December 17, 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 | >
December 17, 2012 [00:03:32] *** wdent has joined #smartos[00:09:14] *** haydock has joined #smartos[00:17:58] *** wdent has quit IRC[00:35:06] *** haydock_ has joined #smartos[00:37:14] *** haydock has quit IRC[00:49:30] *** haydock_ has quit IRC[00:49:40] *** siezer has joined #smartos[00:50:13] *** montyz0 has joined #smartos[00:58:47] *** ryancnelson has quit IRC[01:06:32] *** wdent has joined #smartos[01:08:38] *** montyz0 has joined #smartos[01:13:44] *** montyz0 has quit IRC[01:43:11] *** mrvisser has joined #smartos[01:43:26] <mrvisser> hi all, I'm having a heck of a time trying to compile the latest version of cairo[01:44:14] <mrvisser> has anyone had any success? pkgsrc repos have version 1.20, but I need a later version. trying to install 1.12.8[01:45:27] <mrvisser> first I download and compile pixman, and then I download and compile cairo.. but cairo gives an elf_getarsym error[01:57:29] <nahamu> mrvisser: which dataset are you using / which pkgsrc tree?[02:00:03] <mrvisser> I'm using a "base" machine in the joyent cloud. pkgsrc repo is http://pkgsrc.joyent.com/sdc6/2012Q1/x86_64/All that is 1.20[02:00:09] <mrvisser> it installs successfully from there[02:00:27] *** wdent has quit IRC[02:00:33] <mrvisser> err, it has 1.10, not 1.20[02:00:34] <wesolows> what base revision? is that 1.8.1?[02:00:55] *** wdent has joined #smartos[02:01:00] <mrvisser> yes 1.8.1[02:01:11] <wesolows> ok, hang on a sec[02:05:26] *** wdent has quit IRC[02:05:36] <wesolows> yeesh, what a mess this stuff is.[02:06:09] <mrvisser> cairo is a mess?[02:06:18] <wesolows> pixman even is a horrid mess of goop[02:06:42] <wesolows> it's obvious that no one who writes this stuff cares about anything other than GNU/Linux[02:07:01] <mrvisser> ah[02:07:19] <wesolows> anyway, installing pixman now, then on to cairo so I can see how it's failing for you[02:07:31] <mrvisser> thanks wesolows[02:08:46] <mrvisser> wesolows: for cairo, using —enable-xlib=no on the configure step gets you a bit further[02:11:27] <wesolows> need to file a bug: tar hardcodes /usr/bin/xzcat but it's not in the platform[02:13:25] <mrvisser> ah, is that the error?[02:14:28] <wesolows> no, just an annoyance along the way[02:14:38] <wesolows> but it's definitely a bug[02:14:43] <mrvisser> I see[02:19:16] <wesolows> this machine isn't all that fast... still building[02:20:24] <mrvisser> yea takes some time[02:22:35] <wesolows> ok, on base-1.8.1, I do pkgin in png xz gmake pkg-config libtool-base automake autoconf gcc47, and then build pixman with MAKE=gmake ./autogen.sh; gmake install[02:22:40] <wesolows> then I do:[02:22:53] <wesolows> MAKE=gmake PKG_CONFIG_PATH=/usr/local/lib/pkgconfig CPPFLAGS="-I/opt/local/include -I/usr/local/include" LDFLAGS="-L/usr/local/lib -L/opt/local/lib -R/opt/local/lib -R/usr/local/lib" ./configure --prefix=/usr/local --enable-xlib=no --enable-xlib-xrender=no; gmake[02:22:58] <wesolows> and it builds successfully[02:23:41] <mrvisser> *blink*[02:23:52] <mrvisser> going to give that a try[02:24:58] <wesolows> though strangely, my pkgsrc seems to be 2012Q2, not Q1, and this is definitely 1.8.1[02:25:23] <wesolows> ah, crap. You're building 64-bit, aren't you?[02:25:34] <mrvisser> I'm not on base64[02:25:39] <mrvisser> I'm on the 32-bit machine[02:25:41] <wesolows> ok[02:25:43] <wesolows> so am I[02:25:50] <mrvisser> nice, that's promising :)[02:25:59] <mrvisser> I've been banging my head on this off and on the last couple days[02:26:03] <wesolows> makes me wonder why your pkgsrc is Q1 x86_64 and mine is Q2 i386[02:26:26] <mrvisser> wesolows: I copied the pkgsrc from the website at that time[02:26:30] <mrvisser> I'm in i386[02:26:33] <mrvisser> *on[02:26:35] <wesolows> ok[02:27:01] <wesolows> well, let's see if it works for you. Also, my path is probably different from yours, so I'll note it:[02:27:06] <wesolows> /home/wesolows/i386-smartos/bin:/home/wesolows/bin:/usr/local/bin:/usr/local/sbin:/usr/bin:/usr/sbin:/opt/local/bin:/opt/local/sbin:.[02:27:21] <wesolows> obviously the first 2 aren't interesting[02:29:47] <mrvisser> noted, installation is going, we'll see how it fares[02:30:14] *** jesser has quit IRC[02:32:01] <mrvisser> wesolows: pixman doesn't have an autogen.sh script in it[02:32:13] <mrvisser> pixman-0.28.2[02:43:43] <mrvisser> wesolows: same error :([02:43:45] <mrvisser> ld: elf error: file ../util/cairo-missing/.libs/libcairo-missing.a: elf_getarsym[02:43:45] <mrvisser> ld: warning: file ../src/.libs/libcairo.so: linked to /tmp/cairo-1.12.8/src/.libs/libcairo.so: attempted multiple inclusion of file[02:43:45] <mrvisser> ld: fatal: file processing errors. No output written to .libs/cairo-analyse-trace[02:48:52] <mrvisser> trying now with —enable-trace=no :S[02:54:39] *** mrvisser has quit IRC[02:58:59] <wesolows> I'm not sure what's different here. I just provisioned this zone a short time ago.[03:08:38] *** rc10 has joined #smartos[03:09:29] *** enmand has joined #smartos[03:14:53] *** xforest has joined #smartos[03:15:36] <xforest> HI,when i mdb my nodejs with http.js ,it seem not work,[03:16:12] <xforest> how to fix it?[03:16:41] <rmustacc> What exactly are you tring to run with mdb?[03:17:40] <xforest> monitor http request parameter[03:18:08] <rmustacc> mdb doesn't actively monitor things, it's a debugger.[03:18:14] <rmustacc> So what command are you trying to run with mdb?[03:18:57] <xforest> yes, i want to debug http request, mdb -p pid[03:20:17] <rmustacc> Okay, and once you are attached with mdb, what do you try to do?[03:20:25] <rmustacc> Note that when you are attached like that the process is not running.[03:24:34] <jamesd> xforest, sounds like a better job for dtrace, but would be easier with dtrace probes embeded in it....[03:25:24] *** enmand has quit IRC[03:36:08] *** jelmd has joined #smartos[03:41:09] *** wdent has joined #smartos[03:48:52] <xforest> i run jsstack[03:48:59] <xforest> 8047b88 libc_hwcap1.so.1`_portfs+7[03:49:02] <xforest> 8047c18 port_poll+0x87[03:49:05] <xforest> 8047c58 uv__run+0x8d[03:49:07] <xforest> 8047c68 uv_run+0x17[03:49:10] <xforest> 8047cc8 node::Start+0x1cf[03:49:12] <xforest> 8047ce8 main+0x1b[03:49:15] <xforest> 8047d0c _start+0x83[03:49:17] <xforest> 8047b88 libc_hwcap1.so.1`_portfs+7[03:49:20] <xforest> 8047c18 port_poll+0x87[03:49:22] <xforest> 8047c58 uv__run+0x8d[03:49:25] <xforest> 8047c68 uv_run+0x17[03:49:27] <xforest> 8047cc8 node::Start+0x1cf[03:49:30] <xforest> 8047ce8 main+0x1b[03:49:32] <xforest> 8047d0c _start+0x83[03:49:48] <xforest> l cound't jsframe walk .[03:50:10] <xforest> my demo programe is run setInterval[03:50:42] <xforest> i want to debug the stack.[03:51:09] <xforest> dtrace just do sample the status.[03:51:43] <xforest> is that right?[03:52:20] <jamesd> xforest, you can ask dtrace to store or output ustack 1000's of times per second if you like[03:53:29] <jamesd> What are the top 5 kernel stack traces on CPU (shows why)?[03:53:29] <jamesd> dtrace -n 'profile-997hz { @[stack()] = count(); } END { trunc(@, 5); }'[03:53:47] <jamesd> you can change it to do ustack instead of the kernel stacktrace[03:54:09] <jamesd> and add a predicate that your process is on cpu at time its agregated.[03:55:02] <xforest> ustack can view the varaiable struct in memory like mdb?[03:56:16] <jamesd> yes i think it uses the same library calls to process the user stack[03:57:06] *** richlowe has joined #smartos[03:57:49] <xforest> great job,l'll try[03:58:19] *** richlowe has left #smartos[03:58:23] <jamesd> http://docs.oracle.com/cd/E19253-01/817-6223/chp-actsub-ustack/index.html[03:58:28] *** richlowe has joined #smartos[03:58:29] *** richlowe has left #smartos[03:58:46] *** richlowe has joined #smartos[04:08:43] *** wdent has quit IRC[04:16:28] <xforest> ustack can't look inside struct, it just sample the stack and static.[04:17:00] <xforest> i think if want to run jsframe,i must stop process[04:22:46] <xforest> does Dapsays in this IRC channel ?[04:28:14] *** wdent has joined #smartos[04:29:55] *** ryancnelson has joined #smartos[04:40:03] *** enmand has joined #smartos[04:40:44] *** xforest has quit IRC[04:53:22] *** sachinsharma has joined #smartos[04:59:45] *** ryancnelson has quit IRC[05:01:56] *** xforest has joined #smartos[05:02:21] *** xforest has quit IRC[05:02:37] *** xforest has joined #smartos[05:09:15] *** sachinsharma has quit IRC[05:09:30] *** sachinsharma has joined #smartos[05:18:38] *** jbergstroem has quit IRC[05:21:04] *** jbergstroem has joined #smartos[05:23:17] *** enmand has quit IRC[06:37:40] *** montyz0 has joined #smartos[07:03:43] *** wolstena has joined #smartos[07:05:40] *** rc10 has quit IRC[07:10:40] *** montyz0 has quit IRC[07:13:07] *** wolstena has left #smartos[07:21:43] <jbergstroem> I'm having shutdown issues with smartos. Everytime i halt/reboot, all zones gets improperly shutdown. I notice this by getting recovery on all drives. ubuntu's, other linuxes, freebsd, openbsd, openwrt, whatnot. How do I provide useful enough information for a bug report? If i for instance do vmadm stop <uuid> , the vm is properly shutdown[07:25:25] <rmustacc> Are you telling the system to do an clean shutdown?[07:25:30] <rmustacc> Or are you just typing reboot?[07:31:15] <MerlinDMC> newlix, rbenv works without much problems ... rvm doesn't work that easily just to be complete :)[07:34:15] <jbergstroem> rmustacc: define clean shutdown[07:36:29] <rmustacc> What you seem to want.[07:36:34] <rmustacc> One where every service is cleanly disabled.[07:36:39] <rmustacc> Every zone is orderly shut down.[07:39:22] *** rc10 has joined #smartos[07:39:59] *** rc10 has left #smartos[08:00:38] *** Cpt-Oblivious has quit IRC[08:19:03] *** jamesd has quit IRC[08:30:01] *** sachinsharma has quit IRC[08:44:58] *** alucardX has joined #smartos[08:45:07] <alucardX> morning[08:48:29] *** kschiess has joined #smartos[08:50:27] <jbergstroem> rmustacc: sorry for the afk. i use reboot or halt, both tells me they're about to stop zones. I can see zones stopping but then i loose access to both zones (doh) and the host, so it's hard to say whats going on. It's apparently not shutting them down clean though.[09:03:38] <MerlinDMC> jbergstroem, there is a timeout for that shutdown to happen that is to short for your vms ... but i guess thee is also another command to shutdown the host in a cleaner way[09:08:27] <konobi> iirc, shutdown -y -i5 -g0[09:08:49] <konobi> or -i6 for reboot[09:23:08] *** texarcana has quit IRC[09:24:24] *** texarcana has joined #smartos[09:25:16] *** masked has quit IRC[09:26:50] *** sachinsharma has joined #smartos[10:01:26] *** bens1 has joined #smartos[10:19:12] *** mamash has joined #smartos[10:23:53] *** alcir has joined #smartos[10:24:01] <alcir> how to boot[10:24:06] <alcir> and disable acpi?[10:24:13] <alcir> disable-acpi=true?[10:26:58] *** riceo has joined #smartos[10:38:47] <alcir> good[10:48:54] *** xenol has quit IRC[10:49:01] *** xenol has joined #smartos[10:55:23] <konobi> huh, so no Aram for a while... *sigh*[10:55:27] *** mamash has left #smartos[11:07:19] *** xinkeT has quit IRC[11:08:32] *** xinkeT has joined #smartos[11:18:01] <MerlinDMC> aram?[11:18:40] *** relling has quit IRC[11:20:29] <konobi> MerlinDMC: guy who was debugging a common usb/acpi style issue[11:23:38] *** Cpt-Oblivious has joined #smartos[11:26:42] *** tanngo has joined #smartos[11:28:17] <tanngo> has anyone gotten ipv6 working inside a virtual machine running ubuntu under smartos? i just cant figure it out[11:30:58] <MerlinDMC> allow_ip_spoof: true and configure the ipv6 inside the VM[11:31:48] <tanngo> I just did that. I think. I'll have another look.[11:32:06] <tanngo> vmadm update 91a10b3f-476f-46d6-afb2-660dbcbb35d4 nics.*.allow_ip_spoofing=true[11:32:38] <tanngo> and then rebooted the VM[11:34:23] <MerlinDMC> you would need to use a update_nics json ... subproperties can't be updated on the command line afaik[11:34:35] <tanngo> ah, that explains it[11:34:48] <MerlinDMC> just check if you have allow_ip_spoofing if you do a vmadm get <uuid>[11:35:46] <tanngo> you're right, I don't[11:35:53] <tanngo> tnx for the help[11:40:24] <tanngo> it works now, sweet =)[11:51:27] *** mamash has joined #smartos[11:51:42] *** alucardX has quit IRC[11:52:29] *** alucardX has joined #smartos[11:55:55] *** KermitTheFragger has joined #smartos[12:05:24] *** relling has joined #smartos[12:12:53] *** mrvisser has joined #smartos[12:18:29] *** xinkeT has quit IRC[12:18:46] *** xinkeT has joined #smartos[12:37:21] *** mamash has left #smartos[13:03:57] *** tanngo is now known as linuxprofessor[13:14:57] *** killfill_ is now known as killfill[13:29:25] *** wdent has quit IRC[13:38:05] *** khushildep has joined #smartos[13:47:25] *** xmerlin has joined #smartos[13:48:08] *** mrvisser has quit IRC[13:53:42] *** mrvisser has joined #smartos[13:55:18] *** ripratm has quit IRC[14:00:00] *** enmand has joined #smartos[14:00:36] <mrvisser> Hi all, when I try and install cairo on a Joyent base 1.8.1 smartos node, I end up with this error: http://pastebin.com/3t0FtRCk[14:00:45] <mrvisser> I can't seem to get around it. any advice?[14:01:20] <jperkin> solaris ld cannot handle empty files[14:02:01] <jperkin> I already fixed that in pkgsrc, see https://github.com/joyent/pkgsrc/blob/trunk/graphics/cairo/patches/patch-ad[14:02:54] <mrvisser> ahh, thanks jperkin. does joyent have any desire to upgrade their pixman+cairo versions?[14:03:51] <xmerlin> hi, what's the correct json syntax to remove only 1 parameter (such gateway) in a nic?[14:04:01] <jperkin> mrvisser: pkgsrc creates a stable branch every 3 months and we generally follow that, I believe pkgsrc has the latest cairo now.[14:04:42] <jperkin> so, 1.8.x won't get a cairo upgrade but there will be an updated dataset with a newer one.[14:05:13] <mrvisser> jperkin: I see this: http://pkgsrc.joyent.com/sdc6/2012Q2/i386/All/cairo-1.10.2nb5.tgz[14:05:32] <mrvisser> the latest is 1.12.8 I believe[14:05:55] <jperkin> right, pkgsrc current is at DISTNAME= cairo-1.12.8[14:06:17] <mrvisser> ah, so the Q3 should have the latest?[14:06:25] <mrvisser> (when it is released?)[14:06:29] *** alcir has quit IRC[14:06:33] <jperkin> Q4, I don't recall what Q3 had..[14:06:37] <mrvisser> *doesn't understand this stuff well..*[14:06:50] <mrvisser> there doesn't seem to be a link for Q3[14:07:07] <mrvisser> atleast not joyent's[14:07:08] <jperkin> filip had Q3 mostly done but then had to finish off some other stuff first[14:07:19] <mrvisser> I see[14:07:36] <mrvisser> jperkin: is there a way I can have a look at the "current" ?[14:07:44] <jperkin> alternatively you could use http://www.perkin.org.uk/posts/pkgsrc-2012Q3-packages-for-illumos.html which has 1.12.2[14:08:17] <mrvisser> ok, I think I'll give that a try[14:08:24] <mrvisser> thanks jperkin, it's all clear to me now[14:09:34] *** wdent has joined #smartos[14:09:51] <jperkin> it's somewhat confusing due to a number of factors, but we hope to simplify and streamline things over time[14:10:32] <mrvisser> yea.. package management ain't easy I imagine.[14:23:49] *** kamilr has joined #smartos[14:24:11] <kamilr> Hi, is there any official WishList for smartos ?[14:29:53] *** enmand has quit IRC[14:33:25] *** enmand has joined #smartos[14:48:41] *** khushildep has quit IRC[14:49:22] *** wdent has quit IRC[14:54:27] *** wdent has joined #smartos[14:55:16] <MerlinDMC> kamilr, the mailing list ... and if approved a github issue? ;)[14:55:28] *** khushildep has joined #smartos[15:12:29] *** nikolam has joined #smartos[15:17:10] *** jim80net has joined #smartos[15:20:57] *** kamilr has quit IRC[15:21:02] *** wolfeidau has joined #smartos[15:21:42] <wolfeidau> Anyone know of the top of their head which package g++ is installed via? Cant find the damn thing..[15:21:54] <jperkin> gcc47[15:22:26] <wolfeidau> jperkin: Cool thanks will give it a go[15:26:17] <mrvisser> Do you expect that the packages from http://pkgsrc.smartos.org/packages/SmartOS/2012Q3-multiarch/All will work on a 32-bit base 1.8.1 image?[15:27:37] <jperkin> mrvisser: multiarch is 'special', I wouldn't recommend you use them right now[15:28:40] <mrvisser> jperkin: then the sngl?[15:29:03] <jperkin> mrvisser: no, not that either :)[15:29:25] <mrvisser> hehe.. stick to the instructions on the blogpost you sent me earlier?[15:29:29] <jperkin> yep :)[15:29:36] <mrvisser> alright sounds good :)[15:29:50] <jperkin> I should probably just remove those, they won't really be useful until Q4.[15:48:26] *** CarlosC has joined #smartos[15:50:33] <mrvisser> hmm.. somehow the machine doesn't have make. I've pkgin in gmake automake[15:51:56] <CarlosC> use gmake[15:52:41] <MerlinDMC> mrvisser, reload your shell and try again ... maybe just some hickup?[15:53:49] <mrvisser> nah, didn't work, maybe the place it added isn't on my path[15:53:56] <mrvisser> no make in /opt/local/gnu/bin /opt/local/bin /opt/local/sbin /usr/bin /usr/sbin[15:54:29] <mrvisser> gmake doesn't seem to be on the path either[15:55:16] <CarlosC> gmake should be in /opt/local/bin[15:55:21] <mrvisser> after logging out and logging back in those packages I installed are gone..[15:55:25] <mrvisser> *scratches head*[15:55:34] <CarlosC> pkgin ls[15:56:07] <mrvisser> that's what I did, no automake or gmake anymore[15:56:34] <CarlosC> try installing just gmake[15:57:01] *** mamash has joined #smartos[15:57:52] <mrvisser> my pkgsrc was reverted as well. I had just set up jperkin's repo: http://www.perkin.org.uk/posts/pkgsrc-2012Q3-packages-for-illumos.html[15:58:12] <jperkin> mrvisser: right, those are installed to /opt/pkg so you'll need to add /opt/pkg/bin:/opt/pkg/sbin to $PATH[15:58:35] <jperkin> mrvisser: see http://www.perkin.org.uk/posts/9000-packages-for-smartos-and-illumos.html for more background[15:58:42] <mrvisser> ok, thanks[15:59:09] <mrvisser> jperkin: is it expected that after I log out and in again, all the packages I downloaded would no longer be found in pkgin as well?[16:00:00] <jperkin> mrvisser: it'll be because using an unmodified $PATH you are calling the 'pkgin' which came with the base machine, which isn't the one you used to install those extra packages[16:00:28] <jperkin> /opt/local/bin/pkgin for the base ones, /opt/pkg/bin/pkgin for the extra ones[16:00:41] *** neophenix has joined #smartos[16:02:16] <mrvisser> ah I see, didn't realize I installed a new pkgin[16:02:22] <mrvisser> thanks again[16:02:30] *** iyp has quit IRC[16:02:50] <jperkin> yeh, I should probably add a small explanatory section to each update, I just figured most people who read that post had already read the previous one[16:10:18] *** Cpt-Oblivious has quit IRC[16:18:38] *** iyp has joined #smartos[16:20:49] *** linuxprofessor has quit IRC[16:25:31] <mrvisser> jperkin: would there be any problem with just using the Q3 as the base?[16:28:39] *** jim80net has quit IRC[16:30:12] <jperkin> mrvisser: generally no, the main difference will be if you want to run daemons then you will have to do it by yourself as those packages do not come with SMF bits.[16:30:38] *** nikolam has quit IRC[16:38:12] *** nikolam has joined #smartos[16:38:41] *** jamesd has joined #smartos[16:39:14] *** rancor has joined #smartos[16:41:13] *** daleg has joined #smartos[16:41:13] *** d[^_^]b has quit IRC[16:50:25] <mrvisser> looks like installing gcc47 doesn't create a g++ binary[16:51:49] *** nikolam has quit IRC[16:52:17] <jperkin> it should set up a symlink in /opt/local/bin, if not you can always get to it directly in /opt/local/gcc47/bin[16:52:46] <jperkin> ah, though, if you are using the /opt/pkg package set then yeh, that doesn't set up a symlink (another of the differences)[16:52:57] <jperkin> so will need to call it from /opt/pkg/gcc47/bin[16:53:31] <mrvisser> hmm, right[16:53:32] *** relling has quit IRC[16:53:37] <mrvisser> I usually don't call that manually[16:53:49] <mrvisser> it's normally in other scripts that I compile.[16:54:21] <mrvisser> I'm having a hard time figuring out how I'm going to get puppet to do this for 8 application nodes as well :P I may have to find a better way to automate this[16:54:56] <mrvisser> maybe DL, patch compile[16:55:48] <mrvisser> jperkin: do you have any idea of an eta to have updated packages on the base joyent pkgsrc?[16:55:57] *** tonyarkles has joined #smartos[16:56:19] <jperkin> mamash: ^^ :)[16:59:33] *** alucardX has quit IRC[17:01:10] *** linuxprofessor has joined #smartos[17:08:26] *** newlix_ has joined #smartos[17:08:42] <newlix_> what is the default root password of the percona server ?[17:11:20] *** montyz0 has joined #smartos[17:13:47] *** jim80net has joined #smartos[17:18:01] <CarlosC> newlix_: are you on JPC?[17:23:08] *** ismell has joined #smartos[17:24:01] <chris---> so i'm back to imgadm not being able to talk through a proxy server[17:27:33] <xmerlin> is it possible to use more than 1 zpool in smartos? ...if so how can I tell to vmadm to use the pool 2 instead of pool 1?[17:29:02] <mrvisser> jperkin: those packages seem to be 64bit, assuming there is no 32bit alternative?[17:30:03] *** kfr- has quit IRC[17:32:20] *** kfr- has joined #smartos[17:34:06] <xmerlin> xmerlin ...I've found the answer ;)[17:34:11] <chris---> so there is a patch in bithub for this[17:34:23] <jperkin> mrvisser: no, but I do build 32bit packages, it'd just be a matter of publishing them[17:35:03] <mrvisser> that would be awesome if you could do that[17:36:45] <jperkin> and they are from pkgsrc current, so will be newer[17:37:05] <mrvisser> even better, I think! :)[17:38:55] <jperkin> ok, it'll probably be wednesday before they are available though[17:41:17] <mrvisser> ok, that will be helpful. the reason we're using 32bit in this environment is because c++ symbols are obfuscated in d-trace. so if that can be fixed alternatively, then that would work too :)[17:41:28] <mrvisser> no translator for 64bit[17:41:58] <rmustacc> C++ symbols are always obfuscated.[17:42:04] <rmustacc> It's based on the design of the language.[17:42:18] <rmustacc> Which translator is it specifically you're looking for with 64-bit support?[17:42:48] <mrvisser> http://blog.nodejs.org/2012/04/25/profiling-node-js/[17:42:58] <mrvisser> c++filt[17:43:03] <rmustacc> That isn't the issue.[17:43:08] <rmustacc> There is no ustack support for 64-bit node.[17:43:12] <rmustacc> Err, no jstack[17:43:31] <rmustacc> That's the unfortunate problem at this time.[17:44:17] <mrvisser> ah I see, thanks for the tip, assumed it was the filter[17:44:47] <rmustacc> The problem you're probably seeing is that you're not getting JS frames there, but you do get mangled C++ frames.[17:44:59] <rmustacc> The filter is pretty much the same regardless of bitness.[17:51:09] *** mamash has left #smartos[17:51:36] *** newlix_ has quit IRC[17:53:46] *** tonyarkles_ has joined #smartos[17:56:36] *** tonyarkles has quit IRC[17:56:37] *** tonyarkles_ is now known as tonyarkles[18:09:41] *** tonyarkles has quit IRC[18:18:03] *** wramthun2 has joined #smartos[18:18:41] *** tonyarkles has joined #smartos[18:21:16] *** wramthun has quit IRC[18:21:32] *** dap has joined #smartos[18:28:50] *** tonyarkles has quit IRC[18:29:10] *** daleg has quit IRC[18:29:13] *** notmatt has joined #smartos[18:31:30] *** tonyarkles has joined #smartos[18:32:09] *** mrvisser has quit IRC[18:36:18] <iyp> Can anyone point me at some docs for making my own package in pkgsrc?[18:36:34] <iyp> I have the joyent pkgsrc repo and pk installed already[18:37:45] *** tonyarkles has quit IRC[18:38:51] <xmerlin> iyp, I've found these guides http://wiki.smartos.org/display/DOC/Building+Packages and https://github.com/mamash/pk/wiki[18:39:53] *** wdent has quit IRC[18:40:50] <iyp> Yeah, those are solid. They've helped me as far as compiling packages that are already in the system. I haven't found much help on adding creating new packages to the pkgsrc tree. Guess thats just reading I have to about the layout of pkgsrc dirs[18:42:15] <jperkin> there is the pkgsrc guide which goes into detail, but generally the best way is just to copy+edit an existing package with similar functionality[18:44:02] <iyp> Gotchya. I'm trying to add a database (http://www.monetdb.org/), so I've been looking around in the mysql directory. Still can't figure out where the source is :/[18:45:14] <jperkin> mysql is somewhat complicated as it's spread across multiple directories, something like databases/sqlite3 would be easier[18:48:45] *** KermitTheFragger has quit IRC[18:49:22] <iyp> Roger that. Thanks for the tip[18:49:45] *** mamash has joined #smartos[18:51:16] <iyp> Err. Still can't find the sources. Closest thing I've found are patches.[18:51:53] <jperkin> you don't store the actual distfiles in pkgsrc, it fetches them on demand[18:52:10] <iyp> Well that makes sense now.[18:52:28] <jperkin> you just put directives on where to fetch from, $MASTER_SITES/$DISTNAME.$EXTRACT_SUFX (where EXTRACT_SUFX defaults to tar.gz so you generally don't need to set it)[18:53:08] <iyp> Which file is that set in?[18:53:10] <jperkin> a 'bmake makesum' will fetch the intial version and create a 'distinfo' file with the size/sha1 etc[18:53:13] <jperkin> Makefile[18:53:37] <iyp> Ah, now I see it.[18:53:42] <iyp> Cool[18:55:39] <iyp> So more or less, it's as simple as setting those variables properly in Makefile and setting the proper build options in options.mk?[18:55:55] *** gkyildirim has joined #smartos[18:56:35] <jperkin> options.mk is only for if you want tuneables in how the package is built. common options go in e.g. CONFIGURE_ARGS[18:57:04] <jperkin> i.e. if you want to give users the option to build with gui support, or whatever.[18:57:30] <iyp> Yeah, there are a set of options I had to feed to configure to compile by hand properly on smartos. Those would go in CONFIGURE_ARGS, yeah?[18:57:41] <jperkin> yep, and they'll get passed to ./configure[18:57:46] <iyp> perfect[18:59:46] <iyp> How would I go about rebuilding a package. I keep getting "skipped 1 packages built previously" when I run pk build[19:00:11] <jperkin> sorry, I haven't actually used pk[19:00:44] <jperkin> in native pkgsrc as long as a work directory isn't still lying around then a 'bmake' inside it will rebuild the package[19:01:05] <iyp> Just what I was looking for. Thank you[19:01:14] <jperkin> tea time, bbl.[19:01:28] <iyp> Sounds good. Thanks for the help.[19:04:25] <mamash> iyp: you can do 'pk build -f' to force a build[19:04:31] *** tonyarkles has joined #smartos[19:05:32] <iyp> mamash: yup that works too[19:05:50] <iyp> still haven't figured out why pkgin won't fetch my pkg_summary[19:07:11] *** xforest has quit IRC[19:08:53] <iyp> mamash: is there any particular web server you guys use when serving up a repository?[19:09:20] <iyp> I'm using the node and the http-server package from npm. Think that could be part of the problem?[19:18:21] <mamash> we use lighttpd or nginx[19:19:20] <mamash> hm could be part of the problem[19:19:39] *** wramthun3 has joined #smartos[19:19:51] *** wramthun2 has quit IRC[19:19:54] *** mrvisser has joined #smartos[19:20:22] <linuxprofessor> is there any way to save the contents of /root on smartos so that it survives a reboot? would be nice to keep my authorized_keys for ssh[19:23:02] *** tonyarkles has quit IRC[19:23:09] <MerlinDMC> linuxprofessor, you can configure authorized keys in the config file[19:23:29] <linuxprofessor> ah, sweet[19:23:37] <MerlinDMC> config_inc_dir=config.inc[19:23:38] <MerlinDMC> root_authorized_keys_file=authorized_keys[19:24:05] <MerlinDMC> inside the config ... and put your authorized keys file to /usbkey/config.inc/authorized_keys[19:24:31] <MerlinDMC> you could also change the folder/file names ... but ... yeah[19:25:31] <linuxprofessor> thanks[19:29:41] *** wdent has joined #smartos[19:31:33] *** kschiess has quit IRC[19:38:17] *** wolstena has joined #smartos[19:42:07] <chris---> merlinDMC: just tried your patched imgadm.js for the www proxy stuff, worked[19:42:58] <MerlinDMC> chris---, yep ... but i heard somewhere that imgadm is getting a rewrite ... so that patch is not included in the release ;)[19:44:01] <chris---> will the rewrite do proxy though? :)[19:44:20] <chris---> i started to do what you did, then found what you had done, gets me by for now[19:45:06] <chris---> wondered why it was out there for a month and not picked up yet[19:45:56] <MerlinDMC> chris---, i did the pull req later than i had the code ... if i remember correctly[19:46:22] <MerlinDMC> and if the new one (if there is one) will have proxy support ... i don't know[19:46:26] <chris---> ah i just see the "a month ago"[19:46:53] <chris---> without it I'm stuck[19:47:08] <MerlinDMC> i did that patch after someone here requested proxy support ... i really never need that tho[19:47:26] <MerlinDMC> but nice that it is usable for someone else as well :P[19:47:31] <chris---> these systems wont have a real address on them at all[19:47:37] <chris---> yeah it helps me a bunch[19:48:09] <chris---> import just puked, see what i can make of that[19:48:34] <chris---> Error: zfs receive exited non zero: 1[19:49:27] <MerlinDMC> you should kill partial received datasets first ... zfs list -t all ... and mostly everything with a -partial is broken ;)[19:50:04] *** bens1 has quit IRC[19:50:20] <chris---> i dont see anything with a -partial[19:52:12] <MerlinDMC> and the dataset you want to import is not already there in a broken form?[19:53:20] <chris---> not as far as zfs is concerned, first import i've tried on this thing[19:54:55] *** daleg has joined #smartos[19:55:21] <chris---> tried another random img and it did the same[19:55:35] <chris---> heh, this has been a battle every step of the way[19:55:55] *** mrvisser has left #smartos[19:56:20] <MerlinDMC> and you did patch imgadm by mounting the patched imgadm.js over the original one?[19:56:22] <konobi> MerlinDMC: a patch to put the images under a child dataset of zones would be nice though =0)[19:57:19] <chris---> yeah i mounted the usb, then the boot_archive, then inside the usr.lgz and diddled it[19:57:42] <MerlinDMC> konobi, another zpool would be possibly by just changing one variable on ENV or parameter inside imgadm ... but in a child dataset ... hmm[19:58:43] <MerlinDMC> i had some other things for imgadm as well but i don't want to put time into that while it gets a rewrite[19:59:27] <konobi> zones/imported_images/<uuid>[19:59:35] <chris---> i guess no i didnt mount over the old one, i peeled it open and did it inside[20:01:06] <MerlinDMC> konobi, but i guess that will also break vmadm[20:01:38] <MerlinDMC> someone wanted to be able to import to a zpool named different than zones because he wanted to spin up vms on that one[20:01:56] <MerlinDMC> and cloning does not work between pools i guess :P[20:02:10] <rmustacc> I don't really know that we want to support a different pool though.[20:02:18] <konobi> yeah[20:02:45] <konobi> i'm suggesting child dataset, just to keep the images and zone datasets out of the same namespace[20:03:50] <rmustacc> More of a flag day pain than anything else.[20:03:58] <konobi> mmm[20:04:30] <MerlinDMC> i really never look into /zones or at zfs list ... so i mostly don't care about where they are saved :P[20:04:44] <MerlinDMC> as long as my zones work[20:04:47] *** wdent has quit IRC[20:05:52] <konobi> things like having zones/<uuid>/base and zones/<uuid>/diskN would also be nice, but that'd be a bigger task[20:06:28] <MerlinDMC> rmustacc, are you guys internally working on ipv6? ... especially the vnic_dhcp thing[20:07:53] <rmustacc> MerlinDMC: I don't know if the work is currently going on, but I know it's work we want to go do.[20:09:32] <MerlinDMC> is there any txt material for this work ... specifications for file naming inside zones for example? ... i did cleanup joshies ip6 stuff for myself already and was then sitting infront of the qemu dhcp thing to give up[20:09:55] <MerlinDMC> but it would be nice to have a scheme for filenames that would match your later stuff already[20:10:11] <rmustacc> With respect to inside of kvm or smartos zones?[20:10:31] <rmustacc> For kvm there is no separate file that gets used for this stuff.[20:10:36] <MerlinDMC> right now i did not introduce any new files for configuring zones for ipv6[20:10:56] <rmustacc> If you have a smartos zone you should just configure it as a static ip.[20:11:04] <rmustacc> Whatever the 'correct' way is to do that.[20:11:15] <rmustacc> Probably just using the existing hostname.net0 stuff there is.[20:11:22] <rmustacc> For qemu you need to have the dhcp get smarter.[20:11:28] *** xforest has joined #smartos[20:11:57] <nahamu> MerlinDMC: joshie would remember better, but I think there was a desire to make it look kind of like it would on other solaris-like systems (e.g. using /etc/hostname6.<interface> or whatever...[20:13:25] <konobi> ipadm?[20:13:37] <chris---> may be it for me today, have to save up some energy to peel this apart and put in some debugging[20:14:05] <MerlinDMC> rmustacc, k ... that's the part i'm missing currently - was hoping for a already available starting point :P but it's ok ... i'll try to get that smart in an (in)sane way[20:14:43] <rmustacc> Starting point for dhcpv6?[20:14:43] <MerlinDMC> nahamu, he told me yep ... but at that time i had my stuff running already ;)[20:14:54] <konobi> for qemu, it's a question of wether you can do discovery via the vnic[20:15:09] <rmustacc> Well, I don't know enough of what has changed with v4 versus v6 and dhcp.[20:15:24] <rmustacc> I don't know if we need network discovery or if we'll find we want to do something entirely different.[20:16:51] <MerlinDMC> konobi, it seems to be some code that just answers for the single vm on the vnic ... so possibly there should be a possibility to send almost everything on the wire[20:16:51] *** bens1 has joined #smartos[20:17:37] <MerlinDMC> if i get some vacation i'll play with that ^^[20:19:43] *** bens1 has quit IRC[20:25:01] <chris---> how can i get to the log.debug() stuff alrady in imgadm?[20:29:06] *** xer0x has joined #smartos[20:31:45] *** tonyarkles has joined #smartos[20:35:33] *** xforest has quit IRC[20:37:14] <iyp> Not sure what to make of this error I get when compiling something in the pkgsrc tree.[20:37:15] <iyp> bmake: don't know how to make /tmp/pkgsrc/databases/monetdb/work/.package_makevars.mk. Stop[20:40:30] *** xforest has joined #smartos[20:41:29] <konobi> think it'd be bad form to send a link to my blog entry to that guy on the mailing list?[20:44:04] *** russiane39 has joined #smartos[20:46:23] <MerlinDMC> konobi, not if it's helpful :)[20:47:20] <konobi> done[20:49:00] *** ryancnelson has joined #smartos[20:52:46] <jbergstroem> konobi, MerlinDMC : ok, will try. thanks.[20:53:25] <chris---> merlindmc: found my mistake, seems to work fine now, thanks again[20:53:38] <jperkin> iyp: sounds like you are setting something you shouldn't, can you dump your Makefile somewhere?[20:53:43] *** riceo has quit IRC[20:55:25] <konobi> jbergstroem: eh?[20:56:56] *** Cpt-Oblivious has joined #smartos[20:58:26] *** strebe has joined #smartos[20:59:35] *** sachinsharma has quit IRC[21:00:26] <iyp> jperkin: Yeah, i'm pretty sure I messed something up. Please excuse the gratuitous naievity.[21:00:38] <strebe> I'd like to import an existing pool readonly from a livecd. A friend told me that smartos would work well for this, but booting it, I get something that looks like an installer, which is asking me to select disks for the storage pool. Is there a way to do what I want to from the smartos cd, or am I using the wrong tool entirely?[21:00:44] <iyp> Make file https://gist.github.com/f2b17aec9d47ab582a67[21:00:52] <iyp> options.mk https://gist.github.com/2c1621eaf80addd24e24[21:00:57] <rmustacc> strebe: You'll need to specify noimport=tue[21:01:11] <strebe> rmustacc: in the grub options, or..?[21:01:19] <rmustacc> Yes[21:01:22] <strebe> Thank you[21:01:32] <rmustacc> I think that'll kick you out of the installer.[21:02:04] <jperkin> iyp: you must include bsd.pkg.mk[21:02:12] <jperkin> that's the 'guts' of pkgsrc[21:02:13] <rancor> correct. there's also a menu entry in grub for noinstall=true.[21:02:24] <rancor> second one, iirc.[21:03:05] <iyp> jperkin: gotchya. That seemed to have helped. Different error: *** Error code 255[21:03:16] <jperkin> iyp: also if you aren't doing any OPSYS-based work then you don't need to include bsd.prefs.mk, and if this is for current pkgsrc then you don't need PKG_DESTDIR_SUPPORT[21:03:45] *** montyz0 has quit IRC[21:04:11] <iyp> jperkin: same error if I cut those two out.[21:04:30] <jperkin> iyp: I'd move the CONFIGURE stuff in options.mk into Makefile, remove the /opt/local bits, and un-include options.mk, it's probably an issue if you include bsd.options.mk without setting PKG_* at the top[21:05:39] <iyp> Install went to /usr/local if I didn't specify prefix when I was hand compiling.[21:05:55] <jperkin> iyp: as for --with-unixodbc, all you need to do is add .include "../../databases/unixodbc/buildlink3.mk"[21:06:08] <jperkin> yes, /usr/local is the ./configure default, but pkgsrc will ensure /opt/local is used[21:07:15] <jperkin> and the buildlink stuff ensures that unixodbc is pulled in as a dependency and that the headers/libraries are available for configure to find[21:08:06] <iyp> Awesome![21:08:23] <iyp> Ok getting farther now. Getting to the fetch the code step.[21:08:31] <iyp> Not resolving the domain.[21:08:33] <iyp> d'oh[21:10:02] <jperkin> afk for a while, one thing you'll need to do is build a PLIST file, so once 'bmake' is working, do 'bmake print-PLIST >PLIST', then hopefully a 'bmake package' will work - set PKG_DEVELOPER=yes in /opt/local/etc/mk.conf to add extra debug checks.[21:10:29] <iyp> jperkin: Gotchya. I really appreciate all the help.[21:14:26] *** mamash has left #smartos[21:24:16] *** relling has joined #smartos[21:24:30] <iyp> Cant ping anything not on my smartos box.[21:24:39] <iyp> GZ can, but any other zone can't.[21:25:11] <iyp> Wasn't doing this an hour ago…arg[21:37:31] *** xforest has quit IRC[21:42:51] *** xforest has joined #smartos[22:15:21] *** bens1 has joined #smartos[22:17:48] *** xer0x has quit IRC[22:21:46] *** relling has quit IRC[22:29:05] *** ira has joined #smartos[22:29:07] <jbergstroem> konobi: referred to you pointer re proper shutdown. shutdown -y -i5 -g0 didn't work.[22:29:27] <jbergstroem> trying reboot -i6 now[22:34:35] *** wolfeidau has quit IRC[22:39:11] *** bens1 has quit IRC[22:42:29] <konobi> ah[22:43:21] <jbergstroem> (reboot doesn't accept -i6)[22:43:53] *** xer0x has joined #smartos[22:45:06] *** relling has joined #smartos[22:47:23] *** bdha_ is now known as bdha[22:48:34] *** jamesd_ has joined #smartos[22:49:17] *** CarlosC has quit IRC[22:49:53] <konobi> jbergstroem: shutdown -i6 (for reboot)[22:50:05] <konobi> along with the rest of the options[22:52:54] *** jamesd has quit IRC[22:53:44] <jbergstroem> konobi: ah. sorry. will try, but I'm pretty sure I'll get the same problems[22:54:45] <konobi> jbergstroem: may also be that your vm's aren't supporting acpi properly[22:55:27] <jbergstroem> konobi: already checked that. all respond to acpi if i do vmadm stop <foo> (and that also works as intended)[22:58:55] <jbergstroem> konobi: Same result, unfortunately[23:03:47] *** drew has joined #smartos[23:06:56] <konobi> jbergstroem: "Using the init program to restart a system causes Solaris to perform a sequence of steps to shutdown services gracefully", that's what the -i option is[23:07:47] <konobi> are you on a recent enough smartos version?[23:08:09] *** xer0x has quit IRC[23:08:47] *** drew has quit IRC[23:11:51] <jbergstroem> konobi: joyent_20121203T193049Z[23:12:09] <jbergstroem> konobi: i've experienced this for a pretty long time[23:12:38] <jbergstroem> konobi: not sure how to give meaningful feedback though :([23:15:16] *** tonyarkles has quit IRC[23:16:07] <konobi> jbergstroem: does it take 5 mins for a vm to shutdown?[23:16:56] <ryancnelson> it's only graceful if the service shutdown command is graceful… it's not magic[23:18:09] <konobi> oh, i worked out the acpi-ism for windows... i should note that on the wiki somewhere[23:18:59] <jbergstroem> konobi: no, all of them are pretty quick. the host machine stops responding within a minute, too.[23:19:38] <konobi> i haven't had any issues with the linux vms on my box i have to say[23:21:49] *** wolfeidau has joined #smartos[23:22:39] <konobi> jbergstroem: so if you vm stop/start all of your VMs, they all come up cleanly?[23:26:00] <jbergstroem> konobi: that's it. i'll try all of them again[23:26:27] <konobi> (without rebooting that is)[23:27:25] <jbergstroem> konobi: yep.[23:30:22] *** wolfeidau has quit IRC[23:30:47] *** ryancnelson has quit IRC[23:32:13] *** wdent has joined #smartos[23:43:18] *** iyp has quit IRC[23:44:37] *** xer0x has joined #smartos[23:45:23] <jbergstroem> konobi: my obsd5.2 vm seems to take an awefully long time to return (from vmadm stop <foo>), but the acpi signal is sent and vm stops responding almost immediately. could this affect all other shutdowns too?[23:47:20] *** wdent has quit IRC[23:50:26] *** wdent has joined #smartos[23:52:52] *** tonyarkles has joined #smartos[23:53:43] <konobi> jbergstroem: sounds like it's timing out (5 minutes), at which point it will be force killed[23:53:57] *** wolfeidau has joined #smartos[23:57:14] <konobi> jbergstroem: it may be that openbsd is halting, but not powering down[23:58:36] <konobi> jbergstroem: http://www.openbsd.org/faq/faq10.html # check rc.shutdown