joyent Automation.
[23:19:10] *** Kurlon <Kurlon!~Kurlon@cpe-67-253-141-249.rochester.res.rr.com> has joined #illumos
[23:25:43] <LeftWing> Have I missed something :P
[23:27:37] <Reinhilde> is there a distributor's handbook?
[23:30:59] <jlevon> LeftWing: the last of gerrit getting migrated out
[23:31:03] <LeftWing> Ah
[23:35:50] <Reinhilde> i guess if i am writing a distribution i would need to "wing it", so to speak
[23:43:05] *** hemi770_ <hemi770_!~hemi666@unaffiliated/hemi770> has joined #illumos
[23:43:21] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has quit IRC (Quit: Textual IRC Client: www.textualapp.com)
[23:45:38] *** hemi770 <hemi770!~hemi666@unaffiliated/hemi770> has quit IRC (Ping timeout: 276 seconds)
[23:48:26] *** hemi770_ <hemi770_!~hemi666@unaffiliated/hemi770> has quit IRC (Ping timeout: 240 seconds)
[23:49:33] *** hemi770 <hemi770!~hemi666@unaffiliated/hemi770> has joined #illumos

   November 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 | >