July 3, 2012 
< | 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 | >
July 3, 2012 [00:04:43] *** _alram has quit IRC[00:07:23] *** alfism has quit IRC[00:15:48] *** enmand has joined #smartos[00:19:12] *** neophenix has quit IRC[00:19:21] <Licenser> is it natural (or wanted) for SmartOs that it keeps hogging memory for ARC in a insane way?[00:20:08] <ryancnelson> it's not hogging it.[00:20:23] <ryancnelson> it will give it back if it's needed.[00:20:57] <ryancnelson> arc will use most memory that's otherwise sitting idle to improve disk perf[00:21:06] <ryancnelson> are you seeing it not give back memory?[00:22:15] *** reiddraper has quit IRC[00:23:41] <rmustacc> Licenser: Is something wanting memory and not getting it?[00:24:01] *** enmand has quit IRC[00:24:15] <Licenser> no not yet but I'm at 30/32 gb of wich 21 are ARC[00:24:35] <rmustacc> So, would you rather it was unused?[00:24:37] <Licenser> and when I remember that meanted I can't provision a machine with say 5GB right?[00:24:43] <rmustacc> No.[00:24:56] <rmustacc> The arc is merely a cache.[00:25:02] <rmustacc> If something else needs that memory, it gets released.[00:25:33] <Kush-> How exactly should a 2nd physical interface be configured in /usbconfig/config file?[00:25:48] <Kush-> I added something like this to that file:[00:25:48] <Kush-> be_nic=0:23:8b:8a:4c:e8[00:25:48] <Kush-> be_ip=10.64.61.35[00:25:49] <Kush-> be_netmask=255.0.0.0[00:26:01] <Kush-> (just those 3 lines, but that doesn't seem to get picked up after a reboot)[00:26:47] <ryancnelson> first, unless you want the global zone to have a plumbed-up interface on that nic, you only need the first line[00:27:05] <Kush-> I did get a little luck if I change be_ip and be_netmask to be: be0_ip and be0_netmask, leaving be_nic as-is. It seemed to have named the interface be0.[00:27:05] <ryancnelson> second, see http://wiki.smartos.org/display/DOC/Managing+NICs[00:27:26] <ryancnelson> the example there, the name is "storage"[00:27:49] <Kush-> I saw that, but I need to vnic it to the e1000g0!?[00:28:02] <ryancnelson> and you need to say "be0_nic" and "be0_netmask"[00:28:21] <Kush-> That wiki needs updating, I was seeing "storage_nic=0:c:29:18:ec:10[00:28:24] <Kush-> =)[00:28:33] <ryancnelson> "vnic it to the d1000g0" doesn't make any sense.[00:28:38] <ryancnelson> and i meant what i said.[00:28:48] <ryancnelson> storage_nic is correct[00:28:57] <ryancnelson> the nic definition doesn't need a number after it[00:29:03] <ryancnelson> the layer3 info does[00:29:36] <ryancnelson> what's on the page is correct[00:30:31] *** hij1nx has quit IRC[00:30:42] <Kush-> Yea, I did this earlier... but it wanted to name the new interface be0. This isn't really a problem, but I was wondering if it's possible name it using e1000gX[00:31:12] <ryancnelson> in the global zone?[00:31:15] <Kush-> yes[00:31:29] <ryancnelson> only if you name your nic_tag "e1000g" … which is crazy[00:33:52] <ryancnelson> i mean, you can manually create a vnic named something like that, bound to some other physical interface, but that's just UGLY[00:36:20] <Kush-> That example on the page, is creating a storage0 vnic over e1000g0 interface. I wanted create a interface distinct from e1000g0, a 2nd one: e1000g2[00:36:37] <Kush-> [root@00-23-8b-8a-4c-e6 ~]# dladm show-link[00:36:37] <Kush-> LINK CLASS MTU STATE BRIDGE OVER[00:36:37] <Kush-> e1000g0 phys 1500 up -- --[00:36:37] <Kush-> e1000g2 phys 1500 up -- --[00:36:37] <Kush-> e1000g1 phys 1500 unknown -- --[00:36:37] <Kush-> e1000g3 phys 1500 unknown -- --[00:36:38] <Kush-> be0 vnic 1500 up -- e1000g0[00:37:48] *** knweiss has quit IRC[00:37:54] <ryancnelson> e1000gX is a physical nic … be0 is a vnic[00:38:13] <ryancnelson> e1000g2 you don't need to create. just plumb it up[00:38:18] <Kush-> http://pastebin.com/DtzVaVeu[00:38:38] <Kush-> currently, e1000g2 is the *primary*[00:38:57] <Kush-> I want to plumb a new interface, e1000g0 (dedicated phys interface)[00:39:18] <Kush-> I can do it w/o problem using ifconfig plumb, but I need something that is persistent at boot, so how do I configure this in /usbkey/config file?[00:39:33] <ryancnelson> so plug it in, and plumb it up with a service that starts at boot[00:40:13] <Kush-> ok, that's what I figured as much. are there specific rc scripts that get called at boot where I can do this type of thing?[00:40:47] <ryancnelson> a smf service xml placed in /opt/custom/smf/whatever.xml will be imported at boot[00:40:54] <ryancnelson> … so those are persistent[00:41:17] <Kush-> yes[00:41:20] <ryancnelson> see http://www.psychicfriends.net/blog/archives/2012/03/21/smartosorg_run_things_at_boot.html[00:41:41] <Kush-> that's what I need, thanks.[00:44:48] *** enmand has joined #smartos[00:46:17] *** Licenser has quit IRC[00:46:25] <Kush-> Is there a reason smtp:sendmail service is enabled by default?[00:47:00] <Kush-> (I see it's configured to run on localhost though... so I guess it isn't a big deal)[00:48:38] *** Licenser has joined #smartos[01:10:06] *** bixu has joined #smartos[01:12:10] *** ryancnelson has quit IRC[01:19:14] *** marsell has joined #smartos[01:19:48] *** darjeeling has quit IRC[01:21:52] <konobi> Kush-: what's wrong with using a vnic?[01:25:49] *** supr has joined #smartos[02:06:52] *** darjeeling has joined #smartos[02:27:41] *** ryancnelson has joined #smartos[02:46:03] *** deirdres has quit IRC[02:49:10] *** jdavis has quit IRC[02:52:13] *** ryancnelson has quit IRC[03:07:48] *** relling has joined #smartos[03:36:30] *** banjiewen is now known as banjiewen|away[03:53:10] *** Teknix has joined #smartos[04:16:07] *** relling has quit IRC[04:17:56] *** relling has joined #smartos[04:22:10] *** codingstream has joined #smartos[04:28:31] *** banjiewen|away is now known as banjiewen[04:49:03] *** banjiewen is now known as banjiewen|away[04:52:35] *** reiddraper has joined #smartos[04:56:37] *** enmand has quit IRC[04:59:57] *** hij1nx has joined #smartos[05:14:36] *** Teknix has quit IRC[05:15:56] *** Teknix has joined #smartos[05:16:04] *** dap has joined #smartos[05:30:03] *** denizr has quit IRC[05:32:24] *** dap has quit IRC[05:34:58] *** relling has quit IRC[05:35:19] *** denizr has joined #smartos[05:36:16] *** denizr has quit IRC[05:37:24] *** denizr has joined #smartos[05:43:22] <scanf> Licenser: latest screenshot of fifo? :)[05:50:25] *** denizr has quit IRC[05:50:56] *** denizr has joined #smartos[06:00:23] *** sam_wan has joined #smartos[06:18:00] *** Teknix has quit IRC[06:19:54] *** Teknix has joined #smartos[06:20:48] *** jamesd__ has quit IRC[06:21:09] *** denizr has quit IRC[06:55:25] *** codingstream has quit IRC[07:53:21] <Licenser> scanf https://www.evernote.com/shard/s4/sh/7277e252-4bbc-4fef-89f9-e77ad627893b/0250d9ac0fd9c0ae9403c4623e198ca2[07:55:44] <MerlinDMC> Licenser, i don't like the gauges![07:55:44] <Licenser> MerlinDMC :([07:56:12] <MerlinDMC> but at least ... if it's just a single host it should be usable[07:56:47] <Licenser> I've to go for breakfast but Im all ears for ideas :)[07:57:10] <MerlinDMC> i have cookies and coke for breakfast -.-[07:58:01] <Licenser> I'll have an omlet[07:58:30] <MerlinDMC> already at the office ... no eggs here[07:58:47] <Licenser> heh hea in the hotel they make them for me :)[08:30:31] *** bixu_ has joined #smartos[08:31:08] <bixu_> Do any of the crontab in the smartos global zone persist across reboots?[08:31:39] <MerlinDMC> crontabs should be saved in /var ... so that would be persistent[08:32:16] <bixu_> MerlinDMC: Where in /var ?[08:32:49] <bixu_> I see logs there in /var/cron, but no crontabs (probably missing something trivial).[08:33:34] <bdha> /var/spool/cron[08:33:46] <bixu_> *forehead smack*[08:33:52] <bixu_> I must be tired.[08:34:35] <bixu_> Thanks - that's it.[08:34:38] <bdha> :)[08:34:59] <MerlinDMC> hmm .. i didn't know where ... just that it is inside /var ;)[08:36:19] <bixu_> Here's your furry animal video reward: http://youtu.be/OOjFYUDTvjI?hd=1[08:38:10] *** Licenser has quit IRC[08:43:39] *** bens1 has joined #smartos[08:53:42] *** _alram has joined #smartos[08:54:10] *** bixu_ has quit IRC[08:54:41] *** bixu_ has joined #smartos[08:58:50] *** bixu_ has quit IRC[09:19:33] *** Licenser has joined #smartos[09:21:07] <Licenser> dun dun dun[09:21:20] <Licenser> MerlinDMC what kind of visualisation you think would be more useful?[09:21:26] *** relling has joined #smartos[09:22:47] <Licenser> this bullet grap things?[09:26:47] <Licenser> bixu when in doubt find / -name cron[09:29:09] *** deirdres has joined #smartos[09:29:36] *** AikoM has joined #smartos[09:29:57] *** relling has quit IRC[09:39:36] *** relling has joined #smartos[09:43:54] *** relling has quit IRC[09:49:09] *** relling has joined #smartos[09:55:53] *** relling has quit IRC[10:05:39] *** relling has joined #smartos[10:10:19] <konobi> Kush-: ?[10:15:52] *** relling has quit IRC[10:21:59] *** relling has joined #smartos[10:26:28] *** relling has quit IRC[10:29:10] *** relling has joined #smartos[10:50:34] *** darjeeling has quit IRC[10:54:05] *** robinsmidsrod has quit IRC[10:58:32] *** relling has quit IRC[11:01:42] *** robinsmidsrod has joined #smartos[11:04:22] *** sam_wan has left #smartos[11:04:52] *** relling has joined #smartos[11:15:12] *** daa2 has joined #smartos[11:20:44] *** hij1nx has quit IRC[11:22:20] *** daa2 has quit IRC[11:25:12] *** dis has quit IRC[11:47:30] *** dis has joined #smartos[12:11:30] *** relling has quit IRC[12:23:44] *** relling has joined #smartos[12:32:51] *** Licenser has quit IRC[12:34:33] *** Licenser has joined #smartos[13:44:43] *** ira has joined #smartos[14:14:46] *** darjeeling has joined #smartos[14:47:17] *** relling has quit IRC[14:54:03] *** relling has joined #smartos[15:40:38] *** _alram has quit IRC[15:42:37] *** xinkeT has joined #smartos[15:45:07] *** sable has quit IRC[15:46:53] *** CarlosC has joined #smartos[15:58:28] *** bens1 has quit IRC[15:59:42] *** neophenix has joined #smartos[16:16:23] *** relling has quit IRC[16:18:23] *** relling has joined #smartos[16:22:12] *** _alram has joined #smartos[16:40:26] *** relling has quit IRC[16:42:46] *** relling has joined #smartos[16:42:51] *** bixu_ has joined #smartos[16:56:10] *** jdavis has joined #smartos[16:58:48] *** relling has quit IRC[17:01:15] *** relling has joined #smartos[17:08:53] *** akole has joined #smartos[17:09:43] *** mwcampbell has joined #smartos[17:10:12] *** hij1nx_ has joined #smartos[17:11:33] <mwcampbell> How does Joyent determine what subset of pkgsrc to provide as precompiled binary packages?[17:11:46] <mwcampbell> And is the source (e.g. Makefiles, patches) for Joyent's package repository available somewhere?[17:13:20] *** relling has quit IRC[17:13:52] *** relling has joined #smartos[17:16:48] *** bixu_ has quit IRC[17:18:04] <rmustacc> mwcampbell: Most of it is just a set of what there has been previous customer demand and what we've already modified to be more smf-friendly / integrated in our environments.[17:18:24] <rmustacc> We're also working on bulk pkgsrc builds to provide more generic stuff, but that's in the works.[17:18:50] <rmustacc> mwcampbell: It is all on github and portions of it are pushed up to pkgsrc by our folks who have been working on it for a while, I'll pull up the urls one sec.[17:19:26] <rmustacc> https://github.com/joyent/pkgsrc-wip https://github.com/joyent/pkgsrc https://github.com/joyent/pkgsrc-joyent[17:19:49] <rmustacc> And then this is what one of our folks put together for trying to build it all: https://github.com/mamash/pk[17:20:45] <mwcampbell> Can I build more packages from pkgsrc and install them alongside the precompiled ones in /opt/local?[17:21:19] <rmustacc> From upstream pkgsrc?[17:21:28] <mwcampbell> yes[17:21:35] <mwcampbell> if there's a package that's not in the Joyent binary repo[17:21:40] <rmustacc> If it's from the same repo, probably.[17:22:04] <rmustacc> But you should also just ask mamash/jperkin what it is you're looking for and they usually can add stuff.[17:22:48] <rmustacc> The goal soon is to have the full bulk everthing that builds from upstream available along side as a secondary source. As for how they interact, not entirely sure, there sorry. I'm mostly at the OS.[17:23:20] <jperkin> mwcampbell: yes, that's possible, though it's not documented right now (I will fix that soon)[17:24:22] <mwcampbell> OK, well, one package that I'll want which isn't currently in the Joyent repo is icecast[17:25:07] <jperkin> alternatively, I am currently building a full pkgsrc set for the new 2012Q2 branch, which will install into a different prefix, so you can have those in addition to the joyent-supplied ones[17:27:25] <jperkin> most of the niggles have been ironed out for 2012Q2, so once we provide the official joyent packages for that and some documentation it should be straight-forward to install our packages and then (attempt to) build any from pkgsrc we didn't package yet.[17:28:58] <mwcampbell> BTW, I notice that the current smartos64 dataset is still using 2011Q4[17:30:08] <mwcampbell> at least, the current smartos64 dataset on JoyentCloud[17:32:19] <mwcampbell> Is this IRC channel for users of SmartOS SmartMachines on a cloud provider like JoyentCloud, or only for administrators of SmartOS host boxes?[17:32:53] <jperkin> yep, I believe 2012Q1 is pretty much ready. also, you can update manually by unpacking http://pkgsrc.joyent.com/sdc6/2012Q1/x86_64/bootstrap.tar.gz and doing a 'pkgin up; pkgin fug' I think..[17:34:08] *** denizr has joined #smartos[17:34:34] <jperkin> mwcampbell: I don't see any distinction, general SmartOS questions are very much welcome irregardless of where it's running, however you'd be better going to joyent ops/support for specific questions relating to our service.[17:34:58] *** bens1 has joined #smartos[17:35:18] <mwcampbell> jperkin: understood. I'll stick to general SmartOS questions[17:35:45] *** trentster1 has joined #smartos[17:36:39] *** trentster has quit IRC[17:37:29] <mwcampbell> One thing that discouraged me the last time I seriously tried to use SmartOS, several months ago, was the very small community compared to major Linux distros. But I understand the advantages of OS virtualization, e.g. bare-metal performance, bursting, and observability, so I'm giving SmartOS another try[17:39:49] *** denizr has quit IRC[17:40:09] *** denizr has joined #smartos[17:41:13] *** relling1 has joined #smartos[17:41:30] *** relling has quit IRC[17:41:50] <Licenser> mwcampbell well small community != bad community[17:42:03] <Licenser> you know in linux a lot of people are already sick of answering questions[17:42:44] <Licenser> I think the SmartOS community is still small but it has a rather active core and lots of interested people - plus the joyent folks are really helpful :)[17:42:50] <mwcampbell> Oh, am I already asking too many? I'll spend some more time reading docs[17:43:19] <Licenser> mwcampbell god no there arent't too many questions :)[17:43:38] <Licenser> I'm just saying that I've experienced the RTFM attitude rather frequently in the Linux communtiy, not having seen it here yet[17:45:28] *** AikoM has quit IRC[17:48:00] *** relling1 has quit IRC[17:50:27] <mwcampbell> Question about zones: If one zone starts forking processes excessively, will the other zones on the system be protected, since they're all running on the same kernel?[17:50:48] <mwcampbell> For example, does each zone have its own run queue?[17:51:38] *** relling has joined #smartos[17:52:17] <Licenser> Good question, I am not 100% sure but I thinki you can limit it[17:59:45] *** hij1nx_ has quit IRC[18:02:13] *** relling has quit IRC[18:03:05] *** ryancnelson has joined #smartos[18:03:24] *** hij1nx_ has joined #smartos[18:04:05] *** hij1nx_ is now known as hij1nx[18:05:21] *** hij1nx has quit IRC[18:06:07] <konobi> mwcampbell: zones are restricted by resource caps that have been around for a very long time (see rcap/rctl)[18:08:00] <konobi> for example the zones that surround kvm instances are highly restricted in what can be done[18:09:21] <mwcampbell> What about standard SmartOS zones? For example, if I were to run a fork bomb on my SmartMachine (I won't!), would it affect other SmartMachines on the same box?[18:09:26] <ryancnelson> there's a limit to the number of threads/processes a zone can run.[18:09:29] *** relling has joined #smartos[18:09:37] <ryancnelson> you'll kill yourself before you bother me[18:10:01] <ryancnelson> go ahead and try it :)[18:10:32] <mwcampbell> Well I'm probably not on the same host box as anyone here[18:10:49] <mwcampbell> I get your point though[18:12:13] *** benbangert has quit IRC[18:12:35] <ryancnelson> for example, if you're running sdc , your package definitions look like this:[18:12:40] <ryancnelson> http://ryan.net/misc/skitch/Operations_Portal-20120703-091204.png[18:13:02] <ryancnelson> … joyent sets the number of processes you can run, given a smartmachine of X size....[18:14:06] <ryancnelson> they tend to be just about "what fraction of the whole box you bought" … so it's not oversold, but yeah, if a fork-bomb could wreck the global zone, our business model wouldn't last very long[18:14:21] <ryancnelson> 'cause people accidentally write fork bombs all the time :)[18:15:20] <jperkin> aka pkgsrc bulk builds :-)[18:17:11] <mwcampbell> ryancnelson: Were you being sarcastic there? I would never intentionally run a fork bomb on a machine on a public cloud; I just wanted to know how susceptible this OS virtualization model would be to abuse.[18:17:19] *** benbangert has joined #smartos[18:18:27] <ryancnelson> no, not really :) if you do it (and really, people do. we cost 8 cents an hour… people try lots of stuff), you'll just wedge your zone.[18:19:43] <ryancnelson> really, running "benchmarks" on us (something *everyone* seems to be doing this month… lots of analysts out there) is basically a DOS against your neighbors, unless you're good about it[18:20:22] <ryancnelson> the converse to that effect in places like amazon are the auto-scale systems that say "spin up 10 servers, test 'em, shut down the worst 8" ….[18:21:05] <ryancnelson> (not to bag on amazon too much. they're just the defacto currency of vm's here, since they're by far the biggest)[18:21:53] <ryancnelson> i wasn't saying "YOU JUST TRY IT, BUDDY!! MAKE MY DAY!" :)[18:22:33] <mwcampbell> I haven't run any benchmarks, and I won't.[18:24:50] *** relling has quit IRC[18:25:12] <Licenser> ryancnelson so howis the exchange course amazon / joyent? :P[18:25:30] <Licenser> I think running a benchmark on a VM is a totally valid action[18:25:36] <konobi> mwcampbell: you'd run out of memory and that wouldn't effect neighbours[18:26:19] <konobi> or you'd hit the zone/project limites[18:26:54] *** relling has joined #smartos[18:27:15] *** bixu_ has joined #smartos[18:28:16] <mwcampbell> Speaking of memory usage, here's a newbie question: On Linux, I can run "free" to check the amount of memory used and the amount available. What's the nearest command to that on SmartOS?[18:29:41] <wesolows> /bin/true[18:30:01] <ryancnelson> :)[18:30:23] <wesolows> the amount of memory used is almost always meaningless from the end user's perspective. /bin/true differs from free only in that it doesn't pretend otherwise.[18:30:26] *** hij1nx has joined #smartos[18:31:23] <wesolows> If you're a kernel engineer, you may find the ::memstat dcmd in mdb(1) useful.[18:31:24] <ryancnelson> mwcampbell: if you're an operator, running in the global zone, it's especially true, since your ram gets used by zfs as a cache to make the disks go faster.[18:31:26] *** bixu_ has quit IRC[18:31:37] <ryancnelson> you'll *usually* be using almost all your ram, and that's not a big deal.[18:31:46] <ryancnelson> it's using what you paid best buy for[18:31:47] <wesolows> along with things like ::kmastat, etc.[18:31:51] *** hij1nx_ has joined #smartos[18:32:30] <ryancnelson> (the ram used by zfs ARC is freed when something else needs it)[18:34:56] *** hij1nx has quit IRC[18:35:02] *** hij1nx_ is now known as hij1nx[18:35:28] *** relling has quit IRC[18:37:57] <Licenser> yea I got really nervouse by that the other day :P[18:38:08] <Licenser> a more interesting measure would be free memory + ARC[18:39:21] *** relling has joined #smartos[18:41:19] <wesolows> how is that actionable, though? it doesn't really tell you how many/much more processes you can run, because (a) processes share a lot of text, and (b) the ARC can't go to zero and in any case performce will degrade rapidly as it approaches zero, constraining memory use anyway[18:42:05] <wesolows> in order to really do anything useful with that, you need to know not only how much memory your processes use but also what they're using, which pmap(1) will tell you.[18:42:09] <Licenser> Well it gives you a rough idea where you stand[18:42:16] <wesolows> And you also need to know what kind of I/O they're doing.[18:42:38] <wesolows> very, very rough.[18:42:39] <Licenser> I mean I can tell if I am running out of memory that way[18:42:42] <Licenser> for example[18:42:53] <Licenser> or if ARC is getting low[18:43:04] *** denizr has quit IRC[18:43:22] *** denizr has joined #smartos[18:45:01] *** relling has quit IRC[18:45:52] *** banjiewen|away is now known as banjiewen[18:47:31] <ryancnelson> right… people use "how much free ram do i have" to answer the question "am i leaking memory, or are my users over-driving this host"… that's not a tremendously direct way to answer that question, and using that otherwise empty ram to improve i/o is a much smarter use for it[18:48:11] <ryancnelson> it's a complicated question, and tools that make it seem simple aren't telling you the truth, really[18:48:57] <Licenser> true[18:49:03] *** jdavis has quit IRC[18:51:43] *** deirdres_ has joined #smartos[18:51:43] <ryancnelson> i'm traveling the world again for awhile, folks… maybe i'll be on IRC from exotic timezones some, but if not, have a good July![18:52:17] *** ryancnelson has left #smartos[18:52:51] *** hij1nx has quit IRC[18:52:58] <elijah-mbp> licenser: i think most of the joyent folk are way used to answering questions instead of snapping RTFMs off - most of us have "we love solaris-derived stuff" as that other hat we wear.... :-) and we all want everyone to be successful.[18:53:26] *** hij1nx has joined #smartos[18:53:41] <Licenser> elijah-mbp yea I can see that, and I enjoy that :)[18:53:47] <elijah-mbp> (jokes about ORCL aside... i'd be happy to see them run illumos too ;-) )[18:54:10] <Licenser> oracle runs illumos o.O?[18:54:52] *** ira has quit IRC[18:55:38] <elijah-mbp> licenser: seems unlikely, love to see them flop over onto the open code like everyone else, though ;)[18:55:46] <Licenser> heh[18:55:48] <elijah-mbp> that would take obeying the license terms. ;)[18:55:52] <elijah-mbp> and possibly a soul.[18:55:54] <nahamu> whew. I'm now running on a spice-enabled VM sitting behind a dedicated firewall/NAT zone over an etherstub.[18:56:50] <elijah-mbp> nahamu: that sounds awesome.[18:56:51] <Licenser> nahamu I think you set the record of technical terms/word with that[19:01:17] *** sebasp has joined #smartos[19:15:42] *** Licenser has quit IRC[19:24:39] *** mwcampbell has quit IRC[19:28:09] *** scanf has quit IRC[19:33:34] <deirdres_> starting now: Max Burning preso on SmartOS, live stream http://www.ustream.tv/channel/illumos-user-group[19:40:00] <akole> yay[19:46:39] *** scanf has joined #smartos[19:48:50] *** hij1nx has quit IRC[19:50:45] *** jdavis has joined #smartos[19:53:25] *** hij1nx_ has joined #smartos[20:00:10] *** lpin has joined #smartos[20:01:09] <lpin> hello guys, what's the default file descriptor limit in SmartOS?[20:06:49] <nahamu> deirdres_: is it being recored too?[20:06:56] <deirdres_> yes[20:07:14] <deirdres_> I think both on a local camera and on ustream[20:09:16] <nahamu> thanks![20:09:35] *** neophenix has quit IRC[20:18:33] *** ryancnelson has joined #smartos[20:19:32] *** scanf has quit IRC[20:26:32] *** bens1 has quit IRC[20:27:09] *** bens1 has joined #smartos[20:27:49] *** hij1nx_ has quit IRC[20:35:55] *** hij1nx_ has joined #smartos[20:37:13] *** scanf has joined #smartos[20:41:57] *** sjorge has joined #smartos[20:44:09] *** neophenix_ has joined #smartos[20:45:56] *** Licenser has joined #smartos[20:50:01] *** jrbing has joined #smartos[20:54:57] *** ryancnelson has left #smartos[20:54:59] *** denizr has quit IRC[20:55:31] *** denizr has joined #smartos[20:56:41] *** scanf has quit IRC[21:16:31] *** hij1nx_ has quit IRC[21:34:04] *** lpin has left #smartos[21:36:02] *** sjorge has quit IRC[21:36:16] *** hij1nx has joined #smartos[21:41:16] *** hij1nx has quit IRC[21:45:57] *** hij1nx has joined #smartos[22:03:26] *** hij1nx has quit IRC[22:06:45] *** denizr has quit IRC[22:07:05] *** denizr has joined #smartos[22:07:40] *** hij1nx has joined #smartos[22:14:50] *** elijah-mbp has quit IRC[22:16:33] *** hij1nx has quit IRC[22:17:27] *** elijah-mbp has joined #smartos[22:20:29] *** denizr has quit IRC[22:20:49] *** denizr has joined #smartos[22:24:36] <joshie> In a joyent branded zone, if I added a hostname6.net0 file with an addif line, what would stop it from working when restarting the zone (it isn't working)[22:24:55] <joshie> i can plumb it and addif with ifconfig just fine[22:37:21] *** Orpheus has quit IRC[22:41:18] <elijah-mbp> joshie: for joyent-branded zones, do the smf services actually read that? all of the examples i just looked at in JPC have the IP in the zone xml.[22:41:33] *** bens1 has joined #smartos[22:41:44] *** denizr has quit IRC[22:41:58] <joshie> i made the assumption it did since you can use the hostname.interface file to change the ipv4 address[22:42:02] *** denizr has joined #smartos[22:42:16] <joshie> the assumption was faulty since the statechange script reads that file instead and adds it to the zone's xml[22:43:15] <elijah-mbp> ;)[22:43:28] <elijah-mbp> in /lib/svc/method/net-physical:[22:43:41] <elijah-mbp> https://gist.github.com/ce7d59b8c4813414d5fb[22:43:56] <elijah-mbp> (neat, limechat knows how to do gists - i didn't know it did that.)[22:44:25] <elijah-mbp> the net-physical script in general is worth a read - it's cool stuff going on in there![22:45:47] <joshie> my last statement was also wrong, it will grab an ip from hostname.interface IF one doesn't already exist in the zonecfg for that interface[22:48:04] <nahamu> elijah-mbp: is that from the service for the NGZs or for the GZ?[22:48:16] <elijah-mbp> ah, yeah, i wondered how it would handle that - the net0/net1 interface layout gives me some headaches @ the global zone (because multiple zones all have a net0...)[22:48:58] <elijah-mbp> nahamu: it's out of a zone on one of our headnodes, it may have a funny origin compared to a zone that most people would see. if so - sorry :)[22:49:34] <deirdres_> nahamu: http://smartos.org/2012/07/03/smartos-max-brunings-talk-at-nosig/[22:49:49] <nahamu> deirdres_: thanks! :)[22:50:29] <deirdres_> 2:15 hrs, some of it dead air I'm afraid[22:50:40] <deirdres_> they were also shooting to tape so I hope will do a nicer edit[22:50:46] <nahamu> as long as there is a fast forward button, it's all good.[22:52:57] <nahamu> elijah-mbp: for some reason I didn't realize that /lib also comes from the GZ.[22:58:22] *** CarlosC has quit IRC[23:01:31] *** bens1 has quit IRC[23:04:16] <joshie> Can someone explain to me why in the event that a nic is added without an ip address, statechange rips it out of the hostname file and adds it to the zonecfg?[23:04:56] <joshie> this seems to allow for a one time exploit of some sort if someone had access to that zone to give it an approved ip without interaction from the head node[23:05:21] *** relling has joined #smartos[23:05:29] <nahamu> but only in the instance that a zone was configured without an explicit IP address.[23:05:48] <nahamu> which, to be fair, is an odd case.[23:06:00] <joshie> right, I would assume without explicitly declaring one, I would be assuring it didn't have a usuable ip address that could make it through spoofing[23:06:05] <joshie> but that isn't the case[23:06:48] <nahamu> I think the antispoof only works if an address is defined.[23:08:05] <joshie> well, in that case, antispoofing would be off until someone set that hostname file, which potentially breaks something[23:08:15] <joshie> once again, that seems hidden and unfavorable[23:13:03] <elijah-mbp> that would, i think, be a bug in whatever provisioner is being used. ;) you shouldn't try to deploy zones with interfaces that don't do anything, imho ;)[23:55:43] <supr> on global zone, how can i setup pkgsrc?[23:56:40] <Teknix> supr: http://wiki.smartos.org/display/DOC/Installing+pkgin maybe?[23:57:23] *** scanf has joined #smartos[23:59:03] <supr> Teknix: thanks. These packages seem to be precompiled. Is it possible to compile the packages manually?[23:59:30] <elijah-mbp> supr: yes. github.com/mamash/pk i think is filip's repo for the tools you need for that.