September 24, 2013 
< | 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 | >
September 24, 2013 [00:02:53] *** neophenix has quit IRC[00:14:19] <potatosalad> Quick side note/sanity check: I think kstat's zones:::avenrun_* can also be used to grab the loadavg values for guest zones from the global zone.[00:14:33] <potatosalad> Which avoids the overhead of zlogin[00:24:22] *** jsm has joined #smartos[00:31:04] *** xenol has quit IRC[00:33:18] *** mikl has quit IRC[00:34:19] <rmustacc> potatosalad: Yes, they can, but do remember the kstats aren't a stable interface.[00:34:57] <rmustacc> Sounds like we should consider an uptime -z[00:35:37] *** mary5030 has quit IRC[00:36:02] <potatosalad> rmustacc: Does that just mean that they stat names/modules/classes could change?[00:36:27] *** xenol has joined #smartos[00:36:30] <rmustacc> potatosalad: Or they could go away and we could get the data through some other eay.[00:36:33] <rmustacc> *way[00:36:50] <rmustacc> Probably worth doing something like a getzloadavg[00:38:00] <rmustacc> That said, the kstats are unlikely to change, just worth mentioning it is what it is.[00:39:07] <potatosalad> yeah, I'm using getloadavg from sys/loadavg.h right now for grabbing it from the global zone, I noticed that pset_getloadavg is used in some places from within the zone with PS_MYID, but that's about as far as I got before my understanding of what's going on underneath ran out[00:39:44] *** xenol has quit IRC[00:39:55] *** xenol has joined #smartos[00:40:19] *** alfredodeza has left #smartos[00:44:33] <potatosalad> (this is in erlang) the kstat method is averaging around 0.0004ms while the zlogin uptime method is around 0.15ms per zone[00:47:27] <rmustacc> Well, using a library versus forking and execing will always be faster.[00:47:31] <rmustacc> Or at least, using the kstats should.[00:47:44] <jesse_> potatosalad, that's the problem with the java binding/test, too[00:50:05] *** wolstena has quit IRC[00:55:19] *** elijah-mbp has quit IRC[00:55:37] *** elijah-mbp has joined #smartos[01:15:35] *** deirdres has joined #smartos[01:18:10] *** cpn has joined #smartos[01:20:47] *** wolstena has joined #smartos[01:25:53] *** xenol has quit IRC[01:28:40] *** jim80net has quit IRC[01:30:22] *** jim80net has joined #smartos[01:34:38] *** jim80net has quit IRC[01:36:30] *** xenol has joined #smartos[01:36:43] *** xenol has quit IRC[01:53:43] *** jellydonut has quit IRC[01:57:51] *** wolstena has quit IRC[01:59:35] *** wolstena has joined #smartos[02:06:00] *** jim80net has joined #smartos[02:09:46] *** potatosalad has quit IRC[02:29:57] *** mary5030 has joined #smartos[02:32:49] *** lloydde has quit IRC[02:33:15] *** lloydde has joined #smartos[02:36:37] *** mary5030 has quit IRC[02:37:10] *** mary5030 has joined #smartos[02:37:31] *** lloydde has quit IRC[02:48:37] *** wolstena has quit IRC[02:49:50] *** FatDarrel has quit IRC[02:54:39] *** LnxTx has quit IRC[02:54:51] *** LnxTx has joined #smartos[03:02:09] *** deirdres_ has joined #smartos[03:02:15] *** deirdres has quit IRC[03:02:15] *** deirdres_ has quit IRC[03:02:33] *** deirdres has joined #smartos[03:03:48] *** lloydde has joined #smartos[03:12:11] *** lloydde has quit IRC[03:15:28] *** mary5030 has quit IRC[03:17:13] *** mary5030 has joined #smartos[03:17:41] *** dap has quit IRC[03:19:16] *** dap has joined #smartos[03:20:56] *** mary5030 has quit IRC[03:51:55] *** zr0 has left #smartos[03:53:08] *** potatosalad has joined #smartos[03:56:36] *** potatosalad has quit IRC[03:58:23] <BugeyeD> hi all. accustomed to using supervise from daemontools; installed via pkgin; how do i go about having it start automatically? is there something similar to bsd's rc.local, or do i have to create a manifest for smf?[03:59:53] *** mary5030 has joined #smartos[03:59:55] <BugeyeD> hmmm ... /etc/inittab?[04:03:07] *** nefilim has joined #smartos[04:07:37] <wesolows> normally if something's in pkgsrc, it should already come with an SMF manifest.[04:08:28] <wesolows> If not, you'll need to create one. Something like github.com/davepacheco/smfgen might be helpful[04:08:54] <wesolows> You can get a list of the services that already exist via svcs(1)[04:09:04] <wesolows> You can import and modify manifests and properties using svccfg(1m)[04:09:11] <wesolows> You can view properties using svcprop(1)[04:09:19] <wesolows> And you can start with smf(5)[04:10:25] <BugeyeD> wesolows: thanks for the pointers. pkgsrc didn't contain a manifest. i remember some of this from before i left solaris 10, but was never comfortable with creating manifests and such. being a bsd guy, i'm used to plain simple config files. :)[04:11:47] <wesolows> smfgen might be right up your alley then. A few lines of JSON and a shell script should get you there.[04:12:08] <wesolows> There are other tools people also use -- I think one is called manifold. I've not used it myself.[04:12:36] *** szaydel has quit IRC[04:13:38] *** lloydde has joined #smartos[04:17:50] *** lloydde has quit IRC[04:19:45] *** mary5030 has quit IRC[04:34:18] *** elasticdog has quit IRC[04:36:23] *** mary5030 has joined #smartos[05:00:12] *** jim80net has quit IRC[05:04:03] *** deirdres has quit IRC[05:05:22] *** jim80net has joined #smartos[05:14:44] *** jsm has quit IRC[05:24:25] *** lloydde has joined #smartos[05:34:18] *** szaydel has joined #smartos[05:36:10] *** ipalreadytaken has joined #smartos[05:52:27] *** potatosalad has joined #smartos[06:05:35] *** szaydel has quit IRC[06:08:44] *** szaydel has joined #smartos[06:11:27] *** mary5030 has quit IRC[06:12:23] *** julianduque has quit IRC[06:16:27] *** lloydde_ has joined #smartos[06:17:18] *** lloydde has quit IRC[06:32:25] <arekinath> is there a manpage for sysevent_evc_subscribe somehwere?[06:32:37] <arekinath> zoneevent is dying every time I try to create a new vm at the moment[06:33:03] <arekinath> when I run it on its own it's saying "failed to subscribe to channel: 12"[06:33:36] <arekinath> 12 is ENOMEM... so I'd like to know why sysevent_evc_subscribe is consistently returning that[06:33:41] <arekinath> there's plenty of free ram around[06:33:58] <arekinath> (assuming the return values are errno, I don't even know that, no manpage)[06:37:36] <konobi> arekinath: svcs -x ?[06:37:48] <arekinath> konobi: nothing[06:38:37] <konobi> try restarting syseventd[06:39:03] *** jim80net has quit IRC[06:39:23] *** jim80net has joined #smartos[06:40:14] <arekinath> ok, have done. still the same output from zoneevent[06:40:27] <arekinath> this is happening on more than one machine if that's any help[06:41:35] <konobi> past any experiences that i've had[06:43:09] *** jim80net_ has joined #smartos[06:43:29] <arekinath> hm. it looks like the only way to get ENOMEM out of that is if (evch_dl_getnum(&chp->ch_subscr) >= chp->ch_maxsubscr)[06:43:54] *** jim80net has quit IRC[06:43:54] *** jim80net_ is now known as jim80net[06:44:56] <arekinath> ohhh[06:45:02] <arekinath> project-fifo's chunter[06:45:09] <arekinath> seems to be using up all the slots[06:45:11] <arekinath> hm[06:45:21] <arekinath> if I kill chunter everything is back to normal[06:45:46] * arekinath wonder if there are mdb walkers for this stuff[06:46:52] <arekinath> EVCH_MAX_SUBSCRIPTIONS defaults to 32[06:57:35] *** szaydel has quit IRC[07:03:45] <arekinath> yeah, chunter is trying to create 1 subscription to that per zone on the system + some extras by the looks of it[07:03:53] <arekinath> I have 112 zones on that machine[07:04:12] <arekinath> it wants to create 167 subscriptions all up, but of course 32 is the max[07:04:15] <arekinath> "awesome"[07:04:38] <arekinath> I can just tweak the max with mdb to something > 168 and it works ok[07:04:55] <arekinath> but it's probably set to 32 for performance reasons[07:05:24] <arekinath> Licenser: ^ when you're awake at some point[07:07:36] *** nefilim has quit IRC[07:12:40] *** xinkeT has joined #smartos[07:14:05] *** Teknix has quit IRC[07:14:58] <Licenser> arekinath i don't see how that directly happen how did you trace the number to chunter?[07:15:22] <Licenser> I kind of suspect the zonemon.sh script, if you can make that return just a blank instead of running the dtrace and retry it would be great[07:15:25] *** xinkeT has quit IRC[07:15:41] *** xinkeT has joined #smartos[07:21:21] <arekinath> Licenser: it's the zonedoor processes that own all the subscriptions[07:21:53] <Licenser> then it's a gerneal smartos thing[07:22:09] *** dap has quit IRC[07:22:12] <arekinath> like /opt/chunter/lib/chunter-0.3.11/priv/zonedoor[07:22:14] <arekinath> that zonedoor[07:22:32] <arekinath> chunter starts one for every zone[07:22:46] <Licenser> yup[07:22:58] <Licenser> yea has to since it's zone specific[07:23:39] <arekinath> yeah. so libzdoor is opening the subscription[07:23:46] <arekinath> where's the code for that again...[07:23:46] <Licenser> yup[07:23:50] * arekinath get sick of changing repos all the time[07:24:11] <Licenser> hidden deeply in the guts of SmartOs :)[07:32:16] <konobi> https://github.com/joyent/illumos-joyent/tree/master/usr/src/lib/libzdoor/common[07:32:32] <arekinath> it doesn't seem to call the sysev stuff directly? hmm[07:34:29] <konobi> nope[07:34:56] <arekinath> libsysevent.so.1`sysevent_evc_subscribe[07:34:56] <arekinath> libzonecfg.so.1`zonecfg_notify_bind+0x12f[07:34:56] <arekinath> libzdoor.so.1`zdoor_handle_init+0x4a[07:34:56] <arekinath> zonedoor`main+0x19[07:34:56] <arekinath> zonedoor`_start+0x6c[07:35:00] <arekinath> comes via libzonecfg[07:35:14] <arekinath> well then[07:35:18] <arekinath> what shall we do about this[07:37:25] <arekinath> Licenser: is there an easy way to stop chunter starting zonedoors until this is sorted?[07:37:48] <Licenser> arekinath replace the binary with a dummy? :P[07:37:52] <arekinath> Licenser: will that only break the ssh pubkey stuff?[07:38:04] <Licenser> yup[07:38:24] <Licenser> or just increase the number untill it becomes a problem[07:38:31] <Licenser> those values usually are ages old defaults[07:38:45] <Licenser> I recently got an alert becayuse a system had more then 300 processes :P[07:39:01] <arekinath> yeah... I bumped it to 256 temporarily on one of my boxes and it doesn't seem to be suffering as a result[07:39:11] <arekinath> but changing it permanently requires recompiling the kernel[07:39:16] <arekinath> since it's a preprop macro[07:39:45] <Licenser> ouch[07:40:04] <Licenser> no dynamic changing?[07:40:45] <arekinath> you have to walk down from the zone_t to the zds_entry through to the evch_globals and then along the list until you find the one for com.sun:zones:status[07:40:54] <arekinath> and then edit its ch_maxsubscr[07:41:16] <arekinath> I don't have the mdb-foo to automate that.. how do you do a string comparison?[07:42:05] <Licenser> no clue[07:42:32] <konobi> shouldn't be the case[07:42:58] <arekinath> konobi: I can't see any code to change the max in an ioctl or anything like that... did I miss it?[07:43:14] <konobi> arekinath: i think it may be a library misuse[07:43:29] <arekinath> konobi: ah, you mean to attack it there[07:43:50] <konobi> but i can't find a zdoor client to compare it to[07:45:26] *** mikl has joined #smartos[07:47:18] <Licenser> sadly libzdor isn't documented so it's hard to say if I implemented it correctly[07:47:28] <konobi> file it as a bug[07:50:10] <Licenser> I agree[07:50:18] <MerlinDMC> morning[07:57:02] *** lloydde_ has quit IRC[07:57:36] *** lloydde has joined #smartos[07:58:27] *** mikl has quit IRC[08:01:41] *** lloydde has quit IRC[08:08:28] *** jim80net_ has joined #smartos[08:09:18] *** jim80net has quit IRC[08:09:18] *** jim80net_ is now known as jim80net[08:09:56] <arekinath> http://hastebin.com/raqosibeya.txt[08:09:59] <arekinath> ^ mdb to work around it[08:10:00] <arekinath> haha[08:14:57] *** cwo has joined #smartos[08:16:13] *** hadret has joined #smartos[08:35:43] *** jim80net_ has joined #smartos[08:36:46] *** jim80net has quit IRC[08:36:47] *** jim80net_ is now known as jim80net[08:39:19] *** ilovezfs has quit IRC[08:43:07] *** jim80net has quit IRC[08:43:46] *** mgls has joined #smartos[08:45:58] *** ilovezfs has joined #smartos[08:46:08] *** cwo has quit IRC[08:52:36] *** CaTtleyA has joined #smartos[09:02:07] *** nikolam has joined #smartos[09:03:01] *** bahamas10 has quit IRC[09:03:59] *** aortmann has joined #smartos[09:05:53] *** mikl has joined #smartos[09:08:23] *** lloydde has joined #smartos[09:14:16] *** lloydde has quit IRC[09:14:29] *** nikolam has quit IRC[09:15:28] *** mikl has quit IRC[09:18:39] *** bahamas10 has joined #smartos[09:20:01] *** texarcana has quit IRC[09:21:57] *** alcir has joined #smartos[09:22:07] *** texarcana has joined #smartos[09:23:44] *** jim80net has joined #smartos[09:24:40] *** jim80net has quit IRC[09:30:33] *** jim80net has joined #smartos[09:35:03] *** KermitTheFragger has joined #smartos[09:59:09] *** bahamas10 has quit IRC[10:08:56] *** bahamas10 has joined #smartos[10:11:48] *** leecallen35 has joined #smartos[10:15:01] *** leecallen has quit IRC[10:23:52] *** aortmann has quit IRC[10:25:44] *** bens1 has joined #smartos[10:39:32] *** aortmann has joined #smartos[10:43:03] *** mikl has joined #smartos[10:53:16] *** marsell has quit IRC[10:53:47] *** marsell has joined #smartos[10:56:07] *** yruss972 has joined #smartos[10:58:23] <CaTtleyA> Bonjour, je veux installer FusionInventory-Agent . mais je ne sais pas comment chercher si le package existe merci[11:01:58] <jperkin> BugeyeD: by default packages come with a lukem BSD style rc.d script, which may or may not work after installing the rc.subr package (it's been a while since I used it)[11:04:14] <CaTtleyA> i not found the package pkg_add in p5-FusionInventory-Agent[11:04:47] <CaTtleyA> i found for netbsd : http://pkgsrc.se/search.php?so=-FusionInventory-Agent[11:05:09] <jperkin> CaTtleyA: install a 13.2.* image, it will be available there.[11:06:23] <jperkin> ah no, it doesn't package[11:06:31] <jperkin> I'll fix it[11:08:00] <CaTtleyA> are you the link for the package source of 'p5-FusionInventory-Agent[11:08:19] <jperkin> "link"?[11:12:40] <CaTtleyA> How install the fusion-agent without paquet? .~where found the source code de fusion-agent .[11:13:08] <jperkin> if you want to build from source, start by following http://www.perkin.org.uk/posts/pkgsrc-on-smartos-zone-creation-and-basic-builds.html[11:13:28] <jperkin> the fix for p5-FusionInventory-Agent is to create a net/p5-FusionInventory-Agent/PLIST.SunOS file containing 'bin/memconf'[11:13:37] <jperkin> after that, a 'bmake install' should complete successfully[11:14:07] <jperkin> I will integrate that file into upstream pkgsrc so that updated images include that package.[11:14:57] <MerlinDMC> jperkin, how would I set PKG_OPTIONS for www/ap-php? show-options indicates to use PKG_OPTIONS.ap24-php53 but that doesn't add any options[11:18:22] <jperkin> I think that's just an artefact of including php's options.mk[11:20:20] <jperkin> not sure, would need to dig through to see which particular prefix will set it - for now you could just set PKG_DEFAULT_OPTIONS[11:20:46] <jperkin> if that package specifically has build options it should have its own options.mk rather than showing php's[11:21:02] <jperkin> as not all php modules are e.g. going to have ssl or dtrace support[11:22:22] <MerlinDMC> I just try to get a threadsafe php apache module ... I did build by using PKG_DEFAULT_OPTIONS but it seems to compile a non-threadsafe version ... will cleanup completely and recompile[11:31:32] *** mikl has quit IRC[11:31:58] *** tonist has joined #smartos[11:32:43] *** potatosalad has quit IRC[11:33:08] *** mamash has joined #smartos[11:35:03] *** potatosalad has joined #smartos[11:51:43] *** d[^_^]b has quit IRC[12:12:15] *** aortmann has quit IRC[12:13:12] *** aortmann has joined #smartos[12:14:56] *** ipalreadytaken has quit IRC[12:28:20] *** aortmann has quit IRC[12:29:22] *** aortmann has joined #smartos[12:32:13] *** CaTtleyA has quit IRC[12:33:07] *** yruss972 has quit IRC[12:34:01] *** CaTtleyA has joined #smartos[12:34:53] *** yruss972 has joined #smartos[12:42:21] *** xenol has joined #smartos[12:45:55] *** bens1 has quit IRC[12:46:52] *** bens1 has joined #smartos[13:15:05] *** socketwiz has quit IRC[13:25:21] *** socketwiz has joined #smartos[13:32:35] *** socketwiz has quit IRC[13:33:42] *** tonist has quit IRC[13:33:55] *** socketwi has joined #smartos[13:34:28] *** szaydel has joined #smartos[13:37:33] *** socketwi has quit IRC[13:40:11] *** CaTtleyA has quit IRC[13:41:59] *** CaTtleyA has joined #smartos[13:55:20] *** mamash has left #smartos[13:57:27] *** mamash has joined #smartos[14:07:42] *** xenol has quit IRC[14:07:50] *** xenol has joined #smartos[14:11:27] *** sa5bke has joined #smartos[14:26:07] *** xenol has quit IRC[14:26:26] *** xenol has joined #smartos[14:27:04] *** yruss972 has quit IRC[14:30:37] *** yruss972 has joined #smartos[14:37:09] *** jsm has joined #smartos[14:42:44] *** oddsignals has joined #smartos[14:42:52] *** jsm has quit IRC[14:43:13] *** jsm has joined #smartos[14:43:27] *** tonist has joined #smartos[14:46:14] *** mgls has quit IRC[14:46:19] *** Tekni has quit IRC[14:46:28] *** ivan\ has quit IRC[14:46:29] *** Guest87035 has quit IRC[14:46:43] *** gregburd has quit IRC[14:46:58] *** nshalman has quit IRC[14:47:08] *** ivan\ has joined #smartos[14:47:13] *** Tekni has joined #smartos[14:47:14] *** ivan\ has quit IRC[14:47:14] *** ivan\ has joined #smartos[14:47:14] *** gregburd has joined #smartos[14:47:23] *** Jared_M has joined #smartos[14:49:51] *** nshalman has joined #smartos[14:52:48] *** mamash has left #smartos[15:00:48] *** mamash has joined #smartos[15:05:47] *** xinkeT has quit IRC[15:07:17] *** Teknix has joined #smartos[15:10:02] *** Teknix has quit IRC[15:10:19] *** Teknix has joined #smartos[15:10:19] *** ira has joined #smartos[15:12:42] *** mamash has left #smartos[15:16:21] *** CaTtleyA has quit IRC[15:18:24] *** CaTtleyA has joined #smartos[15:19:16] *** marsell has quit IRC[15:19:44] *** marsell has joined #smartos[15:25:06] *** sebasp_ has joined #smartos[15:25:21] *** sebasp_ has quit IRC[15:25:33] *** sebasp_ has joined #smartos[15:25:34] *** sebasp_ has quit IRC[15:26:09] *** sebasp_ has joined #smartos[15:26:09] *** sebasp_ has quit IRC[15:28:47] *** yruss972 has quit IRC[15:35:21] *** jsm has quit IRC[15:37:59] *** hadret has quit IRC[15:40:42] *** sebasp_ has joined #smartos[15:40:54] *** sebasp_ has quit IRC[15:41:22] *** sebasp_ has joined #smartos[15:58:07] *** CaTtleyA has quit IRC[15:59:51] *** yruss972 has joined #smartos[16:03:46] <jperkin> pkgsrc.{joyent.com,smartos.org} is now geoip-based, so should be faster for folks not in us-east, let me know if you hit issues.[16:05:45] <wiedi> will the pkgsrc-eu-ams.joyent.com hostname still work in the future or should i change it to just pkgsrc.joyent.com?[16:06:24] <jperkin> that will always work, yeh.[16:07:16] <wiedi> both are returning the same ip for me, so geoip seems to work - nice :)[16:07:17] *** CaTtleyA has joined #smartos[16:08:00] <xenol> jperkin: I am from Central Europe and I am getting US-based IP. I guess I should get from Joyent's DC in Amsterdam? :P[16:08:12] *** yruss972 has quit IRC[16:08:39] <jperkin> yes[16:10:37] <xenol> xenol@selene ~ % geoiplookup 165.225.128.155[16:10:38] <xenol> GeoIP Country Edition: US, United States[16:11:33] *** jim80net has quit IRC[16:13:19] *** prasm has quit IRC[16:13:55] *** prasm has joined #smartos[16:21:04] <opeth__> jperkin: hooray ;][16:21:11] *** nikolam has joined #smartos[16:21:14] <opeth__> that's really welcome[16:23:14] *** marsell has quit IRC[16:23:36] *** marsell has joined #smartos[16:30:22] *** yruss972 has joined #smartos[16:30:33] *** neophenix has joined #smartos[16:33:39] *** jim80net has joined #smartos[16:33:47] *** jsm has joined #smartos[16:39:39] *** nefilim has joined #smartos[16:40:17] *** lloydde has joined #smartos[16:41:43] *** tonist has quit IRC[16:42:18] *** aortmann has quit IRC[16:46:11] *** potatosalad has quit IRC[16:55:32] *** KermitTheFragger has quit IRC[16:56:40] *** yruss972 has quit IRC[16:58:37] *** mary5030 has joined #smartos[16:59:17] *** mary5030 has joined #smartos[17:07:59] *** potatosalad has joined #smartos[17:08:31] *** CaTtleyA has quit IRC[17:09:08] <jzu_> w 44[17:18:58] *** pde has joined #smartos[17:20:19] *** axonpoet has joined #smartos[17:22:20] *** Ymarg has joined #smartos[17:30:37] *** alcir has quit IRC[17:32:07] *** dap has joined #smartos[17:33:36] *** theCzar has joined #smartos[17:33:58] *** nikolam has quit IRC[17:36:52] *** jwk404 has joined #smartos[17:38:13] *** lloydde has quit IRC[17:38:34] *** dap has quit IRC[17:50:39] *** szaydel has quit IRC[17:50:56] *** lloydde has joined #smartos[17:59:13] *** szaydel has joined #smartos[18:04:34] *** yruss972 has joined #smartos[18:09:07] *** yruss972 has quit IRC[18:18:16] *** dap has joined #smartos[18:21:17] *** ira has quit IRC[18:22:38] <opeth__> jzu_: you can't have 44 windows open in your irssi can you now? :)[18:27:51] *** noahmehl has quit IRC[18:32:17] *** FatDarrel has joined #smartos[18:34:26] <nathan7> opeth__: at 30 right now, more isn't rare[18:35:06] *** ryancnelson has joined #smartos[18:35:52] *** asonge has joined #smartos[18:36:28] <asonge> so, i've got a machine with smartos on it, just restarted it, and it upgraded to 20130629, and now my network stopped working completely[18:36:54] <asonge> before, i had to open the network card in promiscuous mode with snoop to get more than 4 vlans running[18:37:10] <asonge> and i'm kinda lost in the illumos sea for how to troubleshoot this again[18:37:43] <rmustacc> asonge: What did you upgrade from and what card do you have?[18:38:35] <asonge> it's an intel card if memory serves correctly...and i can't remember what i upgraded from...this is a PXE-boot from ovh[18:39:21] <asonge> i remember filing a bug report about this a few months back[18:40:03] *** yruss972 has joined #smartos[18:40:12] <asonge> pci8086,2002 (pciex8086,1503) [Intel Corporation 82579V Gigabit Network Connection], instance #0[18:40:42] <asonge> https://github.com/joyent/smartos-live/issues/206 was the issue i filed[18:41:06] <ryancnelson> you're sure your networking woes aren't OVH-induced?[18:41:20] <rmustacc> ryancnelson: This is pretty obviously a driver bug given the snoop.[18:41:40] <asonge> rmustacc: the difference now is that the snoop workaround no longer works[18:41:43] <ryancnelson> tops out at 4 vlans?[18:41:43] *** Alek has joined #smartos[18:41:48] <rmustacc> asonge: Huh, okay.[18:42:30] <asonge> any tips on what i can do?[18:42:42] <rmustacc> So there have basically been no changes to the 82579V code since it was first integrated.[18:43:23] <asonge> should i ask to be moved to a box that has a different intel network chipset?[18:43:26] <rmustacc> So there are a few questions to go answer.[18:43:30] <rmustacc> No, I wouldn't.[18:43:43] <rmustacc> First off is if you have no zones up, you can get to the GZ right?[18:43:50] <asonge> i get to the GZ fine[18:44:06] <asonge> i'm ssh'd into gz now[18:44:16] <rmustacc> And it's when you add the fourth mac address that everything goes off the rails?[18:44:16] *** ipalreadytaken has joined #smartos[18:44:23] <asonge> well, that was the *old* behavior[18:44:33] <asonge> let me shut down a bunch of vm's[18:44:33] <rmustacc> Okay, so what is the *new* behavior specifically.[18:44:39] <rmustacc> When does it actually kick in.[18:44:42] <asonge> well, the box just restarted and i can only reach GZ[18:48:05] <opeth__> nathan7: nice, I seldom hit the dozen-mark myself[18:48:07] <asonge> rmustacc: ah, it's a network issue[18:48:15] <asonge> routing table thing[18:48:19] <rmustacc> Okay.[18:48:22] <asonge> otherwise same behavior as before, re: 4[18:48:35] <rmustacc> Well, I can try to help you dig into that for a little bit if you have some time.[18:48:35] <asonge> apparently "route -p" doesn't mean persistent?[18:48:39] *** mamash has joined #smartos[18:48:49] *** ipalreadytaken has quit IRC[18:48:50] <asonge> i have a few hours to kill, if you need my help smashing the bugs[18:48:52] <nathan7> opeth__: Nearly all of my communication happens over IRC[18:49:04] <asonge> there's no production anything on the box if you want access[18:49:31] <rmustacc> asonge: Well, for the GZ most likely whatever it's trying to persist it in is not persitent due to the nature of the GZ.[18:49:50] <asonge> rmustacc: nah, this was a route in the VM[18:49:54] <rmustacc> Thanks, but I generally try to avoid access to other folks machines. So give me a few minutes to refamiliarize that bit.[18:50:47] *** elijah-mbp has quit IRC[18:51:46] *** elijah-mbp has joined #smartos[18:51:57] *** chorrell has joined #smartos[18:52:10] <opeth__> nathan7: I also remember doing that :)[18:52:34] <nathan7> asonge: are you using the failover IP shit?[18:52:41] <asonge> nathan7: nope[18:52:44] <nathan7> k[18:53:05] <nathan7> running that currently for my OVH SmartOS box because I don't have an entity to get a RIPE block with[18:53:51] *** elasticdog has joined #smartos[18:53:59] <asonge> i should set the ip's up differently, probably...but i've had issues finding any docs on the crossbow stuff[18:54:24] <asonge> and there's no docs on how OVH does IP address mac filtering or whatever on their switches (i'm not a network guy, as is obvious)[18:56:55] <rmustacc> asonge: So I have a D script I've cooked up that should help us understand what's going on.[18:57:02] <rmustacc> Or at least let me know what the right question to ask next is.[18:57:08] <asonge> awesome.[18:57:16] <rmustacc> You'll want to be set up such that you have all the zones that are working.[18:57:31] <asonge> all zones are up and running[18:57:33] <rmustacc> And then run the script, and boot the zone that will not work without snoop.[18:57:38] <rmustacc> http://us-east.manta.joyent.com/rmustacc/public/tmp/nic.d[18:57:38] <asonge> ah, okay[18:57:56] <rmustacc> Basically the gld should be toggling it into promiscuous mode automatically for you, but it's not for some reason.[18:58:11] <rmustacc> gld is the generic lan driver, basically the framework for nic drivers in illumos.[18:59:26] <rmustacc> Meanwhile I need to grab the datasheet for that part and just confirm that the magic number is actually 4.[18:59:27] <asonge> okay, apparently, i'm mistaken. i stopped snoop and everything's back up.[18:59:34] <asonge> wtf[19:00:07] <asonge> is there some way for me to dump the routes or something? (that way, i don't have to zlogin and add routes everything)[19:00:27] <rmustacc> From the global zone for another zone?[19:00:48] <asonge> well, basically, on reboot it seemed that all my non-global zone routing table entries didn't come back up[19:01:15] <asonge> let me just restart and see.[19:02:55] *** elijah-mbp has quit IRC[19:03:14] *** wolstena has joined #smartos[19:09:52] *** elijah-mbp has joined #smartos[19:10:47] <asonge> rmustacc: okay, the update must've actually fixed it...after i got my routing table straight, all the vm's respond without snoop.[19:11:08] <asonge> should i mark the issue as unable to reproduce again, since there's no actual known cause for the fix[19:11:23] <rmustacc> Well, I did find one bug fix that could address the problem you were seeing.[19:11:29] <rmustacc> But what were you upgrading from?[19:11:46] <rmustacc> Specifically https://www.illumos.org/issues/3492 would be the fix.[19:11:54] *** tonist has joined #smartos[19:13:11] <asonge> rmustacc: i seem to remember a february 2013 release was running before i rebooted[19:13:19] <asonge> or march[19:15:08] <rmustacc> That wasn't merged in until later February.[19:15:09] *** ryancnelson has quit IRC[19:15:12] <rmustacc> So you could have just missed it.[19:15:14] *** tat- has quit IRC[19:32:11] *** lloydde has quit IRC[19:32:45] *** lloydde has joined #smartos[19:33:58] *** lloydde has quit IRC[19:34:04] *** lloydde_ has joined #smartos[19:35:11] *** yruss972 has quit IRC[19:35:44] *** yruss972 has joined #smartos[19:39:58] *** yruss972 has quit IRC[19:45:51] *** wolstena has quit IRC[19:45:56] *** sebasp_ has quit IRC[19:47:05] *** xinkeT has joined #smartos[19:47:13] *** sebasp_ has joined #smartos[19:49:38] *** Tekni has quit IRC[19:49:56] *** sebasp_ has quit IRC[19:50:03] *** sebasp_ has joined #smartos[19:52:00] *** sebasp_ has quit IRC[19:52:09] *** sebasp_ has joined #smartos[19:52:09] <natefoo> has anyone imported a zpool created under freebsd on smartos/illumos before?[19:52:14] *** sebasp_ has quit IRC[19:53:09] *** sebasp_ has joined #smartos[19:54:47] <jzu_> opeth__: I have usually ~48 windows[19:54:49] <jzu_> opeth__: for real :)[19:55:10] <opeth__> massive[20:03:32] *** tonist has quit IRC[20:13:26] *** tonist has joined #smartos[20:15:30] <griffin_> yes natefo[20:16:18] <griffin_> I have my notes natefoo, its really easy as long as you dont read the bad info from guys who dont know better with blogs[20:16:54] <natefoo> yeah i saw some crap about efi labeling each individual device and resilvering. seemed like a bad idea.[20:17:18] <griffin_> yeah don't do that[20:18:44] <natefoo> so, what's the secret?[20:18:51] <griffin_> all you need to do use link the proper devices into a tmp directory and use the use the -d on the zpool command line[20:18:54] *** xmerlin has quit IRC[20:19:33] <griffin_> here is my raw notes[20:19:36] <griffin_> http://pastebin.com/jZWcRq3q[20:20:44] <natefoo> yeah so i am completely unable to export, so this may not work.[20:21:02] <griffin_> you can always import with -f[20:21:06] <natefoo> i may give it a try without the export anyway.[20:21:23] *** wolstena has joined #smartos[20:22:47] <griffin_> but using the -d on a temp directory with links to the proper devices I went freebsd->smartos->freebsd->smartos[20:23:03] <griffin_> so it works well[20:23:05] <natefoo> yeah, pretty cool (and simple).[20:23:22] <griffin_> there is some pretty bad misinformation out there[20:34:51] *** bens1 has quit IRC[21:01:59] *** xenol has quit IRC[21:02:08] *** xenol has joined #smartos[21:03:22] *** lumb has quit IRC[21:11:35] *** ryancnelson has joined #smartos[21:12:11] *** julianduque has joined #smartos[21:13:56] *** pringlescan1 has joined #smartos[21:14:02] <pringlescan1> where do pk log files go when the build_fails?[21:14:24] *** tonist has quit IRC[21:15:07] *** lloydde_ has quit IRC[21:15:23] *** tonist has joined #smartos[21:15:43] *** lloydde has joined #smartos[21:17:03] *** tonist has quit IRC[21:18:01] *** tonist has joined #smartos[21:19:55] *** lloydde has quit IRC[21:21:09] *** wolstena1 has joined #smartos[21:21:39] *** wolstena has left #smartos[21:28:20] *** hotbox has quit IRC[21:33:10] *** FatDarrel_ has joined #smartos[21:34:44] *** FatDarrel has quit IRC[21:43:41] *** mamash has left #smartos[21:46:56] *** tat- has joined #smartos[21:47:06] *** lloydde has joined #smartos[21:52:56] *** julianduque has quit IRC[21:55:05] *** lloydde has quit IRC[22:01:12] *** mamash has joined #smartos[22:03:26] *** lloydde has joined #smartos[22:12:25] *** leecallen has joined #smartos[22:13:22] *** wolstena has joined #smartos[22:14:22] *** leecallen35 has quit IRC[22:18:45] *** Triluch has quit IRC[22:20:28] *** Triluch_ has joined #smartos[22:23:29] *** pringlescan1 has quit IRC[22:45:40] *** tonist has quit IRC[22:51:22] *** wolstena has left #smartos[22:53:30] *** julianduque has joined #smartos[22:54:54] *** theCzar has quit IRC[22:55:13] *** theCzar has joined #smartos[22:56:01] <theCzar> is it possible to run htop in the global zone?[22:56:24] <theCzar> it seems the lxproc stuff is missing. Although it works in an OS zone[22:57:23] <jperkin> you should be able to mount it manually[22:59:35] *** gemelen has left #smartos[23:00:54] <theCzar> jperkin: from where?[23:02:36] <jperkin> mkdir /system/lxproc; mount -F lxproc lxproc /system/lxproc[23:04:13] <theCzar> jperkin: ah, gottcha, thanks[23:04:15] *** bens1 has joined #smartos[23:07:15] *** bens1 has quit IRC[23:09:52] *** Triluch_ is now known as Triluch[23:11:43] *** Fuzai has quit IRC[23:20:09] *** wolstena has joined #smartos[23:22:31] *** yo61 is now known as robinbowes[23:22:41] *** robinbowes is now known as yo61[23:24:51] *** wolstena has quit IRC[23:27:44] *** bixu has joined #smartos[23:27:54] <bixu> Any pkgsrc peeps on right now?[23:30:29] <jperkin> yeh somewhat[23:34:31] *** Fuzai has joined #smartos[23:36:25] <oddsignals> lpsmith: Did you have any luck getting ghc up and running?[23:40:28] *** julian_duque has joined #smartos[23:42:33] *** julianduque has quit IRC[23:42:37] *** julian_duque is now known as julianduque[23:47:18] *** noahmehl has joined #smartos[23:48:15] <lpsmith> oddsignals, I've been busy with more immediately important projects the last few weeks[23:48:40] <lpsmith> I think I have a plan of attack mapped, and have established the feasibility of most of the links[23:49:07] <lpsmith> The big question is how well the cross-compilation support really works[23:49:41] <lpsmith> Of course, there's probably some unknown unknowns still lurking[23:50:13] *** mamash has left #smartos[23:50:51] <lpsmith> but basically, my plan is to put a recent GHC build tree on NFS, then using linux's binfmt_misc and remote execution via ssh to build the cross compilation toolchain[23:51:10] *** andy_js has joined #smartos[23:52:24] <bixu> jperkin: Hi :)[23:53:14] <bixu> Trying to pkg_add our custom lz4c package and getting this error: pkg_add: Invalid package name lz4c[23:53:20] <bixu> A bit cryptic.[23:53:37] <bixu> pkg_info -X agains the tgz seems reasonable.[23:53:39] *** andy_js has quit IRC[23:53:43] <bixu> *against[23:53:48] <jperkin> it has a version suffix?[23:54:07] <bixu> It's getting -0.0 automatically.[23:54:08] <jperkin> gist the pkg_info -X output[23:54:53] <bixu> https://gist.github.com/anonymous/49f3e13495feeeff88b0[23:55:50] <jperkin> yeh it's missing a version[23:56:10] <jperkin> PKGNAME needs to be lz4c-0.0 or so[23:56:39] <bixu> Does it make sense that this package works, OTOH?[23:56:44] <bixu> https://gist.github.com/anonymous/6704372b24deceb84acc[23:56:58] <oddsignals> lpsmith: Sounds like a good approach in theory at least :-) Let me know if I can help with anything, I'm using a linux kvm for Haskell at the moment but would prefer a native port.[23:57:43] <jperkin> in theory no, but perhaps at least having a - is confusing pkg_*[23:58:29] <bixu> jperkin: Do we just make sure the -N.N is in the final argument?[23:58:33] <bixu> to pkg_create[23:58:53] <jperkin> yes[23:59:20] <jperkin> 1[23:59:43] <jperkin> so that version number would break any comparison checks[23:59:56] <jperkin> drop the dash so it's 1.9.3p392 and it should be ok