Switch to DuckDuckGo Search
   October 15, 2019  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >

Toggle Join/Part | bottom
[00:12:17] <MelMalik> rmustacc: am I meant to ln -s /usr/man/man1m /usr/man/man8 ?
[00:28:17] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has quit IRC (Remote host closed the connection)
[00:28:37] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[00:47:51] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has quit IRC (Remote host closed the connection)
[00:51:09] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[01:04:19] <LeftWing> I don't think that would be wise
[01:05:32] *** heroux <heroux!sandroco@gateway/shell/insomnia247/x-ijslxaywhrdcvync> has quit IRC (Read error: Connection reset by peer)
[01:06:06] *** heroux <heroux!sandroco@gateway/shell/insomnia247/x-npcepdwkfdthaoql> has joined #illumos
[01:12:06] <MelMalik> LeftWing: me neither
[01:13:09] <MelMalik> also unwise: trying to formulate an attributes section for something that is not a component of illumos, while not knowing anything about how the attributes sections are actually formulated.
[01:13:13] <MelMalik> or why they exist
[01:20:11] <LeftWing> https://illumos.org/man/5/attributes -- You can at least read about those!
[01:20:14] *** wl_ <wl_!~wl_@2605:6000:1b0c:4e77::87c> has joined #illumos
[01:20:37] <MelMalik> LeftWing: yes.
[01:24:24] <MelMalik> If a piece of software is extremely cross-processor, cross-platform, cross-everything, is it appropriate to put "Any where sizeof(int) > 4" (the sole condition on which the application depends, other than POSIX compliance) in the "Architecture" field of an attributes table?
[01:24:53] <MelMalik> assuming that attributes tables are even meant to be documented by non-core manual pages
[01:32:37] <LeftWing> I'm not sure I would use an attributes table for things outside the OS to be honest.
[01:34:16] <MelMalik> LeftWing: I'm not either, but I'll do one anyway just because I am already writing a manual page to replace the old one.
[01:35:02] <MelMalik> I figure, if it ends up being a gimmick, it ends up being a gimmick
[02:12:27] *** fireglow- <fireglow-!fireglow@fireglow.su> has joined #illumos
[02:12:27] *** fireglow <fireglow!fireglow@fireglow.su> has quit IRC (Ping timeout: 250 seconds)
[02:13:51] *** fireglow- is now known as fireglow
[02:16:56] *** MarcelT <MarcelT!~marcel@tortuga.telka.sk> has quit IRC (Ping timeout: 240 seconds)
[02:26:18] *** jimklimov <jimklimov!~jimklimov@ip-86-49-243-194.net.upcbroadband.cz> has joined #illumos
[02:40:07] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has quit IRC (Remote host closed the connection)
[02:50:24] *** atrius <atrius!~atrius@2600:1700:22d1:518:6d1a:d1e9:cac3:e490> has quit IRC (Ping timeout: 246 seconds)
[02:50:59] *** atrius <atrius!~atrius@99-160-139-232.lightspeed.nsvltn.sbcglobal.net> has joined #illumos
[03:11:14] *** Yogurt <Yogurt!~Yogurt@c-73-189-45-147.hsd1.ca.comcast.net> has joined #illumos
[03:15:17] *** Yogurt <Yogurt!~Yogurt@c-73-189-45-147.hsd1.ca.comcast.net> has quit IRC (Ping timeout: 240 seconds)
[03:20:00] *** ed209 <ed209!~ed209@165.225.128.67> has quit IRC (Remote host closed the connection)
[03:20:07] *** ed209 <ed209!~ed209@165.225.128.67> has joined #illumos
[03:25:25] *** Kurlon_ <Kurlon_!~Kurlon@cpe-67-253-141-249.rochester.res.rr.com> has quit IRC (Remote host closed the connection)
[03:26:00] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has joined #illumos
[04:04:22] *** Kurlon_ <Kurlon_!~Kurlon@cpe-67-253-141-249.rochester.res.rr.com> has joined #illumos
[04:07:14] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has quit IRC (Ping timeout: 240 seconds)
[05:29:22] *** catalinbostan <catalinbostan!~textual@84.232.132.80> has joined #illumos
[05:30:04] *** catalinbostan <catalinbostan!~textual@84.232.132.80> has quit IRC (Read error: Connection reset by peer)
[06:31:33] *** MarcelT <MarcelT!~marcel@tortuga.telka.sk> has joined #illumos
[06:57:25] *** trinityhex <trinityhex!~puppy@ip70-188-0-187.rn.hr.cox.net> has joined #illumos
[07:10:21] *** trinityhex <trinityhex!~puppy@ip70-188-0-187.rn.hr.cox.net> has quit IRC (Quit: Leaving)
[07:20:46] *** nikolam <nikolam!~Nikolam@unaffiliated/nikolam> has joined #illumos
[07:21:27] <nikolam> I were always under impression that 'sudo' is the problem for itself (instead of RBAC/pfexec). And I was right: https://thehackernews.com/2019/10/linux-sudo-run-as-root-flaw.html
[07:29:44] <LeftWing> Eh. All software has bugs. I would not assume there are zero bugs in RBAC.
[07:34:40] <nikolam> Reminds me of telling to someone, not to rely solely on SSL for authentication, just before some released info on Debian bug ,creating SSL keys, were full of zeros for years.
[07:39:10] *** slx86 <slx86!~Thunderbi@185.98.144.22> has joined #illumos
[07:40:37] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Quit: tsoome)
[07:42:50] <LeftWing> It is certainly a good idea to have a second factor for important things where you can
[07:57:47] *** nbhauke <nbhauke!~hauke@p5DEEF3E8.dip0.t-ipconnect.de> has joined #illumos
[08:03:11] *** tsoome_ <tsoome_!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[08:14:50] *** jimklimov <jimklimov!~jimklimov@ip-86-49-243-194.net.upcbroadband.cz> has quit IRC (Quit: Leaving.)
[08:16:24] <leoric> nikolam: well, it is rather corner case, you should allow user to execute something as anybody besides root... This seems to be uncommon.
[08:24:19] <nikolam> leoric, I was never persuaded that entering the same password twice with sudo, is somehow "more secure" then entering other password, once while logging in as admin user and doing administrative tasks. RBAC were better all the time, maybe missing "feature" of entering same password twice, that is surely large waste of time.
[08:25:40] <leoric> rbac for usual user is worse for the same reason - it allows anything just run with elevated privileges without user being asked or even notified about this
[08:26:25] <leoric> if you don't want sudo to ask password, just configure it as needed
[08:26:46] <leoric> it can be required, for example, for non-interactive applications
[08:27:27] <leoric> but illumos 'Primary Administrator' role is a security nightmare
[08:28:14] <leoric> (if it's granted to regular user, not role)
[08:28:18] <nikolam> having separate account for user and separate account for admin is the best practice (and not counting root)
[08:28:42] <nikolam> Why it's not default during install is beyond me.
[08:29:00] <leoric> what exactly? one more root role?
[08:29:06] <nikolam> sudoer is de facto root user. so default user account during install should not be sudoer
[08:29:31] <tsoome_> the root issue about setuid/setgid is that if you have huge application, you can not make sure it does all correctly related to those setui/setgid attributes. historically the approach has been to perform all privileged stuff as early as possible, so the “blast” radius would be minimal. sudo does “fix” this by creating huge complicated application…
[08:29:41] <nikolam> 2 accounts during install to be created, now just one.
[08:30:11] *** nbhauke <nbhauke!~hauke@p5DEEF3E8.dip0.t-ipconnect.de> has quit IRC (Quit: nbhauke)
[08:38:36] <leoric> tsoome_: Well, not exactly. setuid-aware application will usually drop privileges when not needed, sudo just makes it run as root.
[08:39:11] <leoric> smf privileges is much better way to fix it
[08:39:29] <leoric> most network applications need root just to open privileged port
[08:42:05] <tsoome_> yes, but the point is, you drop them as soon as possible. su/sudo obviously can not drop them, and the issue is, sudo must do much work before it is to execute the actual application.
[08:45:14] <nikolam> Maybe the best for interactive work is not to use sudo at all, but to nicely log in with shell as another, privileged user, just to do admin tasks with pfexec. And use that admin account just for that.
[08:47:30] <tsoome_> am, well, rbac is role based access control, that privileged user is called role, and does not need pfexec because it is using shell with pfexec built in;)
[08:48:31] <nikolam> yes, but user needs separate password and that is better then sudo with the same password. Role is ok but for priviledged user.
[08:48:54] <tsoome_> (you assign role by using su, and you only can do that when you have role listed at your account - usermod -R).
[08:50:21] <nikolam> Yes, I think default user account should not have that role. But to create another account with the role, you can access only with separate password.
[08:50:57] <tsoome_> note that role has password. Role in that sense is that “different user”, even better, because role can not login.
[08:51:45] <tsoome_> you better read up what the role is:)
[08:51:52] *** tsoome_ <tsoome_!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Quit: tsoome_)
[08:51:56] *** slx86 <slx86!~Thunderbi@185.98.144.22> has quit IRC (Quit: slx86)
[08:52:31] <nikolam> That would mean sudo/sudoer is not needed at all, what was the starting point.
[09:01:03] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has joined #illumos
[09:12:23] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #illumos
[09:16:43] *** nikolam <nikolam!~Nikolam@unaffiliated/nikolam> has quit IRC (Quit: Leaving)
[09:26:53] <MelMalik> leoric: ircds don't use root but only because they don't conform to the standard
[09:27:26] *** nbhauke <nbhauke!~hauke@Dachstein.nt.e-technik.tu-darmstadt.de> has joined #illumos
[09:27:57] <MelMalik> leoric: also, should the shell of primary administrator's accounts be showing # instead of %?
[09:31:47] <MelMalik> and i just realised i run ircd as a PA on my illumos box.
[09:40:59] *** andy_js <andy_js!~andy@97e29e78.skybroadband.com> has joined #illumos
[09:41:00] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[09:58:18] *** jimklimov1 <jimklimov1!~jimklimov@31.7.243.238> has joined #illumos
[10:00:39] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Ping timeout: 264 seconds)
[10:01:20] *** man_u <man_u!~manu@89-92-76-134.hfc.dyn.abo.bbox.fr> has joined #illumos
[11:09:30] *** pcd <pcd!~pcd@openzfs/developer/pcd> has quit IRC (Ping timeout: 268 seconds)
[11:15:38] *** joltman <joltman!~znc@unaffiliated/joltman> has quit IRC (Ping timeout: 265 seconds)
[11:21:17] *** joltman <joltman!znc@gateway/vpn/privateinternetaccess/joltman> has joined #illumos
[11:39:31] *** jimklimov1 <jimklimov1!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[11:40:05] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #illumos
[11:43:14] *** merzo <merzo!~merzo@34-7-133-95.pool.ukrtel.net> has quit IRC (Ping timeout: 240 seconds)
[11:48:19] *** N1lux <N1lux!~Thunderbi@2a01:cb18:7f8:8900:6600:6aff:fe6b:cf62> has joined #illumos
[11:48:27] *** N1lux <N1lux!~Thunderbi@2a01:cb18:7f8:8900:6600:6aff:fe6b:cf62> has left #illumos
[11:52:10] *** wiedi_ <wiedi_!~wiedi@ip5b4096a6.dynamic.kabel-deutschland.de> has quit IRC (Read error: Connection reset by peer)
[11:52:46] *** wiedi <wiedi!~wiedi@ip5b4096a6.dynamic.kabel-deutschland.de> has joined #illumos
[11:58:53] *** merzo <merzo!~merzo@108-20-132-95.pool.ukrtel.net> has joined #illumos
[12:20:00] *** ed209 <ed209!~ed209@165.225.128.67> has quit IRC (Remote host closed the connection)
[12:20:07] *** ed209 <ed209!~ed209@165.225.128.67> has joined #illumos
[12:30:06] *** pcd <pcd!~pcd@openzfs/developer/pcd> has joined #illumos
[12:43:55] *** man_u <man_u!~manu@89-92-76-134.hfc.dyn.abo.bbox.fr> has quit IRC (Quit: man_u)
[13:28:19] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[13:29:26] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #illumos
[13:37:40] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[13:38:10] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #illumos
[13:51:00] *** rsmarples <rsmarples!~roy@xenity.marples.name> has joined #illumos
[13:51:06] <rsmarples> hello
[13:51:35] <rsmarples> I'm trying to find something to change the process title on Illumos - like setproctitle from the BSD familty
[13:51:59] <rsmarples> stamping on argv[0] doesn't seem to work in some examples I have found
[13:52:04] <rsmarples> are there any other methods?
[13:57:37] *** wiedi <wiedi!~wiedi@ip5b4096a6.dynamic.kabel-deutschland.de> has quit IRC (Quit: ^C)
[14:02:27] *** wiedi <wiedi!~wiedi@ip5b4096a6.dynamic.kabel-deutschland.de> has joined #illumos
[14:11:52] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has joined #illumos
[14:16:45] <jlevon> rsmarples: what do you mean by work?
[14:17:18] <igork> rsmarples: i think you have to wait couple of hour when more illumos devs will be online and try to ask again
[14:17:42] <igork> probably jbk can help
[14:22:49] *** tozhu <tozhu!~tozhu@101.207.125.86> has joined #illumos
[14:24:43] <tozhu> hello all, is there anyone know vrrp internal on Solaris? seems system detect cycle is 1s, can we reduce it to 50ms or 100ms? where is the code to control the checking interval ? thanks in advance
[14:26:24] <toasterson1> neirac any non 0 or 1 error code comes from the utility inside the zone or su
[14:26:24] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[14:27:04] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #illumos
[14:28:34] *** toasterson1 <toasterson1!~toastmatr@85-195-235-249.fiber7.init7.net> has quit IRC (Remote host closed the connection)
[14:28:49] *** toasterson <toasterson!~toastmatr@85-195-235-249.fiber7.init7.net> has joined #illumos
[14:30:02] *** toasterson <toasterson!~toastmatr@85-195-235-249.fiber7.init7.net> has quit IRC (Remote host closed the connection)
[14:30:15] *** toasterson <toasterson!~toastmatr@85-195-235-249.fiber7.init7.net> has joined #illumos
[14:31:19] *** toasterson <toasterson!~toastmatr@85-195-235-249.fiber7.init7.net> has quit IRC (Remote host closed the connection)
[14:31:37] *** toasterson <toasterson!~toastmatr@85-195-235-249.fiber7.init7.net> has joined #illumos
[14:33:04] * toasterson now has ipv6 again :)
[14:33:21] <toasterson> and updating from pkg.openindiana.org broke :(
[14:33:32] *** tozhu <tozhu!~tozhu@101.207.125.86> has quit IRC (Quit: tozhu)
[14:33:41] <toasterson> nginx does not like being contacted from ipv6
[14:39:46] *** hadfl <hadfl!~hadfl@omniosce.ee.ethz.ch> has quit IRC (Quit: nn o/)
[14:39:59] *** hadfl_ <hadfl_!~hadfl@omniosce.ee.ethz.ch> has joined #illumos
[14:50:29] <toasterson> interestingly pkg.openindiana.org does not even have an ipv6 address. on which nginx do I land 0.o
[15:12:14] <rsmarples> jlevon: as in the ps output isn't affected
[15:12:36] <rsmarples> jlevon: source - http://lxr.nginx.org/source/src/os/unix/ngx_setproctitle.c
[15:12:53] <rsmarples> (not that i tried nginx specifically, i just used that example)
[15:13:07] <jlevon> rsmarples: currently illumos ps purposefully ignores current argv. I have changes to fix that but they were too controversial.
[15:13:37] <rsmarples> jlevon: ok, that's good information thanks
[15:15:35] *** neirac <neirac!~cneira@190.162.109.190> has joined #illumos
[15:21:00] <KungFuJesus> anyone have any ideas on this one? https://www.illumos.org/issues/11815
[15:32:54] <jimklimov> Got a question to fellow IT experts stemming from a friendly argument: are there public/private key schemes where neither key can be calculated from the other (e.g. some algorithm generates them so they are known to match, and that's it), and if yes - how are they called?
[15:33:20] *** jamtorus <jamtorus!~quassel@s91904425.blix.com> has joined #illumos
[15:34:56] *** jellydonut <jellydonut!~quassel@s91904425.blix.com> has quit IRC (Ping timeout: 240 seconds)
[15:41:43] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Quit: tsoome)
[15:46:03] <toasterson> jimklimov (IRC): isn't the point of public key crypto that you can alwys calculate the public key? is tat not the base of their existence?
[15:50:31] <andyf> rsmarples - run `ps auxww` to see the current arguments (versus startup ones)
[15:56:58] <jimklimov> I think I've read someplace recently, but could not find it quickly hence the argument, that the requirement is that data encrypted by one key can be verified by the other
[15:57:31] <jimklimov> and that while many algorithms go a step further to make public key calculatable from the private, it is not a "spec requirement" and some algos do not
[15:59:01] *** neirac <neirac!~cneira@190.162.109.190> has quit IRC (Quit: Leaving)
[16:21:31] *** v_a_b <v_a_b!~volker@p57A272B1.dip0.t-ipconnect.de> has quit IRC (Quit: Leaving.)
[16:22:03] *** tozhu <tozhu!~tozhu@171.210.169.78> has joined #illumos
[16:36:04] *** tozhu <tozhu!~tozhu@171.210.169.78> has quit IRC (Quit: tozhu)
[16:38:33] *** jamtorus is now known as jellydonut
[16:39:42] *** tozhu <tozhu!~tozhu@171.210.169.78> has joined #illumos
[16:48:23] *** rzezeski <rzezeski!uid151901@gateway/web/irccloud.com/x-cvlkuxbimqbxayoi> has joined #illumos
[17:30:06] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has quit IRC (Quit: Leaving)
[17:34:18] *** catalinbostan <catalinbostan!~textual@84.232.132.80> has joined #illumos
[17:34:55] *** tozhu <tozhu!~tozhu@171.210.169.78> has quit IRC (Ping timeout: 268 seconds)
[18:07:39] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[18:30:58] *** nbhauke <nbhauke!~hauke@Dachstein.nt.e-technik.tu-darmstadt.de> has quit IRC (Quit: nbhauke)
[18:31:53] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[19:24:31] *** jimklimov1 <jimklimov1!~jimklimov@31.7.243.238> has joined #illumos
[19:25:37] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Ping timeout: 240 seconds)
[19:47:55] *** catalinbostan <catalinbostan!~textual@84.232.132.80> has quit IRC (Read error: Connection reset by peer)
[19:52:07] *** Kurlon_ <Kurlon_!~Kurlon@cpe-67-253-141-249.rochester.res.rr.com> has quit IRC (Remote host closed the connection)
[19:52:43] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has joined #illumos
[20:00:54] <wilbury> igork: are you in zfs leadership meeting, too? :-)
[20:01:15] <igork> yes :)
[20:01:32] <igork> today is good time to me
[20:09:37] *** freakazoid0223 <freakazoid0223!~matt@pool-96-227-98-169.phlapa.fios.verizon.net> has joined #illumos
[20:51:53] *** nbhauke <nbhauke!~hauke@p5DEEF3E8.dip0.t-ipconnect.de> has joined #illumos
[20:53:28] *** Bryanstein <Bryanstein!~Bryanstei@shellium/admin/bryanstein> has quit IRC (Ping timeout: 246 seconds)
[21:01:34] *** rsmarples <rsmarples!~roy@xenity.marples.name> has left #illumos ("WeeChat 2.5")
[21:07:54] *** Bryanstein <Bryanstein!~Bryanstei@shellium/admin/bryanstein> has joined #illumos
[21:10:41] *** amrmesh <amrmesh!~Thunderbi@109.201.133.238> has joined #illumos
[21:13:02] *** amrfrsh <amrfrsh!~Thunderbi@185.212.171.68> has quit IRC (Ping timeout: 265 seconds)
[21:13:02] *** amrmesh is now known as amrfrsh
[21:18:44] *** jubal <jubal!~jubal@199-6-237-24.gci.net> has quit IRC (Quit: jubal)
[21:50:56] *** lgtaube <lgtaube!~lgt@84.16.224.13> has quit IRC (Ping timeout: 240 seconds)
[21:52:12] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[21:57:37] *** nbhauke <nbhauke!~hauke@p5DEEF3E8.dip0.t-ipconnect.de> has quit IRC (Quit: nbhauke)
[21:57:57] *** MarcelT <MarcelT!~marcel@tortuga.telka.sk> has quit IRC (Ping timeout: 240 seconds)
[22:06:08] *** nbhauke <nbhauke!~hauke@p5DEEF3E8.dip0.t-ipconnect.de> has joined #illumos
[22:10:49] *** nbhauke <nbhauke!~hauke@p5DEEF3E8.dip0.t-ipconnect.de> has quit IRC (Client Quit)
[22:20:00] *** ed209 <ed209!~ed209@165.225.128.67> has quit IRC (Remote host closed the connection)
[22:20:07] *** ed209 <ed209!~ed209@165.225.128.67> has joined #illumos
[22:23:52] <copec> So (ignorant thinking out loud), Is there a synchronization problem with virtual CPUs similar to green threads on top of kernel threads?
[22:26:07] <copec> If you have hardware, and want each VM to have access to all the CPU resources, so you set them each up with the same number of vcpus as real cpus, it seems like there would be a lot of wait for each VM's vcpu threads to all reach the synchronization sections
[22:28:41] <_mjg> you don't need to reach to "green threads" to have this problem
[22:28:50] <_mjg> it already exists with 1:1 mapping
[22:29:13] <_mjg> that said, the problem is well known and is mostly taken care of in things like kvm
[22:29:24] <_mjg> where vcpus contending on locks can tell the hypervisor which vcpu they are waiting on
[22:29:57] <_mjg> to theb est of my knowledge it is not taken care of in bhyve nor kvm as shipped with smartos (too old)
[22:30:18] <_mjg> also neither bsd or illumos vms make use of the feature even if present
[22:37:05] *** lgtaube <lgtaube!~lgt@46.165.203.45> has joined #illumos
[22:37:29] *** danmcd <danmcd!~danmcd@static-71-174-113-16.bstnma.fios.verizon.net> has quit IRC (Ping timeout: 250 seconds)
[22:41:20] <copec> interesting
[22:43:40] <copec> I guess what I meant was "like green threads", where there are concurrency problems abstracted from the underlying model
[22:46:36] *** toasterson <toasterson!~toastmatr@85-195-235-249.fiber7.init7.net> has quit IRC (Read error: Connection reset by peer)
[22:49:54] <copec> _mjg what have you studied on this problem? Could you point me towards anything you found interesting?
[22:52:45] *** CaptainTobin <CaptainTobin!~tobin@c-68-38-10-41.hsd1.in.comcast.net> has joined #illumos
[22:55:05] <jbk> is there any actual documentation on such an interface? or is it just something undocumented kvm does w/ linux?
[22:55:39] *** Kurlon_ <Kurlon_!~Kurlon@cpe-67-253-141-249.rochester.res.rr.com> has joined #illumos
[22:58:20] <_mjg> copec: for the problem in general, with userspace threadingi n particular, i think futex papers are great
[22:58:57] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has quit IRC (Ping timeout: 240 seconds)
[22:58:58] <_mjg> jbk: i'm afraid you willh ave to ask kvm people, i suspect /worst case/ they can provide some so thaty ou don't have look at the code
[23:01:25] *** hadfl_ <hadfl_!~hadfl@omniosce.ee.ethz.ch> has quit IRC (Quit: nn o/)
[23:01:42] *** hadfl <hadfl!~hadfl@omniosce.ee.ethz.ch> has joined #illumos
[23:02:07] <CaptainTobin> Is there a seperate openindiana channel or do you guys just share this one?
[23:02:32] <LeftWing> There are at least two: #openindiana (for users, I guess?) and #oi-dev which is more for people working on OI
[23:02:56] <CaptainTobin> I see
[23:02:59] <CaptainTobin> Thank you
[23:03:07] <LeftWing> Though if you have a question that is effectively about the core illumos bits, here is also a fine place to ask!
[23:03:31] <CaptainTobin> I'll stick around.
[23:04:02] <CaptainTobin> I do actually one one question.. are either runnable in kvm?
[23:04:29] <jbk> _mjg: it's more about making sure any support is implemented correctly
[23:04:33] <LeftWing> Certainly! There are people regularly using various illumos things under QEMU/KVM on Linux desktops, and on SmartOS servers, and other places
[23:05:06] <CaptainTobin> any special bits I need to know to get it a-working? jbk
[23:06:59] *** dgtlcmo <dgtlcmo!~dgtlcmo@c-24-61-203-74.hsd1.ma.comcast.net> has joined #illumos
[23:07:35] <LeftWing> jlevon: I don't think your changes to ps are too controversial. The distribution runners are just trying to get a handle on what's actually happening. I'm going to improve the descriptions in the bugs, and we might want to split up the mdb-specific bits and the ptools bits.
[23:08:10] <CaptainTobin> sorry i meant LeftWing
[23:08:51] <LeftWing> CaptainTobin: What is the host operating system?
[23:09:28] <CaptainTobin> RHEL 8 so qemu 2.something but I also have qemu trunk.. master w/e
[23:09:50] <LeftWing> I'd recommend using VNC instead of SPICE, if possible, for now.
[23:09:58] <LeftWing> (for the emulated display)
[23:10:26] <LeftWing> There was an issue we had with SPICE which has been fixed, but I'm not sure if it's made it into every distribution yet -- https://www.illumos.org/issues/11501
[23:10:59] <LeftWing> I'd probably also try to use BIOS boot (aka CSM, legacy, etc) instead of UEFI for the time being. Folks are, I believe, still finishing up installer support for UEFI.
[23:11:17] <LeftWing> Other than that, I'm not aware of any speed bumps! We have drivers for Virtio block and network devices.
[23:11:29] <LeftWing> Let us know if you hit any issues!
[23:11:45] <CaptainTobin> well that's good at least
[23:12:13] <CaptainTobin> does it have ga for clipboard sharing?
[23:12:36] <LeftWing> err
[23:12:42] <LeftWing> I do not know!
[23:12:49] <CaptainTobin> guess I will find out
[23:12:50] <CaptainTobin> heh
[23:12:57] <CaptainTobin> I shall let you know soonish
[23:13:21] <LeftWing> I've only really used things like clipboard sharing on VMware Workstation/Fusion, and that required special software installed in the guest. I don't know if/how that works with QEMU/KVM.
[23:13:23] <gitomat> [illumos-gate] 11715 Add MD5 tests to crypto test suite -- Jason King <jason.king at joyent dot com>
[23:16:25] <CaptainTobin> LeftWing: I am looking forward to finding time to give it a try.. I thought solaris was long dead but I was always interested in it when I was younger
[23:16:38] <LeftWing> Well, we're a long way from Solaris here haha
[23:18:01] <CaptainTobin> well evolution, getting away from oracle's stuff (seems to be a huge trend in everything they get their hands on), and such
[23:19:17] <CaptainTobin> so I hope you didn't take it as jab or something
[23:20:35] <jlevon> LeftWing: yeah I'm going to split out the non-ps bits anyway.
[23:20:55] <LeftWing> jlevon: I think that would help!
[23:21:45] <LeftWing> CaptainTobin: It's all good, we've just been on a decade long mission to make sure people realise we're not compatible with Solaris 11
[23:23:15] <CaptainTobin> I see
[23:23:46] <LeftWing> jlevon: It seems like 11689/11690 go together, 11691 could be separate, and then 11692 is obviously the one that people are paying the attention to
[23:24:28] <LeftWing> I think?
[23:27:23] <jlevon> going to separate out all of them
[23:28:18] <LeftWing> Alright!
[23:28:40] <LeftWing> I appreciate it, thank you!
[23:29:12] <LeftWing> Sorry I've been a bit slow to act, I'm trying to get my immigration stuff sorted out
[23:29:42] <jlevon> np
top

   October 15, 2019  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >