September 27, 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 27, 2013 [00:03:50] *** kevinykchan has joined #smartos[00:05:52] *** mary5030 has quit IRC[00:15:23] *** njm has quit IRC[00:27:40] *** khushildep has joined #smartos[00:28:44] *** deirdres has joined #smartos[00:28:46] <khushildep> Hey all[00:28:47] *** AlainODea has joined #smartos[00:29:39] <ryancnelson> hi![00:30:18] <AlainODea> hello :)[00:31:16] <AlainODea> oddsignals: myself and leroux are working on Cross-Compiling GHC. I hear you are bootstrapping from the 32-bit GHC from an earlier pkgsrc[00:33:45] <khushildep> If I'm setting up a new zone, should the gateway be the GZ or the normal GW IP?[00:34:09] <khushildep> Yes it's a noob question :) I'm struggling to get a smarts zone talking to anything at the moment.[00:37:00] <AlainODea> khushildep: normal GW IP in my experience[00:37:24] <AlainODea> khushildep: not at all noob. It's a valid question about the deployment architecture :)[00:37:27] <leroux> AlainODea: I'm a bit farther along. =) oddsignals provided a fix for a problem I was having.[00:37:43] <AlainODea> leroux: awesome teamwork :)[00:37:57] <leroux> http://ghc.haskell.org/trac/ghc/ticket/8361[00:38:04] <leroux> http://ghc.haskell.org/trac/ghc/ticket/8366[00:38:32] <leroux> But now I'm getting a million StgRun errors.[00:39:17] <leroux> oddsignals: Could you tell me what your HOST_OS is? You can find out with `ghc --info | grep "Host platform"`[00:40:23] <leroux> AlainODea: I think he's gotten up to 7.6.3.[00:43:32] <leroux> Actually, I don't need that...[00:44:24] <leroux> I need `./inplace/bin/ghc-stage1 --info | grep "target os"`.[00:46:34] <AlainODea> leroux: Yep, he reported that to smartos-discuss. I'm incorporating his patch into my cross_ghc.sh script.[00:46:45] <leroux> Which patch?[00:46:57] <AlainODea> leroux: http://ghc.haskell.org/trac/ghc/ticket/8366[00:46:59] <leroux> The include?[00:47:01] <leroux> Ok.[00:47:01] <AlainODea> leroux: yep[00:47:42] <leroux> Oh, there's also another "roadblock". http://ghc.haskell.org/trac/ghc/ticket/8362[00:47:57] <leroux> Shouldn't be too hard to fix.[00:48:10] <leroux> But I'll wait till we can build HEAD>[00:49:47] <AlainODea> leroux: I suspect that's a problem with using numbers instead of symbols for expected return codes[00:51:15] <leroux> http://ghc.haskell.org/trac/ghc/attachment/ticket/8362/testsuite_out.txt[00:51:32] *** jim80net has quit IRC[00:52:08] <leroux> I'm going to sleep for a bit then get back to work on this.[00:52:24] <leroux> I didn't sleep till 5am (then woke up at 7am for school).[00:52:36] <leroux> See you later.[00:53:54] *** neophenix has quit IRC[00:53:59] <AlainODea> leroux: sleep for a quite a bit dude. School is important. I think maybe set this aside for tonight and get back at it fresh tomorrow :)[00:54:11] <leroux> Nah. I still perform well in school.[00:54:33] <AlainODea> Good stuff :)[00:54:50] <AlainODea> khushildep: did it work with an an external gateway?[00:55:21] <khushildep> AlainODea - yeah sorry, so used to SDC these days that I didn't bother to check that resolv.conf was correct - bangs head on desk[00:55:24] <khushildep> err oops[00:55:59] <AlainODea> khushildep: I usually put dns_domain and resolvers in the JSON for vmadm create. Simplifies things[00:56:14] <khushildep> *writes notes* thanks AlainODea[00:59:45] <AlainODea> You can't change dns_domain with vmadm update. You have to edit the zone XML or use zonecfg -z uuid[01:00:08] <AlainODea> zonecfg is discouraged, but I imagine directly editing the zone XML is even more discouraged :)[01:01:48] <khushildep> AlainODea - any idea on this http://pastebin.com/ym2UQevB ?[01:02:59] *** wiedi_ has joined #smartos[01:03:00] <AlainODea> khushildep: slight variation. SmartOS has nic tags as an abstraction over vnics. Try nictagadm list[01:03:10] <AlainODea> You use those as the NICs in your JSON[01:03:38] <khushildep> mmm that command isn't found...[01:04:10] <AlainODea> Hmm. What does uname -r say?[01:04:15] *** wiedi has quit IRC[01:04:26] <AlainODea> Useless. Sorry. uname -a[01:05:45] <AlainODea> I'm on joyent_20130822T211238Z. Soon to be joyent_20130919T215407Z[01:06:16] <khushildep> ah never mind - wrong terminal - maybe I should go to sleep! :-)[01:07:47] *** lloydde has quit IRC[01:07:54] <AlainODea> Not a bad idea. The mind works better with rest :) I'm here often and on smartos-discuss and always glad to help[01:08:08] <khushildep> and one more question - promise - what the 'safe' way to change the root pw on the GZ?[01:08:23] *** lloydde has joined #smartos[01:09:06] <AlainODea> I'd actually switch to SSH login. You can force it with /usbkey/config. http://wiki.smartos.org/display/DOC/extra+configuration+options#extraconfigurationoptions-Console%2FLogin%2FSSHkeys[01:09:24] <AlainODea> I can put the specifics in a Gist if you like[01:09:58] <khushildep> Ooo yes please![01:10:27] <AlainODea> Give me a few moments and I will be back with a link... :)[01:10:52] <khushildep> thx[01:11:39] *** noahmehl has quit IRC[01:12:31] *** lloydde has quit IRC[01:16:01] <ryancnelson> don't wanna steal your thunder, AlainODea, but:[01:16:02] <ryancnelson> http://www.perkin.org.uk/posts/smartos-global-zone-tweaks.html[01:17:17] <AlainODea> ryancnelson: no worries at all. Better to spread jperkin's solid docs than have a slew of half-baked Gists :)[01:17:45] <AlainODea> I'll talk to locking it to public key only[01:18:48] <ryancnelson> that post has usurped my own http://www.psychicfriends.net/blog/archives/2012/03/21/smartosorg_run_things_at_boot.html[01:19:07] <AlainODea> sed is giving me trouble in the GZ. It's not the GNU sed[01:19:17] <ryancnelson> i need a better example use-case for /opt/custom/smf , since ssh-key-population is best done another way[01:19:27] <wesolows> correct; it's the real sed. :)[01:23:29] <AlainODea> https://gist.github.com/AlainODea/6722007[01:23:36] <AlainODea> wesolows: LOL. Indeed.[01:23:56] <wesolows> just add -e, right?[01:23:58] <AlainODea> wesolows: real enough to thwart my efforts to use in-place editing[01:24:11] <AlainODea> sed -e 's/PasswordAuthentication yes/PasswordAuthentication no/g' -i bak /usbkey/config[01:24:19] <wesolows> oh that, yeah.[01:24:27] <AlainODea> That doesn't work. I'm sure I've done something stupid, but I can't see it[01:24:34] <wesolows> they did some -i thing but it's totally fucked iirc.[01:25:16] *** andoriyu1 has joined #smartos[01:26:06] <AlainODea> wesolows: I enjoyed your talk at Sun Family Reunion.[01:26:27] <wesolows> oh thanks.[01:26:44] <wesolows> the missing anecdote is C++ symbol names too long to use with DTrace[01:26:51] *** andoriyu has quit IRC[01:26:52] <wesolows> I forgot that one.[01:26:56] <AlainODea> Nice[01:27:11] <AlainODea> I enjoy war stories[01:27:32] <AlainODea> You folks have some many. You got to do a lot of incredible things and still do[01:27:48] <AlainODea> *so many.[01:28:34] <AlainODea> I showed people at work the BOMs you shared publicly, and they were blown away. I was explaining that Open Source is a lot deeper than just source code.[01:29:38] <khushildep> assume this needs a reboot? restarting ssh service doesn't seem to help[01:30:00] <AlainODea> khushildep: it does. /usbkey/config is applied at boot[01:30:21] <khushildep> *grumbles* fingers crossed my shitty super micro IPMI holds out :-)[01:30:22] *** lloydde has joined #smartos[01:30:28] <khushildep> KVM even[01:30:45] <AlainODea> /usbkey/shadow is where root's pw hash is gotten. It can be changed across a boot from there[01:31:35] <AlainODea> IPMI is pretty shade regardless. I had to rescue iDRAC *from* SmartOS using ipmitool. Someone needs to create a bullet-proof and simpler BMC[01:31:57] *** utlemming has quit IRC[01:32:00] <AlainODea> I can't live without it, but I wish it worked better.[01:33:44] *** utlemming has joined #smartos[01:39:30] *** mary5030 has joined #smartos[01:42:56] <ryancnelson> re: sed ... /usr/perl5/bin/perl behaves pretty well for twisting text around.[01:43:34] <wesolows> unfortunately building such a BMC is "not allowed". Sorry.[01:43:36] <rmustacc> And please don't use /usr/perl5/bin/perl or rely on it.[01:43:48] <wesolows> ryancnelson: do not use that perl.[01:43:54] <AlainODea> wesolows: LOL.[01:44:19] <AlainODea> I like my tortured approach that fills /tmp with junk. It's clearly good ;)[01:44:20] <wesolows> it would put a bunch of big guys out of business, so it's not allowed.[01:44:39] <wesolows> this industry has a strict no-disruption policy.[01:44:46] <AlainODea> wesolows: if you build one it should be called Run BMC[01:44:58] <rmustacc> Poor Bryan.[01:44:59] <wesolows> clever, but that wasn't my plan.[01:46:15] <AlainODea> rmustacc: I thought you'd like that :)[01:47:26] <AlainODea> wesolows: apropos http://www.youtube.com/watch?v=_hN1SKVx31s[01:54:28] *** lloydde has quit IRC[01:59:59] *** jlove has joined #smartos[02:04:43] *** szaydel has quit IRC[02:07:39] *** deirdres has quit IRC[02:12:29] *** szaydel has joined #smartos[02:15:21] *** mary5030 has quit IRC[02:15:42] *** lloydde has joined #smartos[02:15:56] *** mary5030 has joined #smartos[02:16:38] *** sebasp is now known as sebasp_[02:22:12] *** lloydde has quit IRC[02:22:38] *** lloydde has joined #smartos[02:24:37] *** jlove has quit IRC[02:27:17] *** lloydde has quit IRC[02:27:30] *** noahmehl has joined #smartos[02:38:09] *** sebasp_ is now known as sebasp[02:39:29] *** Alek has quit IRC[02:48:29] *** khushildep_ has joined #smartos[02:49:26] *** khushildep has quit IRC[02:49:26] *** khushildep_ is now known as khushildep[02:52:42] *** wolstena has quit IRC[02:59:32] *** ilovezfs has quit IRC[03:03:06] *** ryancnelson has quit IRC[03:04:59] <AlainODea> khushildep: did the reboot work out?[03:05:25] <khushildep> AlainODea ys indeed - thanks for the Gist mate.[03:16:55] *** ImTheBitch has quit IRC[03:18:11] *** sebasp is now known as sebasp_[03:18:53] *** khushildep has quit IRC[03:19:32] *** khushildep has joined #smartos[03:25:13] *** ImTheBitch has joined #smartos[03:27:14] <AlainODea> khushildep: my pleasure :)[03:27:59] <AlainODea> GHC has a curious build process. The Cross-Build needs some clean up badly. It requires a brain the size of a planet to understand it[03:29:38] <Licenser> heh[03:31:03] *** kevinykchan has quit IRC[03:32:32] <AlainODea> I like a challenge[03:32:35] <AlainODea> It sharpens the mind[03:33:14] *** lloydde has joined #smartos[03:35:45] *** jaimef has quit IRC[03:39:36] *** lloydde has quit IRC[03:41:46] *** dap has quit IRC[03:41:54] *** jaimef has joined #smartos[03:43:27] *** khushildep has quit IRC[03:52:15] *** jaimef has quit IRC[03:56:31] *** nefilim has joined #smartos[03:56:46] <konobi> huh... LDNS looks pretty cool[03:56:54] *** jaimef has joined #smartos[03:57:48] *** noahmehl has quit IRC[03:59:53] *** jaimef has quit IRC[03:59:58] *** xmerlin has quit IRC[04:02:09] <richlowe> I think we must be looking at different things?[04:04:09] <konobi> thee freebsd lnds?[04:04:22] <konobi> um... ldns[04:05:54] *** jaimef has joined #smartos[04:14:37] *** szaydel has quit IRC[04:15:55] <AlainODea> konobi: how does ldns compare to ISC BIND?[04:16:40] <konobi> freebsd are replacing their use of bind with ldns[04:17:03] <AlainODea> Interesting. That's a pretty major shift.[04:18:52] <AlainODea> What prompted the change?[04:19:14] <konobi> DNSSEC it seems, that's the freebsd10 relnotes anyway[04:19:24] *** kevinykchan has joined #smartos[04:19:36] <AlainODea> Interesting. DNSSEC is a bit of a bear in BIND apparently. I haven't gotten around to trying it[04:19:57] <konobi> it's not too bad[04:20:02] *** ilovezfs has joined #smartos[04:20:12] <konobi> _but_ the tooling around bind makes it a pain[04:21:51] <AlainODea> rndc and friends?[04:22:20] <konobi> dig[04:23:03] <AlainODea> That's a client-side issue though isn't it. How does ldns change things?[04:23:40] <konobi> ldns has all that tooling client side, as well as server side and local forwarding side[04:25:38] *** kevinykchan has quit IRC[04:25:59] <AlainODea> Interesting, so ldns is intended to replace the whole DNS tooling across the board. That's ambitious[04:26:16] *** richlowe has quit IRC[04:26:22] *** richlowe has joined #smartos[04:26:36] *** kevinykchan has joined #smartos[04:28:26] *** szaydel has joined #smartos[04:36:48] <AlainODea> konobi: What got you into it?[04:37:43] <AlainODea> Oh wow. They have a really nice C API for this[04:39:10] <konobi> AlainODea: doing some fbsd cross compiling to get stuff working on freebsd on arm for ipv6[04:45:08] <AlainODea> It seems to be cross-compiling season ;)[04:45:33] <AlainODea> I'm cross-compiling GHC from Ubuntu to SmartOS[04:45:39] <AlainODea> I want 64-bit GHC[04:46:00] <richlowe> the winter of our discontent.[04:46:09] <konobi> time to talk to audrey =0)[04:53:35] *** szaydel has quit IRC[04:56:16] *** szaydel has joined #smartos[05:03:37] <AlainODea> Fault tolerant DNS, DHCP, and PXE. Check. Thank you SmartOS. Good night![05:03:41] *** AlainODea has left #smartos[05:03:57] *** jlove has joined #smartos[05:08:11] *** jlove has quit IRC[05:13:24] *** julianduque has quit IRC[05:15:48] *** julianduque has joined #smartos[05:31:25] *** richlowe has quit IRC[05:31:41] *** richlowe has joined #smartos[06:13:47] *** mary5030 has quit IRC[06:15:43] *** kevinykchan has quit IRC[06:16:12] *** kevinykchan has joined #smartos[06:23:00] *** dap1 has joined #smartos[06:24:13] *** jlove has joined #smartos[06:25:00] *** szaydel has quit IRC[06:27:59] *** kevinykchan has quit IRC[06:28:01] *** kevinykc_ has joined #smartos[06:29:02] *** jlove has quit IRC[06:32:32] *** kevinykc_ has quit IRC[06:33:42] *** kevinykchan has joined #smartos[06:33:55] *** dap1 has quit IRC[06:35:59] *** kevinykchan has quit IRC[06:36:52] *** dap1 has joined #smartos[06:36:53] *** kevinykchan has joined #smartos[06:44:34] *** kevinykchan has quit IRC[06:50:24] *** nefilim has quit IRC[06:56:56] *** dap1 has quit IRC[06:59:51] *** mary5030 has joined #smartos[07:04:39] *** wolstena has joined #smartos[07:05:15] *** wolstena has left #smartos[07:13:00] *** mary5030 has quit IRC[07:22:51] *** andoriyu has joined #smartos[07:25:50] *** andoriyu1 has quit IRC[07:54:03] *** jesser has joined #smartos[08:12:41] *** mamash has joined #smartos[08:32:12] <oddsignals> leroux: "i386-unknown-solaris2"[08:32:28] <leroux> oddsignals: Thanks. The patch should work fine then.[08:32:40] <leroux> http://ghc.haskell.org/trac/ghc/attachment/ticket/8366/0001-Include-termios.h-for-solaris2-host.-Fixes-8366.patch[08:35:57] <oddsignals> leroux: sorry, just saw your second message - "OSSolaris2"[08:36:03] <leroux> Mhm.[08:36:15] <leroux> That's fine. the host_os should be the same.[08:36:23] <leroux> Any luck today?[08:36:48] <oddsignals> Not yet, just got out of bed :-)[08:37:02] <leroux> Oh, yeah. I forgot, timezones...[08:38:26] <oddsignals> I'm not sure whether I should attempt to fix the HEAD build first, or just try to get 7.6.3 set up to cross-compile for 64-bit[08:38:36] <oddsignals> http://ghc.haskell.org/trac/ghc/ticket/8365[08:39:09] <leroux> I ran into issues with StgRun.[08:39:33] <leroux> You're error but with StgRun symbols.[08:39:56] <leroux> Umm, ghc-7.6.3 may be more useful.[08:40:34] <oddsignals> Then I'll aim for that for now[08:41:31] <leroux> Sounds good.[08:44:29] <oddsignals> regarding the IO_ensureIOManagerIsRunning etc., I'm hoping it's just a case of a missing #ifdef or something - from what I could tell it's platform-specific, so it could just be that it's not matching any of the architecture checks and thus not being compiled into the library. But I haven't really looked into it properly yet.[08:47:11] <leroux> I think we're both facing the same issues.[08:47:22] <leroux> I compiled with threads, so we may be getting a bit different messages.[08:47:42] <leroux> But my errors occurred in libHsrts so yours.[08:49:36] <oddsignals> leroux: probably - is it missing the same symbols? ensureIOManagerIsRunning, ioManagerCapabilitiesChanged and unpackCString?[08:50:01] <leroux> I don't quite remember. I should have tee'd the output.[08:50:13] <leroux> I should probably just tee ALL output when I'm working on ghc.[08:51:04] <oddsignals> I always forget myself - or like yesterday, start a script session for logging in one window and then start the build in another :-/[08:51:56] <leroux> What do you use for that?[08:53:09] <oddsignals> just script - "script filename" will log all input and output into that file until you Ctrl-D or exit[08:54:12] <leroux> Ahh, okay.[08:54:40] <leroux> And then you just `tail -f` that file?[08:57:39] <oddsignals> leroux: no, everything is still sent to the terminal as well[08:57:45] <leroux> Oh, okay.[08:58:09] <oddsignals> leroux: so like tee, but for a whole session instead of just a single command[08:58:15] <leroux> Right.[09:05:33] *** tonist has joined #smartos[09:07:22] <oddsignals> has there been a change to the pkgsrc servers? I'm getting twice the speed I'm used to all of a sudden :-)[09:15:24] <jperkin> oddsignals: yes, it's now geoip based[09:17:12] <oddsignals> jperkin: I heard, but I used to get the same speeds from eu-ams earlier as from the us servers[09:17:38] <oddsignals> consistently so close to 0.5Mbit that I just assumed it was capped[09:17:56] <oddsignals> but now I'm suddenly seeing 3.5MB/sec :-)[09:18:16] <oddsignals> eh, 5Mbit, not 0.5[09:18:30] *** alcir has joined #smartos[09:20:02] *** texarcana has quit IRC[09:20:48] <oddsignals> ah fantastic - my 32-bit 7.6.3 runs under multiarch :-)[09:20:58] <oddsignals> that should save a lot of time[09:21:22] *** andoriyu has quit IRC[09:22:01] *** texarcana has joined #smartos[09:51:04] *** mamash has left #smartos[09:58:24] *** disharmony has joined #smartos[10:13:08] *** ilovezfs has quit IRC[10:13:16] *** ilovezfs_ has joined #smartos[10:13:36] *** mamash has joined #smartos[10:15:38] *** leecallen35 has joined #smartos[10:17:32] *** leecallen has quit IRC[10:22:51] *** noahmehl has joined #smartos[10:28:52] *** CaTtleyA has joined #smartos[10:37:59] *** bens1 has joined #smartos[10:46:10] *** noahmehl_ has joined #smartos[10:47:52] *** Triluch has quit IRC[10:48:50] *** noahmehl has quit IRC[10:49:43] *** noahmehl_ has quit IRC[10:50:22] *** noahmehl has joined #smartos[10:51:02] *** swaagie has quit IRC[10:54:24] *** Triluch has joined #smartos[11:02:49] *** khushildep has joined #smartos[11:05:23] <CaTtleyA> Bonjour, j'ai télécharger un programme a installer tar.gz / peux pas installer il me manque la commande make. j'ai fais un pkgin install make | ~/installation_glpi/FusionInventory-Agent-2.3.2]# pkgin install make[11:05:27] <CaTtleyA> make is not available on the repository[11:05:29] <CaTtleyA> calculating dependencies... done.[11:05:32] <CaTtleyA> nothing to do.[11:05:46] <jperkin> 'gmake'[11:08:10] <CaTtleyA> jperkin: ok, i test[11:11:41] <CaTtleyA> the command is : gmake && gmake install ?.[11:11:45] <CaTtleyA> jperkin: ?.[11:16:45] <jperkin> well, depends on the software - I meant that you want 'pkgin install gmake' to get it, there is no 'make' package.[11:35:31] *** khushildep_ has joined #smartos[11:35:53] *** hadret has joined #smartos[11:36:37] *** noahmehl has quit IRC[11:39:42] *** khushildep has quit IRC[11:39:42] *** khushildep_ is now known as khushildep[11:48:49] *** noahmehl has joined #smartos[11:50:10] *** tonist has quit IRC[11:53:44] <alcir> there is a way to change the uuid of a vm?[11:56:22] <yofuh> as everything vmadm uses is stored in /etc/zones, i would assume it is[11:56:39] <jperkin> it's unlikely to be pretty, though[11:56:46] <jperkin> as everything is keyed off uuid[11:57:58] <JT-EC> Can just create a new vm and move the disk of course.[11:58:29] <alcir> but the uuid in zfs remain the same[11:58:35] <alcir> move? what do u mean?[11:58:37] <JT-EC> zfs rename[11:59:39] <yofuh> the result would be the same as changing the /etc/zone stuff + zfs rename[11:59:51] <yofuh> nothing in the dataset would change[12:01:07] <JT-EC> yofuh: Yep, it's just if you think messing in /etc/zones isn't pretty you can get vmadm to do it correctly.[12:01:37] *** swaagie has joined #smartos[12:06:53] *** noahmehl has quit IRC[12:25:42] *** khushildep has quit IRC[12:42:38] *** russiane39 has quit IRC[13:05:41] *** rodgort has quit IRC[13:06:17] *** tonist has joined #smartos[13:08:04] *** rodgort has joined #smartos[13:23:42] *** nikolam has quit IRC[13:29:39] *** AlainODea has joined #smartos[13:31:23] *** szaydel has joined #smartos[13:37:28] *** sebasp_ has quit IRC[14:21:20] *** CaTtleyA has quit IRC[14:23:04] *** CaTtleyA has joined #smartos[14:34:22] *** nathan7 has quit IRC[14:45:32] *** nathan7 has joined #smartos[14:55:32] *** pringlescan has joined #smartos[14:56:08] *** xenol has joined #smartos[14:56:27] <pringlescan> Are there any recommendations for running a redundant syslog server? Also are those writes synchronous in that if it's a remote machine will it slow everything down?[14:58:34] *** andy_js has joined #smartos[15:02:54] *** bens1 has quit IRC[15:04:00] *** bens1 has joined #smartos[15:05:21] *** nikolam has joined #smartos[15:05:31] *** sebasp has joined #smartos[15:20:18] <jperkin> oddsignals: do you happen to have a working 64-bit ghc?[15:20:29] <oddsignals> jperkin: not yet, sorry :-)[15:20:40] <oddsignals> I've gotten over a couple of snags and hit a few new ones[15:20:46] <jperkin> I'm having trouble persuading the 32-bit one to build me a 64-bit bootstrap[15:21:00] *** sebasp has quit IRC[15:22:07] *** sebasp has joined #smartos[15:22:56] <oddsignals> jperkin: me too - looking at http://wiki.alpinelinux.org/wiki/Porting_GHC_to_Alpine that Alain mentioned on the mailing list to see if it can get me any further[15:26:25] <oddsignals> I did at one point get a stage1 compiler that _claims_ to be targeting x64, but I'm not sure it was built correctly and ./configure didn't want to accept it for --with-ghc[15:29:07] <oddsignals> jperkin: just started a new build, what I'm doing differently now is specifying --host= as well, the alpinelinux page claims it's necessary for ghc < 7.7[15:31:30] <oddsignals> what I'm doing is running the 32-bit 7.6.3 I have on a multiarch base, with an alien script that just runs on the same machine and x86_64-unknown-solari2-gcc etc just being scripts that run gcc -m64[15:32:10] <oddsignals> jperkin: I did have to patch a few things to get it to compile on multiarch>[15:32:16] *** jim80net has joined #smartos[15:39:36] <oddsignals> jperkin: http://lpaste.net/93488[15:43:07] <oddsignals> jperkin: OK, that actually built a stage2 compiler - but it dumps core when run :-([15:43:23] <oddsignals> file /opt/src/ghc-7.6.3/inplace/lib/ghc-stage2[15:43:24] <oddsignals> /opt/src/ghc-7.6.3/inplace/lib/ghc-stage2: ELF 64-bit LSB executable AMD64 Version 1, dynamically linked, not stripped[15:43:45] <oddsignals> ldd says it's linked to the 64-bit libs as well[15:47:19] <jperkin> ok, getting closer ;)[15:48:23] *** neophenix has joined #smartos[15:50:56] *** hadret has quit IRC[16:08:12] *** xenol has quit IRC[16:08:20] *** xenol has joined #smartos[16:14:52] *** tonist has quit IRC[16:22:12] *** andy_js has quit IRC[16:22:30] *** andy_js has joined #smartos[16:23:48] *** mamash has left #smartos[16:24:35] *** andy_js has quit IRC[16:27:05] *** andy_js has joined #smartos[16:31:28] *** andy_js_ has joined #smartos[16:33:31] *** andy_js has quit IRC[16:33:31] *** andy_js_ is now known as andy_js[16:34:15] *** mary5030 has joined #smartos[16:46:58] *** andy_js has quit IRC[16:48:09] *** andy_js has joined #smartos[16:51:42] *** nefilim has joined #smartos[17:19:49] *** chorrell has joined #smartos[17:26:40] *** Teknix has quit IRC[17:26:57] *** Teknix has joined #smartos[17:30:27] *** pringlescan has quit IRC[17:37:59] *** alcir has quit IRC[17:44:02] *** nikolam has quit IRC[17:57:06] <AlainODea> oddsignals: Good stuff. This is really exciting progress![18:00:19] *** pringlescan has joined #smartos[18:01:19] *** nikolam has joined #smartos[18:12:20] *** theCzar has joined #smartos[18:13:15] *** andoriyu has joined #smartos[18:23:44] *** julianduque has joined #smartos[18:25:57] *** ryancnelson has joined #smartos[18:31:31] *** dap has joined #smartos[18:39:19] *** lloydde has joined #smartos[18:50:16] *** dap has quit IRC[18:57:11] *** mariusp has joined #smartos[19:07:09] *** jamiealquiza has joined #smartos[19:08:25] <oddsignals> AlainODea: It would be more exciting if it actually wanted to work :-)[19:09:58] *** wolstena has joined #smartos[19:20:42] <oddsignals> pstack from the core dump - any ideas? http://lpaste.net/93503[19:21:20] <oddsignals> I'll write a status message to the mailing list later, after I've put the kids to bed.[19:22:36] <rmustacc> Is that trying to use C++ code?[19:22:45] <rmustacc> Or does it just have its own mangling.[19:23:00] <rmustacc> Or just rather long and weird function names? ;)[19:35:14] *** Alek has joined #smartos[19:37:02] *** jamiealquiza has quit IRC[19:38:12] <pringlescan> is getting a ton of "5245c244 daemon: accept(8) failed errno=130 (Software caused connection abort)" a concern in openldap? I enabled start tls and now the service starts in maintenance mode… however when I start it manually and do -d −1 (verbose logging) it works… but I see many of those errors[19:38:17] *** jamiealquiza has joined #smartos[19:40:50] <rmustacc> You're getting a bunch of ECONNABORTED[19:40:57] <rmustacc> Seems like something to figure out.[19:41:05] <pringlescan> yeah… I can't figure out the exit 1 though[19:41:34] <pringlescan> I need my SMF to work, and I'm not getting any errors in svcs -x and slapd by default has no logging of it's own it appears[19:43:15] <pringlescan> also a slaptest works, which is my first troubleshooting step, so it's not a config issue[19:43:21] <wesolows> sometimes if you're not sure how something logs, it can be useful to pfiles it[19:43:32] <wesolows> that, or read the manual. but let's not get carried away![19:45:01] <pringlescan> i know i'm talking to the wrong crowd, but I've been doing 120 hours of overtime every 2 weeks, i don't have time to do laundry let alone read :-)[19:45:16] <pringlescan> when I remove the TLS part I don't get an exit 1 / maintenance state[19:45:35] <rmustacc> Did you check the service log?[19:45:36] <pringlescan> but now I want to run it interactively and make sure the connections aren't screwy without TLS[19:45:48] <pringlescan> svcs -x is empty, it just says start returned exit code 1[19:45:58] <pringlescan> *the log in svcs -x is empty, it does report a maintenance state[19:47:16] <pringlescan> ok, the connection issue isn't TLS related[19:47:39] <pringlescan> I wonder if Atlassian Crowd does something really dirty with connections...[19:48:33] <pringlescan> I think this has to do with connection pooling[19:50:20] <wesolows> typically when I've had issues with TLS it's because I've not set up certificates and keys properly.[19:50:39] <wesolows> But you've an uphill battle for sure if you can't find the log files or at least figure out how to get it to log what and where you want.[19:51:15] <wesolows> I've set up slapd before but I don't remember the config file syntax offhand. I'm available as a contractor, but I won't do your job for you no matter how overworked you may be. :)[19:51:44] <pringlescan> wesolows: thanks, it works when I run it by hand, but not from the smf, using the same exact run command[19:52:03] <wesolows> but only with TLS enabled...[19:52:25] <wesolows> which would seem to rule out the wrong daemonisation settings, but worth checking perhaps[19:52:26] <pringlescan> yes, connection issue is definitely crowd timing out the connections after 30 seconds, but I don't think openldap is set to close the connections[19:52:49] <wesolows> you could use truss -f to observe startd running the start method and children, and seeing why it dies...[19:52:50] <pringlescan> I'm using one I didn't make, but it runs as root, so it shouldn't be permission related, but I do have the permissions for the certs set to root:ldap[19:53:16] <wesolows> this problem will seem pretty easy after a few hours of sleep.[19:54:24] <pringlescan> as do most[19:55:07] <pringlescan> *shrugs* start tls and ldaps:// work fine[19:56:10] <pringlescan> I even changed to the same directory before running it and it works, there's nothing in the smf that should do this[19:57:32] <pringlescan> yay! i figured it out[19:57:45] <pringlescan> not sure where to report this or if you guys supply an openldap smf manifest[19:57:59] <pringlescan> but the package sets the ldap group to ldap and the smf sets it to root[19:58:27] <pringlescan> oh… wait now it thinks it's running but isn't accepting connections[20:00:42] *** chorrell has quit IRC[20:03:33] <pringlescan> slapd.conf ignores my log settings even though it's chowned and touched ahead of time, the logs are empty, even if the first lines of the config file are the log directives[20:07:11] *** chorrell has joined #smartos[20:09:36] <pringlescan> slapd will take the path to the certificates in relation to the work directory instead of the config file's path[20:22:22] <oddsignals> rmustacc: Haskell function names with a bit of decoration :-)[20:22:28] *** ryancnelson has left #smartos[20:22:57] <jesse_> pringlescan, do you have permissions on the path to the conf file blocking access?[20:24:38] <pringlescan> I just did -d −1 in the smf so it'd go to the smf log to troubleshoot the problem[20:24:52] <pringlescan> no idea why slapd can't write to the log file, but i can't worry about it[20:25:13] <pringlescan> our webfilter "support ldap" but it doesn't let you specify the objectclasses to search for or anykind of mapping so I'm not sure how that's supporting ldap...[20:25:45] <pringlescan> it only supports posixGroup and groupOfNames not the unique one I use… bleh[20:30:15] <oddsignals> At least I've established that all executables compiled with the cross-compiling GHC crash - not sure if it's the code generation or the linker that's the problem. Going to try again with an unregisterised build.[20:34:29] <rmustacc> oddsignals: Do the non-cross compiled ones work?[20:35:01] <oddsignals> rmustacc: yes, the 32-bit compiler works fine on 32-bit SmartOS now[20:35:21] <oddsignals> but 64-bit is still proving problematic[20:36:01] <oddsignals> so I'm trying with an unregisterised build so I can compile to C, and test from there[20:41:39] *** Meths has quit IRC[20:43:26] *** Meths has joined #smartos[20:47:52] *** tonist has joined #smartos[20:54:10] *** Alek has quit IRC[20:58:15] *** kele has joined #smartos[21:05:06] *** volci has quit IRC[21:05:32] *** volci has joined #smartos[21:07:13] *** sebasp is now known as sebasp_[21:18:19] <oddsignals> ghc --info | grep Host[21:18:19] <oddsignals> ,("Host platform","x86_64-unknown-solaris2")[21:18:29] *** jim80net has quit IRC[21:18:35] <oddsignals> +o/[21:18:44] <oddsignals> eh, \o/ :-)[21:20:06] *** lloydde has quit IRC[21:20:39] *** lloydde has joined #smartos[21:21:50] <jperkin> cool, can I get a tarball? :)[21:24:08] <oddsignals> jperkin: certainly, but I'd like to try to get it to the next stage first - it's built on multiarch, so it requires copying a few libs into the "wrong" directories and having gcc symlinked to x86_64-unknown-solaris2-gcc before it will run on plain base64.[21:24:35] <jperkin> ok[21:24:39] <oddsignals> jperkin: next step is trying to use it to build a ghc on a plain base64 system[21:25:03] *** lloydde has quit IRC[21:26:00] <jesse_> jperkin, just like java, multiarch required for 64bit=)[21:27:01] *** julianduque has quit IRC[21:29:32] <jperkin> multiarch isn't required[21:37:56] *** DonDaniello has joined #smartos[21:39:11] <jesse_> jperkin, for 64bit it is+[21:39:12] <jesse_> ?[21:39:39] <jesse_> it's built in two passes and requires the underlying 32bit to work[21:39:44] *** richlowe has quit IRC[21:40:20] <jesse_> (and all the java binaries are 32bit, and dynamically switch bitness according to options)[21:40:31] *** khushildep has joined #smartos[21:41:23] <DonDaniello> quick design question: is it possible to create a zpool/zfs on NFS and use it as a storage for VMs (zones)?[21:42:02] <jesse_> no[21:42:30] *** bixu has quit IRC[21:42:39] <jesse_> I think even using iscsi disks for zones pool has been explicitly disabled[21:42:58] <khushildep> guys where can I find a good reference for the json you need to pass to vmadm?[21:43:19] <jesse_> man vmadm?[21:43:32] <jesse_> and to get started, wiki.smartos.org?[21:43:53] <jperkin> jesse_: nope.[21:45:59] <jesse_> jperkin, nope to what?[21:46:08] <jperkin> openjdk requiring multiarch[21:47:04] <jesse_> 32bit is ok on 32bit arch, but there is no 64bit version?[21:47:17] <jesse_> unless you went through much pain to get it working[21:47:30] <jperkin> it wasn't much pain, just a few patches to disable multiarch[21:47:37] <jperkin> else we'd never be able to provide it for the 64bit images[21:48:19] <jesse_> so you basically built the "32bit build" as 64bit?[21:48:40] <jperkin> probably, I forget the details[21:48:59] <MerlinDMC> khushildep, just for the basic stuff in the json or do you want a reference of all the available options?[21:50:10] <khushildep> MerlinDMC would like to see some examples really, a bit more than whats on the wiki[21:53:05] *** lloydde has joined #smartos[21:56:04] <oddsignals> ok, that unregisterised compiler is *slow* :-([21:58:36] <killfill> anyone knows if smartos will run on this supermicro x9dri-ln4f+ mobo?[22:03:31] *** khushildep has quit IRC[22:05:01] <AlainODea> oddsignals: I'll be back at this in a few hours.[22:06:07] <oddsignals> AlainODea: good luck![22:06:46] <opeth__> jesse_: however I think it's possible to define a zpool other than zones in the payload and I know that you can build other pools on iscsi since I've done and am using that[22:06:48] <oddsignals> AlainODea: this build will probably still be going by then :-)[22:06:55] <AlainODea> oddsignals: you too. I'm getting closer. I'm hoping the HEAD version of the GHC can be used as a seed[22:07:13] <AlainODea> oddsignals: the GHC build takes a metric eternity to run[22:07:16] <opeth__> that might lead to having guests on networked storage[22:08:18] <oddsignals> AlainODea: well, with a proper ghc it wasn't _that_ bad - 20-30 mins - but with the unregisterised compiler it's... different.[22:08:21] <opeth__> yeah there is a zpool key[22:10:21] <opeth__> and yes, man approves[22:10:29] *** khushildep has joined #smartos[22:10:36] <opeth__> I'll have to try this at some stage[22:10:44] <oddsignals> AlainODea: I was also hoping to use HEAD as a starting point, but ran into problems and figured I might have better chances with cross-compiling from the 32-bit 7.6.3[22:12:51] <AlainODea> oddsignals: it's a bit clumsy, but the cross-build has had a LOT of improvement on it in 7.7+[22:13:45] <jesse_> opeth__, there is a setting here and there to tell about other-pool-than-zones, but I think I came across some hardcoded "zones" here and there a year ago[22:14:21] <oddsignals> AlainODea: That's what I heard, however I didn't find a quick way of getting past this: http://ghc.haskell.org/trac/ghc/ticket/8365[22:15:05] <jesse_> opeth__, probably fixed by now if someone uses it=)[22:16:13] <AlainODea> oddsignals: fun times :)[22:16:19] *** leecallen35 has quit IRC[22:16:38] *** leecallen has joined #smartos[22:17:07] *** Meths_ has joined #smartos[22:18:15] *** Meths has quit IRC[22:21:35] *** ilovezfs_ has quit IRC[22:25:14] *** ilovezfs has joined #smartos[22:28:30] *** mary5030 has quit IRC[22:29:58] *** mary5030 has joined #smartos[22:32:41] *** khushildep has quit IRC[22:36:23] *** Alek has joined #smartos[22:36:25] *** papertigers__ has joined #smartos[22:36:48] *** mariusp has quit IRC[22:37:11] *** richlowe has joined #smartos[22:38:00] *** papertigers has quit IRC[22:38:00] *** papertigers__ is now known as papertigers[22:56:04] *** tonist has quit IRC[22:56:41] *** DonDaniello has quit IRC[23:01:20] *** pringlescan has quit IRC[23:10:15] *** bens1 has quit IRC[23:14:09] *** Alek has quit IRC[23:17:00] *** julianduque has joined #smartos[23:17:21] *** kele has quit IRC[23:21:17] *** AlainODea has quit IRC[23:39:18] *** mary5030 has quit IRC[23:51:50] *** xenol has quit IRC[23:52:16] *** MerlinDMC has quit IRC[23:52:40] *** jelmd has quit IRC[23:52:43] *** LeftWing__ has joined #smartos[23:52:55] *** chorrell has quit IRC[23:53:01] *** LeftWing has quit IRC[23:53:16] *** Azbruh_ has joined #smartos[23:53:46] *** dumfries has quit IRC[23:53:50] *** nikolam has quit IRC[23:55:11] *** dumfries has joined #smartos[23:55:23] *** xenol has joined #smartos[23:57:07] *** MerlinDMC has joined #smartos[23:57:36] *** nikolam has joined #smartos[23:57:44] *** jamiealquiza has quit IRC[23:58:03] *** Azbruh has quit IRC