September 13, 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 13, 2013 [00:12:07] *** jim80net has joined #smartos[00:13:11] *** Alek2 has quit IRC[00:19:40] *** tbatchelli_m has quit IRC[00:21:52] *** tbatchelli_m has joined #smartos[00:27:58] <yo61> I've always wondered why vmadm lists a VM for the host machine[00:28:57] <yo61> I inadvertently did a "vmadm delete ..." on it yesterday, then hit ctrl-C quickly when I realised[00:29:20] <yo61> It's now showing as "incomplete" and I've noticed it's got type "KVM"[00:29:38] <rmustacc> yo61: I'm not sure I understand what you mean by a VM for the host machine.[00:29:55] <rmustacc> I don't see it list anything for the GZ.[00:30:09] <yo61> vmadm list shows a VM with the same name as the SmartOS host[00:30:21] <rmustacc> What is the output of vmadm get for that?[00:31:03] *** socketwi has quit IRC[00:31:17] <yo61> https://gist.github.com/robinbowes/6544696[00:31:28] <Meths> Compare with zoneadm list output which does give you the normal extra 'global'[00:31:39] *** socketwiz has joined #smartos[00:32:12] <Meths> That just looks like a vm with an alias that matches the host.[00:32:17] <yo61> OK[00:32:48] <yo61> I don't recall how it got created...[00:33:24] <rmustacc> Yup, Meths has it.[00:33:55] <yo61> It's gone.[00:33:59] <yo61> Thanks[00:35:42] <yo61> I wish I could create native zones - I fancy running some stuff that could do with the extra performance (logstash/elasticsearch)[00:36:48] <Meths> Why can't you?[00:37:04] <yo61> Provisioning fails[00:37:26] <yo61> No idea why[00:37:35] <Meths> No error?[00:37:40] <yo61> Get's stuck really early in the process[00:37:49] <yo61> vmadm create just times out[00:38:11] <Meths> Is the image you're trying to use available?[00:38:18] <yo61> Yes, it is[00:38:26] <Meths> and imported?[00:38:41] <yo61> I can zlogin to the zone when it's sat there waiting for the provisioning to finish[00:38:47] <Meths> Oh[00:40:17] <Meths> That sounds like something that's been mentioned over the last couple of days. If you run svcs in the zone is there just one online and one offline and the rest just -[00:40:28] <yo61> Yeah - it was probably me mentioning it[00:40:49] <Meths> Ah, okay.[00:41:37] <yo61> The last bit of output from svcs -xv is: https://gist.github.com/robinbowes/6544775[00:41:45] *** ImTheBitch is now known as There[00:42:22] <Meths> Anything in the mentioned log file?[00:42:30] *** There is now known as ImTheBitch[00:42:37] <yo61> Trying to gist it now[00:42:40] <Meths> Have you already popped all this in a bug report?[00:42:45] <yo61> No[00:43:04] <yo61> No-one's been able to reproduce it[00:43:19] <yo61> Here you go: https://gist.github.com/robinbowes/6544781[00:43:24] <rmustacc> That doesn't mean don't file a bug.[00:43:29] <yo61> True[00:43:36] <yo61> I've also been a bit busy[00:43:46] <rmustacc> That's fine.[00:44:03] <rmustacc> I just want to make sure that folks know that we'd much rather bugs be filed than not filed.[00:44:07] <yo61> Where's best place to file a report?[00:44:26] <yo61> github?[00:44:35] <rmustacc> Yeah, that's the only real place at the moment.[00:44:48] <Meths> Is that the entire log? Looks like the important bit would be above that. The rest looks like effect not cause.[00:44:57] <yo61> That's all there is[00:45:21] <rmustacc> Well, importantly it's telling us to look at the zone console.[00:45:33] <rmustacc> So in your zone, I would first run passwd for root.[00:45:40] <rmustacc> Set it to something that makes sense.[00:45:47] <rmustacc> Then zlogin -C <zone name>[00:45:56] <rmustacc> It may tell us why we're seeing a failure there now.[00:46:05] *** cjones_ has quit IRC[00:46:27] <yo61> No, I don't see anything[00:46:47] <rmustacc> It just gives you a standard log in prompt?[00:46:57] <yo61> Enter user name for system maintenance (control-d to bypass):[00:47:00] <rmustacc> Oh, yes.[00:47:02] <rmustacc> Please do that.[00:47:10] <yo61> I entered root, and logged in[00:47:13] <rmustacc> That'll help us understand what happened.[00:47:14] <rmustacc> Nothing?[00:47:17] <yo61> No[00:47:24] <yo61> Sep 12 22:46:17 su: 'su root' succeeded for root on /dev/console[00:47:25] <rmustacc> And if you exit that session?[00:47:34] <yo61> Aha[00:48:04] <yo61> https://gist.github.com/robinbowes/6544834[00:48:59] *** tbatchelli_m has quit IRC[00:49:02] <yo61> As Meths says, that looks like effect rather than cause[00:49:21] <yo61> Something is not happening early in the provisioning process and leaving stuff broken[00:49:25] <rmustacc> No, those are very important.[00:49:35] <rmustacc> Because startd starts configd.[00:49:45] <rmustacc> But if configd can't come alone then we will never try and enable EMI.[00:50:11] <rmustacc> What image is that?[00:50:17] <yo61> In /var/svc/log/system-early-manifest-import:default.log is says it can't find /etc/svc/profile/generic.xml[00:50:29] <yo61> It's base64 13.2.0[00:51:08] <yo61> (what's the break key for the console?)[00:51:52] <rmustacc> CTRL-][00:51:55] <rmustacc> At the start of the line.[00:52:35] <yo61> Mmm, not working[00:52:56] <rmustacc> Oh, sorry, missed a line in the manual page.[00:53:06] <rmustacc> To end the serial console session hit CTRL-]. For OS VMs, you'll need to do this at the start of a line, so generally this means pressing: ENTER then CTRL-] then a dot character.[00:53:15] *** tbatchelli_m has joined #smartos[00:55:24] <yo61> I appear to be stuck Am stuck at Enter user name for system maintenance (control-d to bypass):[00:56:34] <yo61> Can always ssh in and kill from another session, I suppose[00:59:53] *** bsdguru has quit IRC[01:03:00] <rmustacc> Well I finally got a chance to grab latest bits and provision base64 of 13.2.0 and I'm not seeing this which is weird.[01:03:20] <rmustacc> I have no generic.xml.[01:03:42] *** FatDarrel has quit IRC[01:03:43] <rmustacc> So that is itself probably a red herring.[01:10:23] <yo61> https://github.com/joyent/smartos-live/issues/261[01:10:50] <yo61> I think I captured most things in there[01:12:57] <yo61> rmustacc: I'm happy to give you access to the system if you think it might help you[01:13:10] *** chorrell has quit IRC[01:18:30] <rmustacc> Stupid question, what happens if you run svcadm enable on the EMI service?[01:19:00] <yo61> In the zone?[01:19:04] <rmustacc> Yeah[01:19:16] <yo61> Let me start it again...[01:20:25] <yo61> I can't see an EMI service?[01:21:24] *** andy_js has quit IRC[01:21:51] <yo61> Should that show up in svcs -a ?[01:22:38] <rmustacc> yo61: It's svc:/system/early-manifest-import:default[01:22:46] <yo61> Ah, sorry[01:22:59] <rmustacc> No worries. I should be clearer.[01:23:13] <yo61> It doesn't appear to do anything[01:23:24] <yo61> svcadm enable svc:/system/early-manifest-import:default[01:24:21] <yo61> Same fatal error in the console[01:24:46] <yo61> svc.configd: Fatal error: Backend copy failed: opening /etc/svc/volatile/fast_repository.db: No such file or directory[01:34:10] *** tbatchelli_m has quit IRC[01:34:25] *** elijah-mbp has quit IRC[01:34:54] <rmustacc> Oh well, worth a shot.[01:37:06] *** Alek has joined #smartos[01:37:51] <yo61> Yeah, I've pretty much run out of shots to give :)[01:39:05] *** tbatchelli_m has joined #smartos[01:39:32] *** elijah-mbp has joined #smartos[01:43:31] *** FatDarrel has joined #smartos[01:44:18] *** deirdres has quit IRC[01:49:26] *** FatDarrel has quit IRC[01:50:15] *** potatosalad has quit IRC[01:51:18] <yo61> <sigh> I rebooted the box and can't ssh to it now.[01:54:14] *** sebasp has quit IRC[01:55:19] *** Daemonik has joined #smartos[01:55:36] *** sebasp has joined #smartos[01:55:37] <Daemonik> I would like to modify /lib/svc/method/ssh for one of my zones. How can I got about doing that?[01:57:09] <richlowe> I think the answer is "You can't, create a new instance of network/ssh, have it run a copy of that script in a place that's editable"[01:57:14] <richlowe> but I'm not sure, so probably wait for someone who is[01:57:24] <rmustacc> That's pretty much the answer.[02:01:11] <richlowe> presumably the zone SMF state persists[02:01:18] <richlowe> so you could probably also go hack the path to the start method with svccfg[02:01:20] <richlowe> without the new instance[02:02:28] <rmustacc> Oh yeah, that works as well.[02:10:57] *** chorrell has joined #smartos[02:19:34] *** wolstena has quit IRC[02:19:42] <yo61> My node is failing on boot with: svs:/system/identity:node failing[02:19:54] <yo61> Any idea what causes that?[02:29:51] *** potatosalad has joined #smartos[02:31:19] *** tbatchelli_m has quit IRC[02:31:45] <yo61> Well, I stopped the scrub that was underway and rebooted and it came back OK[02:31:55] *** tbatchelli_m has joined #smartos[02:33:02] <yo61> Just the "can't create zones" issue then. Time for bed...[02:40:27] *** chorrell has quit IRC[02:42:01] *** potatosalad has quit IRC[02:46:03] *** jaimef has quit IRC[02:51:17] *** jaimef has joined #smartos[02:52:37] *** chorrell has joined #smartos[02:56:40] *** Alek has quit IRC[02:57:08] *** hile_ has quit IRC[02:59:38] *** wolfeidau has quit IRC[03:01:27] *** hile_ has joined #smartos[03:01:57] *** Prudhvi_ has quit IRC[03:02:40] *** Prudhvi has joined #smartos[03:03:20] <fxhp> Hi all[03:03:33] <fxhp> L5520 supports EPT[03:03:46] <fxhp> so KVM should work well correct?[03:07:27] <konobi> you can't nessecarily trust ark[03:07:42] <fxhp> was that directed to me?[03:07:52] *** jaimef has quit IRC[03:09:44] *** nefilim has joined #smartos[03:10:45] <konobi> fxhp: yup... looks like the Gainestown doesn't support EPT[03:11:07] <fxhp> http://ark.intel.com/Products/VirtualizationTechnology[03:11:25] <konobi> http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Gainestown.22_.2845_nm.29[03:12:13] <konobi> fxhp: yes... and ark is generally incorrect with regards to EPT support[03:12:27] <fxhp> I see[03:12:48] <fxhp> bummer[03:13:06] <fxhp> konobi thank you, glad I posted here[03:13:08] <fxhp> : )[03:13:44] <konobi> fxhp: that doesn't nessecarily mean wikipedia is correct... but it seems more reliable =0)[03:14:18] *** jaimef has joined #smartos[03:14:25] <fxhp> yeah, its a thousand dollar purchase, and I really want to mess with KVM[03:15:21] <konobi> check your BIOS to ensure everything is enabled, just in case[03:15:36] <fxhp> I don't own it yet, speculating a purchase[03:16:16] <fxhp> http://www.ebay.com/itm/Dell-Poweredge-C6100-2U-8x-XEON-QC-L5520-2-26GHz-4xNODES-4x250GB-96GB-Ram-Tested-/261223126667[03:16:30] <fxhp> 4 nodes in one rack server[03:18:03] <konobi> fxhp: go for it... you can always replace the cpus[03:18:20] <fxhp> replace 8 CPUs?[03:18:28] <fxhp> that would cost over 1,000[03:18:34] <fxhp> : )[03:20:03] <konobi> local e-waste places are always good to check[03:20:14] <konobi> otherwise there's deepdiscountservers.com[03:25:49] <fxhp> konobi - lol they only have one CPU listed... http://www.deepdiscountservers.com/cpu.html[03:25:54] <fxhp> and guess which one it is[03:28:56] <konobi> if you pick one of the servers you can pick a cpu[03:31:28] *** Transformer has joined #smartos[03:31:50] *** Transformer has quit IRC[03:33:06] *** chorrell has quit IRC[03:33:34] <fxhp> konobi - the upgraded cpu (X5570) is also a "Gainestown" chip[03:33:40] <fxhp> same situation[03:34:50] <fxhp> """All models support:"""[03:35:13] <fxhp> I wonder if that means some in that family support more advanced features[03:38:55] <ashley_w_> this with a L5639 seems to support EPT http://www.deepdiscountservers.com/hp-proliant-dl160-g6-491532-b21.html[03:40:18] <ashley_w_> $570 with 72GB (no disks). tempting.[03:41:37] <konobi> huh... someone's tool reports 5520 as ept capable[03:41:48] <konobi> using host-cpu-info[03:42:04] <richlowe> Remember that ARK lied.[03:42:14] <richlowe> so if they pulled down that data when it was wrong, they probably say that about everything.[03:42:58] <arekinath> you mean the E5520? that's a nehalem chip isn't it?[03:43:48] <arekinath> oh, silly gainestown[03:43:55] *** jaimef has quit IRC[03:44:13] <konobi> which is based on nahelam[03:44:22] <arekinath> but lacks ept[03:45:15] <arekinath> "everything nehalem or later, except gainestown"[03:45:15] <arekinath> :([03:45:40] *** vizant0 has quit IRC[03:45:47] <konobi> someone needs to slap intel re: ark[03:45:54] <fxhp> "everything nehalem or later, except gainestown" ... Truly support EPT?[03:49:48] *** jaimef has joined #smartos[03:52:21] <fxhp> ashley_w_ : c6100 with the L5639 and you would literally have your own cloud[03:52:49] <fxhp> ashley_w_: four servers in one, all running SmartOS[03:53:28] <fxhp> ($1,379 no disks)[03:54:32] <ashley_w_> just switch my fios to a business account, and...[03:58:37] <ashley_w_> my coworkers already ask why i have an 8 core box with 32GB of ram (desktop hardware)[03:59:55] *** trentster has quit IRC[04:04:28] *** trentster has joined #smartos[04:33:52] *** izaki has joined #smartos[04:57:35] *** jellydonut has quit IRC[05:18:34] <ismell> wow 4 boxes in one[05:21:46] *** deirdres has joined #smartos[05:26:56] *** Daemonik has quit IRC[05:36:12] *** jsm has quit IRC[05:48:14] *** FatDarrel has joined #smartos[05:55:52] *** jamesog_ has quit IRC[05:56:14] *** jamesog has joined #smartos[06:15:15] *** adhi_outlined has joined #smartos[06:16:14] <adhi_outlined> hi, is anyone knows how to identify physical disk on SmartOS using LED blinking like in nexenta ?[06:16:36] <rmustacc> That depends entirely on your hardware.[06:17:02] <adhi_outlined> make some search in google and found "luxadmin" but didn't find in pkgin search[06:17:16] <rmustacc> Do you have a ses device?[06:17:32] <adhi_outlined> did you mean SAS drive ?[06:17:48] <rmustacc> No, I do not.[06:18:02] <rmustacc> The question really comes down to what hardware do you have in your machine.[06:18:13] <rmustacc> And whether your HBA has a SES device or not.[06:18:35] <rmustacc> If it does, it is a lot easier, if it does not, you have more work to do to get the mappings.[06:18:46] <rmustacc> But LeftWing in here knows that best, but I don't think he's around at the moment.[06:19:16] *** ktkNA has quit IRC[06:19:18] <adhi_outlined> how about "luxadm" ?[06:19:23] *** ktkNA has joined #smartos[06:19:45] *** whyz has quit IRC[06:19:46] *** Jadelrab has quit IRC[06:19:57] *** arekinath has quit IRC[06:20:00] *** arekinath_ has joined #smartos[06:20:05] <rmustacc> All I can really tell you is that it is certainly possible, but it is not currently as easy unless you have certain hardware configurations.[06:20:33] *** arekinath_ is now known as arekinath[06:20:40] *** arekinath has quit IRC[06:20:40] *** arekinath has joined #smartos[06:20:42] <rmustacc> adhi_outlined: Probably can get a better answer by e-mailing smartos-discuss.[06:20:50] <rmustacc> But it won't be immediate.[06:21:29] *** chi has quit IRC[06:21:37] *** chi has joined #smartos[06:22:12] <adhi_outlined> great, thanks for your suggest :D[06:22:17] *** whyz has joined #smartos[06:22:56] <rmustacc> But you should really see if you have a ses device with IIRC fmtopo[06:23:11] *** jihyun has joined #smartos[06:23:18] <rmustacc> Because the one thing I can tell you is that it really depends on your hardware and what information is already available.[06:33:28] *** theCzar has joined #smartos[06:35:14] <adhi_outlined> thanks rmustacc[06:37:12] <rmustacc> You want something like /usr/lib/fm/fmd/fmtopo[06:37:14] <theCzar> So... I'm trying to start up SmartOS on an old AMD phenom II machine that I have. I got it to setup once. But then I had to move the machine. When I set it up again, and booted from the flash-drive, it got to the first boot message (welcome to sunos etc...) and just hung there for ever. I'm not totally sure what to do to trouble shoot. Has anyone else experienced something like this, or know what I can try next?[06:42:56] *** andoriyu has joined #smartos[06:48:31] <theCzar> I just tried booting with the -v flag[06:49:01] <theCzar> it appears to have hung after spewing out a bunch of x86_features[06:49:42] <theCzar> now it's just showing a blinking cursor after 'mem = 4192636K (0xffe5f000)'[06:53:26] *** nikolam has joined #smartos[06:54:05] *** andoriyu1 has joined #smartos[06:55:14] <rmustacc> theCzar: Sounds like you hit exactly this: http://blog.delphix.com/mba/2013/my-dell-t620/[06:57:19] *** andoriyu has quit IRC[06:59:31] *** szaydel has quit IRC[07:01:19] <theCzar> rmustacc: hmm, it does seem similar[07:01:34] <theCzar> granted I haven't tried everything he has yet[07:01:49] <theCzar> but I was able to boot it and set it up once[07:02:06] <theCzar> which makes me really wonder...[07:02:07] *** vaskas has joined #smartos[07:02:40] <rmustacc> Okay, then there is a possibility that it is something different.[07:02:59] *** nefilim has quit IRC[07:04:37] <theCzar> rmustacc: as far as I can tell the only change that has happened since is that I moved it, and plugged in an ethernet cable (which I did not have access to in its previous location)[07:05:20] <theCzar> and changed monitors. obviously, none of that should matter[07:06:16] <theCzar> Currently I'm rebuilding the USB key and wiping the hard drive from another computer[07:06:40] <theCzar> on the off chance it will make it try to redo the setup[07:07:50] *** julianduque has quit IRC[07:19:02] *** andoriyu1 has quit IRC[07:20:03] *** chimera77 has joined #smartos[07:23:54] *** andoriyu has joined #smartos[07:26:05] *** wolfeidau has joined #smartos[07:26:46] *** wolfeidau has quit IRC[07:26:54] *** wolfeidau has joined #smartos[07:29:38] *** vaskas has quit IRC[07:30:18] <theCzar> nope... no luck[07:30:22] <theCzar> man... what the[07:30:43] *** vaskas has joined #smartos[07:31:03] <rmustacc> Do you have ipmi on the machine?[07:31:17] *** FatDarrel has quit IRC[07:35:12] <MerlinDMC> morning[07:36:51] <theCzar> rmustacc: I don't think I do[07:37:36] *** potatosalad has joined #smartos[07:42:42] *** vaskas has quit IRC[07:44:24] *** vaskas has joined #smartos[07:50:59] *** vaskas has quit IRC[07:52:00] *** vaskas has joined #smartos[07:59:20] *** deirdres has quit IRC[08:01:56] *** tbatchelli_m has quit IRC[08:02:35] *** FatDarrel has joined #smartos[08:16:01] *** potatosalad has quit IRC[08:17:24] *** Jadelrab has joined #smartos[08:20:00] *** xinkeT has quit IRC[08:20:55] *** xinkeT has joined #smartos[08:21:31] *** mamash has joined #smartos[08:21:38] *** deirdres has joined #smartos[08:28:31] *** FatDarrel has quit IRC[08:31:01] *** deirdres has quit IRC[08:48:11] *** nikolam has quit IRC[08:56:02] *** alcir has joined #smartos[09:01:42] *** trentster has quit IRC[09:06:22] *** trentster has joined #smartos[09:07:12] *** vaskas has quit IRC[09:07:43] *** theCzar has quit IRC[09:13:42] *** bixu has quit IRC[09:14:18] *** bixu has joined #smartos[09:14:32] *** bens1 has joined #smartos[09:20:01] *** texarcana has quit IRC[09:22:02] *** texarcana has joined #smartos[09:23:54] *** tonist has joined #smartos[09:31:26] *** izaki has quit IRC[09:34:02] <alcir> someone using splunk?[09:44:30] *** wiedi has quit IRC[09:48:03] *** KermitTheFragger has joined #smartos[09:51:55] *** ajacoutot has joined #smartos[09:52:00] <ajacoutot> hi[09:57:49] *** izaki has joined #smartos[10:04:01] *** andoriyu has quit IRC[10:06:36] *** izaki has quit IRC[10:26:35] *** estibi_mobile has joined #smartos[10:27:32] *** estibi_mobile is now known as estibi[10:28:09] *** vizant0 has joined #smartos[10:28:19] *** nikolam has joined #smartos[10:31:19] <simmel> wtf, `route add -host 10.1.1.3 -interface e1000g1` should add a host route to 10.1.1.3 on interface e1000g1 right? How come it decides to add the route with a gateway the same as my default route and on the e1000g0 (which is where the default route is)!? *cries*[10:32:58] *** ktkNA is now known as ktk[10:36:02] *** nikolam has quit IRC[10:39:16] *** vizant0 has quit IRC[10:41:58] *** vizant0 has joined #smartos[10:46:44] *** vizant0 has quit IRC[10:47:04] *** W0rmDr1nk has joined #smartos[10:51:07] *** nikolam has joined #smartos[11:00:20] *** estibi has quit IRC[11:03:00] *** izaki has joined #smartos[11:07:54] *** izaki has quit IRC[11:16:10] *** andy_js has joined #smartos[11:17:40] *** vizant0 has joined #smartos[11:24:14] *** Meths_ has joined #smartos[11:24:53] *** Meths has quit IRC[11:50:51] *** tonist has quit IRC[11:51:55] *** yruss972 has joined #smartos[11:51:56] *** tonist has joined #smartos[11:56:44] *** adhi_outlined has left #smartos[11:59:21] *** vaskas has joined #smartos[12:03:25] *** izaki has joined #smartos[12:05:42] *** mattconn_ has joined #smartos[12:08:07] *** izaki has quit IRC[12:08:20] *** mattconnolly has quit IRC[12:09:47] *** indutny has quit IRC[12:11:19] *** tonist has quit IRC[12:13:04] *** yruss972 has quit IRC[12:13:40] *** vaskas has quit IRC[12:37:54] *** jim80net has quit IRC[13:01:19] *** yruss972 has joined #smartos[13:01:43] *** trentster has quit IRC[13:06:02] *** KermitTheFragger has quit IRC[13:06:34] *** trentster has joined #smartos[13:15:09] *** jellydonut has joined #smartos[13:17:07] *** chakatz has joined #smartos[13:25:51] *** joffe has joined #smartos[13:26:41] *** jellydonut has quit IRC[13:29:34] *** hadret has joined #smartos[13:32:21] *** szaydel has joined #smartos[13:38:18] *** tonist has joined #smartos[13:40:24] *** miksago has joined #smartos[13:41:21] <miksago> hey, should I need to compile libz, libssl, and libcrypto manually to install node.js on smartos base64?[13:43:21] <jperkin> why not use the existing packages?[13:46:14] <miksago> I'm getting errors like:[13:46:14] <miksago> ld: fatal: file /opt/local/lib/libz.so: wrong ELF class: ELFCLASS64[13:46:41] <miksago> (I'm needing to compile node, as it's a custom branch)[13:47:24] <jperkin> ok, so make sure you build node with -m64 in CFLAGS[13:47:47] <jperkin> by default gcc will build 32bit objects, which you can't then link against the 64bit libraries provided by base64[13:47:56] <miksago> ah[13:48:07] <miksago> I did: ./configure-gyp --with-dtrace --gdb --dest-cpu=x64[13:49:37] <jperkin> ok, sounds like a node bug if --dest-cpu isn't adding the appropriate -m32/-m64[13:50:29] <miksago> hmm, okay[13:51:16] <jperkin> as we build node from pkgsrc the infrastructure handling will mask bugs like that[13:51:54] <simmel> Modify the pkgsrc "package" to build your custom branch?[13:52:17] <mamash> jperkin: on base64 gcc should default to -m64 though?[13:53:13] <miksago> looks like for some reason in the generated options.gypi file, it set host_arch to ia32[13:53:15] *** nikolam has quit IRC[13:59:31] *** hadret has quit IRC[14:02:18] *** ira has joined #smartos[14:02:25] <jesse_> jperkin, what's 'standard' image?[14:02:36] <jperkin> base plus a bunch of packages[14:02:50] <jesse_> and still no multiarch?=([14:03:13] <jesse_> "java" is base plus openjdk?[14:03:15] <jperkin> not in JPC[14:03:41] <jesse_> is it going to be part of the jpc set at some point?[14:04:01] <jperkin> I don't know exactly what all the "base plus" images are composed of, but it's mostly just 'pkgin in foo' and possibly some config tweaks.[14:04:52] <jperkin> at some point yes, once I've finished converting a bunch of the more popular packages and it's had some reasonable testing[14:05:02] <jperkin> sngl is multiarch'd so I'm hoping for some feedback via that[14:05:29] <jperkin> however the appetite for multiarch seems very low, you are the only person I know of that uses it.[14:06:36] <jesse_> how does the jdk compile at all, then? it needs multiarch for 64bit hotspot[14:06:47] <jesse_> or did you do something to it?[14:08:37] <MerlinDMC> jesse_, take a look at the mibe config for the java image and you'll see what they change ;) https://github.com/joyent/mi-java[14:09:33] <jperkin> it uses sun-jdk7 for bootstrap[14:09:55] <jesse_> it still shouldn't compile[14:10:13] <jesse_> unless you just compile 32bit jvm[14:10:20] <jesse_> it should work ok without multiarch[14:11:28] <jesse_> (and that mibe says it uses sun-jdk6)[14:12:30] <jesse_> MerlinDMC, I didn't know there were repos like that for the images[14:12:41] *** yruss972 has joined #smartos[14:12:50] <MerlinDMC> jesse_, they did release them over time after getting mibe to the public[14:12:51] <jesse_> "just what I wanted"[14:16:32] *** nikolam has joined #smartos[14:17:07] *** yruss972 has quit IRC[14:17:43] *** hadret has joined #smartos[14:23:22] *** Teknix has joined #smartos[14:23:51] *** xinkeT has quit IRC[14:25:26] *** nikolam has quit IRC[14:37:47] *** izaki has joined #smartos[14:39:04] *** jim80net has joined #smartos[14:41:52] <jesse_> jperkin, is there a newer multiarch image than the 12.4.1?[14:42:16] <jesse_> maybe one with trunk pkgsrc?[14:42:35] <jperkin> yes, there is 13.2.2 in http://pkgsrc.smartos.org/datasets/[14:43:54] <jesse_> I didn't have that source, explains a lot...[14:44:24] <jesse_> imgadm: error (SourcePing): unexpected ping error with image source "http://pkgsrc.smartos.org/datasets/" (dsapi): Error: statusCode 404[14:44:30] <jesse_> and that must explain why not=)[14:44:52] <jperkin> it's just nginx, not an imgadm server[14:48:34] *** chakatz has quit IRC[14:54:37] *** chakatz has joined #smartos[14:59:41] *** marsell has joined #smartos[15:00:50] *** mattconn_ has quit IRC[15:01:01] *** neophenix has joined #smartos[15:01:03] *** nikolam has joined #smartos[15:01:59] *** chorrell has joined #smartos[15:08:42] *** chorrell has quit IRC[15:10:47] *** neophenix has quit IRC[15:31:14] *** potatosalad has joined #smartos[15:45:56] *** theCzar has joined #smartos[15:48:47] *** jim80net has quit IRC[15:54:26] *** jsm has joined #smartos[15:59:50] *** potatosalad has quit IRC[16:01:01] *** chorrell has joined #smartos[16:05:11] *** neophenix has joined #smartos[16:07:51] *** ilovezfs has quit IRC[16:09:52] *** jim80net has joined #smartos[16:14:49] *** nikolam has quit IRC[16:15:37] *** nefilim has joined #smartos[16:16:39] *** nikolam has joined #smartos[16:17:02] *** rodgort has quit IRC[16:18:56] *** rodgort has joined #smartos[16:26:47] *** tonist has quit IRC[16:35:29] *** volci has quit IRC[16:36:21] *** neophenix has quit IRC[16:42:40] *** deirdres has joined #smartos[16:43:24] *** jaimef has quit IRC[16:45:09] *** Trixboxer has joined #smartos[16:45:19] <Trixboxer> hi, do we have hdparm package for smartos ?[16:45:37] *** daleg has quit IRC[16:46:01] *** daleg has joined #smartos[16:48:36] <jperkin> no, it is Linux-only[16:49:53] *** jaimef has joined #smartos[16:50:26] *** potatosalad has joined #smartos[16:51:04] *** szaydel has quit IRC[16:51:14] *** mamash has left #smartos[16:51:28] *** julianduque has joined #smartos[16:51:30] <fxhp> ashley_w_ - I want FIOS so badly. Only Uverse in my area.[16:52:50] *** mary5030 has joined #smartos[16:53:55] *** mary5030 has quit IRC[16:54:31] *** mary5030 has joined #smartos[16:54:46] *** mary5030 has quit IRC[16:55:13] <Trixboxer> jperkin: is it not possible to compile it for smartos ?[16:55:18] *** mary5030 has joined #smartos[16:57:42] *** tbatchelli_m has joined #smartos[16:58:03] <nahamu> Trixboxer: if it depends on linux-only system calls, it won't work.[16:58:31] *** szaydel has joined #smartos[16:59:45] <Trixboxer> ok[17:00:15] *** chorrell has quit IRC[17:01:01] <miksago> where would pkgin install the svcs config for rabbitmq to?[17:02:32] <ira> svcadm export I believe will get it...[17:02:43] <ira> (It'll get you the XML of the file, not the location.)[17:02:54] <simmel> `route add -host 10.1.1.3 -interface e1000g1` should add a host route to 10.1.1.3 on interface e1000g1 right? How come it decides to add the route with a gateway the same as my default route and on the e1000g0 (which is where the default route is)!? *cries*[17:03:21] *** roast_ has quit IRC[17:03:21] *** koaps has quit IRC[17:04:15] *** koaps has joined #smartos[17:06:34] <miksago> hmm, okay[17:07:30] <miksago> svcadm doesn't have an export, but for some reason rabbitmq isn't starting when run via svcadm[17:07:59] <yo61> So, I pulled all the disks from my smartOS node, put in a single empty 500GB disk, reinstalled/rebooted[17:08:12] <ira> try svccfg export[17:08:20] <yo61> I then imported base64 13.2.0[17:08:22] <yo61> imgadm import 0084dad6-05c1-11e3-9476-8f8320925eea[17:08:50] <yo61> And created a zone[17:08:53] <yo61> It worked[17:09:41] <rmustacc> miksago: I think ira meant svccfg export[17:09:56] <ira> rmustacc: Indeed.[17:10:07] <rmustacc> D'oh, missed that a few lines up.[17:10:11] <miksago> rmustacc: yeah, I also found /opt/local/share/smf/rabbitmq/manifest.xml[17:10:19] <rmustacc> miksago: But generally you should already have it imported.[17:10:27] <miksago> but I can't see anything astray in there[17:10:30] <rmustacc> And yeah, I was about to say pkgsrc drops it in /opt/local/share/smf[17:10:32] <ira> yeah, but use the svc comands to modify that stuff for your own sanity ;)[17:10:35] <wesolows> those who were asking about hdparm, what do you actually want to do?[17:10:46] *** tbatchelli_ has joined #smartos[17:10:50] *** tbatchelli_m has quit IRC[17:10:53] <rmustacc> miksago: What is it you're trying to do or what's going wrong?[17:10:55] <nahamu> Trixboxer: wesolows wants to know what you wanted to do with hdparm.[17:11:17] <miksago> I'm trying to run rabbitmq, I installed with `pkgin in rabbitmq`, then `svcadm enable rabbitmq:default`[17:11:45] <Trixboxer> nahamu: thanks for noticing wesolows[17:11:48] <miksago> but then doing `svcs rabbitmq` and I get: offline 2:55:29 svc:/pkgsrc/rabbitmq:default[17:12:01] <rmustacc> miksago: You should run 'svcs -xv rabbitmq'[17:12:01] <Trixboxer> Im trying to use this most easy disk benchmark[17:12:25] *** roast_ has joined #smartos[17:12:28] <wesolows> just use dd, then. hdparm's "benchmark" is at least as useless as dd.[17:12:44] <miksago> https://gist.github.com/miksago/c93008bd2949cbc511b9 < result of that[17:12:52] <wesolows> as in time dd if=/dev/dsk/whatever of=/dev/null bs=whatever count=whatever[17:12:59] <miksago> I guess I'd need to start epmd[17:13:07] <rmustacc> Yup.[17:13:41] <Trixboxer> wesolows: yeah dd is what I can use[17:13:52] <Trixboxer> but it does not give speed info like linux[17:14:10] <Trixboxer> only the number of records[17:14:22] <richlowe> and time[17:14:43] <Trixboxer> yeah so use calculator :)[17:14:53] <Trixboxer> why not give it right there :X[17:15:44] <miksago> rmustacc: looks like I had a dead epmd running, which meant svcs couldn't start a new one[17:15:48] <miksago> or something like that[17:16:01] <rmustacc> What did svcs -xv say for epmd?[17:16:16] <wesolows> you can use GNU dd if you like[17:16:36] <wesolows> the timing stuff is a GNU feature, not a Linux one. GNU dd is available in pkgsrc and works fine.[17:18:13] <Trixboxer> umm may I know the exact package name please ?[17:18:14] <wesolows> when you said hdparm I feared I'd be groveling through sd's inventory of ioctls for some obscure flag setting function[17:18:52] <Trixboxer> pkgin search dd not much helpful[17:19:17] <yo61> rmustacc: I've updated https://github.com/joyent/smartos-live/issues/261[17:19:31] <wesolows> I think it's coreutils[17:19:55] <wesolows> It's very likely already in your zone.[17:23:08] <Trixboxer> ok it was not default installed[17:23:29] <rmustacc> It is in a zone, not if you're using a global zone set of pacakges.[17:25:54] <miksago> rmustacc: said it wasn't running, I disabled that service, and re-enabled it, that worked[17:26:10] <miksago> then had to re-install rabbitmq to wipe bad directory permissions[17:29:52] *** chorrell has joined #smartos[17:30:19] * nahamu points out that to acess the raw disks one needs to be in the GZ where pkgsrc isn't installed by default.[17:30:21] *** chakatz has quit IRC[17:31:32] *** tbatchelli_ has quit IRC[17:32:02] *** tbatchelli_m has joined #smartos[17:32:49] *** szaydel has quit IRC[17:32:55] *** trentster has quit IRC[17:34:54] *** alcir has quit IRC[17:37:25] *** trentster has joined #smartos[17:40:23] *** julianduque has quit IRC[17:43:27] *** chakatz has joined #smartos[17:44:30] *** chakatz has quit IRC[17:45:05] *** chakatz has joined #smartos[17:46:50] *** paul_lamb has joined #smartos[17:50:15] *** FatDarrel has joined #smartos[17:50:39] *** chakatz has quit IRC[18:00:05] <yo61> I've updated my ticket again: https://github.com/joyent/smartos-live/issues/261[18:00:35] <yo61> It seems to be some problem with the persistent configuration stored with my zpool[18:00:46] <yo61> I have no idea what to check though. Any ideas?[18:01:45] <ktk> am I right that ipfilter on solaris is basically the same like on freebsd/netbsd syntaxwise?[18:03:45] <wesolows> probably? You'd have to ask Oracle what changes they've made in Solaris. We wouldn't know.[18:03:52] <wesolows> Certainly it was in the past.[18:04:24] <ktk> ok[18:04:31] <ktk> then I know which manpages to consult ;)[18:06:55] <richlowe> if I recall, it's the same but probably older[18:07:06] <richlowe> darren dropped a _massive_ webrev on us (illumos), and we ignored it[18:07:25] <richlowe> because it's huge, and full of ipfilter code.[18:07:36] *** FatDarrel has quit IRC[18:07:57] <ktk> what do you mean with webrev?[18:08:20] <ktk> great the tutorials linked in the FAQ are 404 :)[18:08:59] *** andoriyu has joined #smartos[18:10:08] <wesolows> A lot of Oracle's docs are 404, or links to useless index pages, because apparently it was more important to them to turn down sun.com as a valid domain name than for their customers to be able to access useful information.[18:10:15] <wesolows> Take that one up with your sales rep. :)[18:11:03] <ktk> yeah bunch of dumbasses[18:19:05] *** julianduque has joined #smartos[18:24:29] *** nikolam has quit IRC[18:30:54] *** mamash has joined #smartos[18:33:16] <jesse_> jperkin, gcc48 doesn't have /opt/local/lib in lib path?[18:33:25] <jesse_> did gcc47 have that either[18:33:37] <jesse_> or did I break something else?[18:34:52] *** Alek has joined #smartos[18:45:03] <rmustacc> jesse_: That should be fixed in 2013Q2.[18:46:10] *** vertigo has left #smartos[18:47:50] <jperkin> it's certainly fixed for gcc47 in 2013Q2, I might not have forwarded the patches to gcc48 though[18:48:06] *** izaki has quit IRC[18:54:38] *** chorrell has quit IRC[18:58:43] *** mamash has left #smartos[18:59:19] <rmustacc> We should make sure to get that forward ported.[19:00:26] <jperkin> yeh definitely.[19:03:17] *** chimera77 has quit IRC[19:08:50] *** bens1 has quit IRC[19:14:31] <rmustacc> yo61: Interesting.[19:14:58] <rmustacc> One thing I would check is that if you zfs send your local version of the dataset to a file and checksum it that it appears correctly.[19:15:38] *** wolstena has joined #smartos[19:23:53] <yo61> OK, will try that[19:24:33] <rmustacc> See if it matches the checksum in the manifest.[19:24:39] <jesse_> jperkin, so, for once, it's not me :P[19:30:55] <yo61> Like this: zfs send zones/0084dad6-05c1-11e3-9476-8f8320925eea > base64-13.2.0.zfs[19:30:58] <yo61> ?[19:31:02] <jperkin> well, it kinda is, just use gcc47 ;P[19:31:13] <yo61> Or is the checksum on the gziped file?[19:32:36] <yo61> Hmm, need to send a snapshot[19:32:46] <rmustacc> There should be one already.[19:32:51] <rmustacc> The @final snapshot is what we imported.[19:33:00] <yo61> OK, using that[19:33:07] <yo61> zfs send zones/0084dad6-05c1-11e3-9476-8f8320925eea@final | gzip > base64.13.2.0.gz[19:34:24] <jperkin> well, that's going to depend on the compression level used[19:34:32] <yo61> Yeah, I wondered that[19:35:35] <rmustacc> Sometimes the manifests have the checksum for the uncompressed version.[19:36:22] <yo61> zfs show says:[19:36:25] <yo61> {[19:36:25] <yo61> "sha1": "53b2f11e3bea1e5fd9052a5ef6f2874ccc88e355",[19:36:25] <yo61> "size": 159268611,[19:36:25] <yo61> "compression": "gzip"[19:36:26] <yo61> }[19:37:31] <yo61> The gzip'd file is[19:37:31] <yo61> If anyone has the Devil and Demons in them , it is these people. Making a living out of bullshit , and Buffy the vampire slayer !! Capitalists to the extreme . As as for our country, we have traditions, and a culture going back thousands of years.[19:37:35] <yo61> Along with foBah[19:37:38] <yo61> wtf?[19:37:55] <yo61> The gzip'd file is 158759130[19:40:35] <yo61> Am I doing something wrong ?[19:45:49] *** tonist has joined #smartos[19:46:21] *** W0rmDr1nk has quit IRC[19:46:59] *** estibi_mobile has joined #smartos[19:48:21] *** estibi_mobile is now known as estibi[19:48:25] *** mamash has joined #smartos[19:57:45] *** Meths_ is now known as Meths[20:00:15] *** estibi has left #smartos[20:00:48] <rmustacc> yo61: Sorry I've been trying to take care of a few other things.[20:00:58] <rmustacc> You don't want the zfs show information.[20:01:40] *** deirdres has quit IRC[20:01:49] *** chorrell has joined #smartos[20:08:39] *** wolstena has quit IRC[20:08:41] *** wolstena1 has joined #smartos[20:22:49] *** bixu_ has joined #smartos[20:22:50] *** bixu has quit IRC[20:26:42] *** mamash has left #smartos[20:29:56] *** mamash has joined #smartos[20:33:08] <yo61> Sorry - back now[20:33:11] <yo61> What do I need?[20:35:28] *** FrankD_Work has quit IRC[20:35:48] *** vizant0 has quit IRC[20:35:57] <rmustacc> Probably first thing is to look at the imgadm show output.[20:36:49] <yo61> Sorry, that's what I meant before - that was the imgadm show output[20:38:12] <rmustacc> Oh okay.[20:38:23] <rmustacc> So it only has the gzip compressed bit.[20:38:53] <rmustacc> I guess stupid thing to rule out, imgadm delete the image and then import it again.[20:38:53] <yo61> Yes[20:39:00] <yo61> OK[20:39:05] <yo61> Have done that, but will do again.[20:39:09] <rmustacc> Ah, you already did that.[20:39:12] <rmustacc> Okay, then that rules that out.[20:39:24] <yo61> Importing again[20:39:34] <rmustacc> Probably not worth it.[20:39:37] <yo61> Will take a few mins[20:39:49] <rmustacc> Just trying to think through where differences would be.[20:40:00] <rmustacc> The differences in /usbkey/config shouldn't do anything.[20:40:20] <yo61> No, I sync'd those and rebooted. They don't make any difference[20:40:54] <rmustacc> And it's the same version of the platform in both cases?[20:41:11] <yo61> It's the same USB key[20:41:21] <yo61> The only difference is the different zpool[20:41:25] <rmustacc> 'Hmm.[20:43:12] *** ryancnelson has joined #smartos[20:44:01] *** hadret has quit IRC[20:44:57] *** nikolam has joined #smartos[20:49:04] *** deirdres has joined #smartos[20:49:27] *** nefilim has quit IRC[20:49:44] *** nefilim has joined #smartos[20:49:58] *** ryancnelson has quit IRC[20:54:55] *** miksago has quit IRC[20:55:53] <yo61> The import has completed[20:56:00] *** Alek has quit IRC[20:56:08] <yo61> I shall try creating a zone again[20:56:50] <yo61> No change[20:59:40] *** nikolam has quit IRC[21:02:32] <yo61> There appear to be a lot of services disabled on my global zone, and ntp is in maintenance[21:05:42] *** patrickmslattery has joined #smartos[21:08:38] *** vizant0 has joined #smartos[21:11:09] <jesse_> do other 13.2 images have /etc/release, or is it just the multiarch missing it?[21:11:29] <yo61> I have to pop out briefly...[21:12:23] <jperkin> jesse_: just multiarch, I fixed it in the most recent sngl but I didn't create an updated multiarch[21:12:45] <jesse_> ok, the build was using it to figure out which SunOS it is[21:13:09] <jperkin> it's not the most stable of interfaces.. ;)[21:13:16] <jesse_> it's the only[21:13:23] <jperkin> what's wrong with uname -v ?[21:13:51] <jesse_> I have no idea what the problem with that was[21:14:02] *** chorrell has quit IRC[21:15:23] *** nikolam has joined #smartos[21:15:45] *** bixu_ is now known as bixu[21:16:16] *** chorrell has joined #smartos[21:20:25] *** patrickmslattery has quit IRC[21:20:39] <jesse_> this "quick 15 minute update" is going to take all night[21:20:49] * jesse_ opens a Magners[21:30:24] <jperkin> opus magners[21:46:26] *** vizant0 has quit IRC[21:49:31] *** vizant0 has joined #smartos[21:55:10] *** mamash has left #smartos[22:01:53] <yo61> Friels is nicer[22:03:35] *** wolstena1 has quit IRC[22:05:35] <bixu> NATURAL ICE[22:05:39] *** deirdres has quit IRC[22:06:01] <paul_lamb> i'd be willing to pay to have a list of packages updated for next release, is there any such thing?[22:06:21] *** deirdres has joined #smartos[22:06:47] <jperkin> you could try offering bounties on pkgsrc-users at netbsd dot org[22:07:22] <jperkin> I'm sure there are a few devs who'd take you up on any offers[22:08:43] <paul_lamb> thanks, getting postgresql 9.3 and postgis 2.1 working in a zone hasn't been very fun[22:08:54] <jperkin> postgresql 9.3 is already done[22:09:22] <paul_lamb> but postgis 2.1 depends on gdal 1.10 which has about a hundred dependencies[22:09:44] <bixu> paul_lamb: Also, if you use Chef, we have https://github.com/wanelo-chef/postgres if you need to build from source.[22:09:46] <yo61> http://www.gwyntcidershop.com/product/dog_dancer/[22:10:54] <bixu> I've also started work on a cookbook to help automate building one's own pkgsrc packages: https://github.com/bixu/pkg_create[22:11:17] *** marsell has quit IRC[22:11:30] <paul_lamb> that would be cool[22:11:56] <bixu> paul_lamb: I've been packging rbenv-style rubies with it.[22:12:34] <paul_lamb> i've been using ansible but chef isn't out of the question[22:13:04] <bixu> Cool. I keep meaning to look at Ansible. How do you like it?[22:13:11] <paul_lamb> trying to convert a few linux and an old openidiana server over to smartos[22:13:30] <paul_lamb> just the fact that you don't have to install anything on the server makes me like ansible[22:13:34] <bixu> I do think that we should be open to more than one configuration management tool.[22:13:42] <bixu> Agreed.[22:13:52] <bixu> Been want to look at OpsMezzo as well.[22:14:10] <paul_lamb> I have lots of ruby code but just getting ruby working can be painful[22:14:27] <bixu> Agreed.[22:15:05] <paul_lamb> ok if pg9.3 has already been done in the repo is there a way to use it now?[22:15:27] <jperkin> if you grab pkgsrc via cvs then you can build it from there[22:15:40] <jperkin> I haven't synced to the git repo for a little while so I don't think it's there yet.[22:16:22] <paul_lamb> i didn't see it in the github repo earlier today, looked in trunk[22:16:39] <bixu> I've also published some stuff for getting the omnibus (isolated) chef-client onto your smartos zone: https://github.com/bixu/chef-bootstrap[22:17:16] <paul_lamb> nice, i will try that out[22:18:33] <bixu> Pull requests are most welcome.[22:18:43] <paul_lamb> and last question for today, are there any magical tools besides rsync to duplicate a zfs filesystem with joyent's storage?[22:18:56] <bixu> And a big thanks to jperkin for really good docs for my pkgsrc endeavors.[22:19:39] <bixu> paul_lamb: If you submit a feature request for zfs dataset delegation to the zone, I'll upvote it :)[22:20:03] *** Alek has joined #smartos[22:20:43] <rmustacc> You can delegate a dataset for management today with smartos.[22:21:22] <rmustacc> 'delegate_dataset' in vmadm[22:22:03] <jesse_> is it still create-only?[22:22:15] <jesse_> I use zonecfg to add/remove datasets from zones[22:23:26] <paul_lamb> i found the zonecfg thing too, that's how I mounted my tank pool to a zone[22:23:38] <rmustacc> It is not currently updateable per the man page.[22:24:10] <jesse_> I think it has a ticket in smartos-live already[22:24:13] *** mamash has joined #smartos[22:24:27] <jesse_> I would have added it a year ago if there weren't one=)[22:26:24] <jperkin> paul_lamb: gdal seems reasonably straight-forward, it's likely that simply updating the existing pkgsrc package shouldn't be too much work[22:27:55] <paul_lamb> because of elfclass64 errors, i've so far had to recompile 18 different libs[22:28:20] <jperkin> yeh, pkgsrc will fixup all that stuff for you[22:28:44] *** mattconnolly has joined #smartos[22:29:01] <paul_lamb> i gave up, created a centos zone, added the gis rpm to it and installed[22:32:22] *** tonist has quit IRC[22:32:47] <rmustacc> paul_lamb: Where you trying to build a 32-bit binary in a 64-bit package set or vice versa?[22:33:04] * yo61 is back[22:33:23] *** tonist has joined #smartos[22:33:53] <paul_lamb> i used a 64 bit zone, i'm about to head out, i'll reproduce the error on monday[22:34:34] <rmustacc> Well if you hit errors compiling stuff, do ask in here, we're happy to help as we can.[22:34:51] *** bens1 has joined #smartos[22:34:59] *** tonist has quit IRC[22:35:41] *** bens1 has quit IRC[22:37:18] *** tonist has joined #smartos[22:38:23] *** deirdres has quit IRC[22:38:39] *** deirdres has joined #smartos[22:41:05] *** vizant0 has quit IRC[22:45:22] <yo61> So, is there anything further I can do to debug this machine I'm having problems with?[22:46:52] *** paul_lamb has quit IRC[22:46:54] <jesse_> just out of curiosity, is that non-working zones-pool old or new?[22:47:49] <yo61> Define "old" and "new"[22:48:56] <jesse_> created with older smartos than you're booting it with now[22:49:08] <yo61> Yes, it's old[22:51:24] <jesse_> did you upgrade your zpool after upgrading smartos?[22:51:54] <yo61> I believe so[22:52:28] *** ira has quit IRC[22:52:28] <yo61> I just checked, and I get: Pool 'zones' already has all supported features enabled.[22:53:07] <jesse_> that was a long shot, anyway, it should fail/complain when importing the image if that would matter[22:53:30] <yo61> It did originally have a different name[22:53:38] <jesse_> shouldn't matter[22:53:47] <jesse_> I have several pools I've renamed into zones[22:54:07] <yo61> zpool history is good, isn't it?[22:54:37] <yo61> 2011-10-15.18:19:56 zpool create -O casesensitivity=mixed -O compress=gzip-1 -O dedup=on -O nbmand=on space01 raidz2 da1p1 da3p1 da5p1 da7p1[22:54:40] <yo61> da9p1 da11p1 da13p1 da15p1 log mirror gpt/slog0 gpt/slog1 cache gpt/cache0 gpt/cache1[22:55:06] *** ira has joined #smartos[22:55:10] <jesse_> *gasp* dedup[22:55:19] <yo61> Heh, yeah[22:55:20] <jesse_> *gasp* gzip-1 not lz4[22:55:51] <yo61> dedup because of how I use the space (saves me loads of storage)[22:56:16] <jesse_> you're in the 0.00000001% gets something from dedup?=)[22:56:22] <jesse_> +that[22:56:22] <yo61> gzip-1 because errrr....[22:56:32] <jesse_> (lz4 is new)[22:56:45] <yo61> Yeah, was going to say lz4 was not around[22:57:06] <yo61> I have a lot of audio files in flac format, that I copy to two places[22:57:22] <yo61> (with only small changes between the two locations)[22:57:38] <yo61> So, I effectively get double the storage by using dedup[22:58:07] *** Alek has quit IRC[22:58:35] <yo61> Is it a global thing? Or can I turn it off and just enable it for specific datasets?[22:58:51] <jesse_> casesensitivity=mixed sounds a bit bad for a zones pool[22:59:06] <jesse_> not sure why it would be a problem, but...[22:59:07] <yo61> Mmmm[22:59:17] *** wolstena has joined #smartos[22:59:31] <yo61> I think I did that because of sharing with Windows clients[22:59:44] <jesse_> dedup had some funny things with it, if you disable it it still lingers[22:59:53] <yo61> (not that I do that currently, but I did when I had all this stuff on OpenSolaris)[23:00:36] <yo61> I need borrow a few large drives so I can backup and re-create my pool[23:01:39] *** tbatchelli_m has quit IRC[23:02:18] <yo61> Actually, looks like compression is off on the pool[23:02:32] *** daleg has quit IRC[23:02:43] <yo61> Just set it to lz4[23:03:08] <jesse_> remember, YMMV=)[23:04:00] <yo61> Tell you what, zpool scrub takes a long time[23:04:37] <yo61> Not sure if that's just how it is, or if my hardware is performing badly[23:05:27] <rmustacc> scrubs generally aren't the fastest thing.[23:05:35] <rmustacc> Due to the nature of the workload vs. disks.[23:06:21] <yo61> I figured as much.[23:06:30] *** W0rmDr1nk has joined #smartos[23:06:44] *** EspressoLuv has quit IRC[23:07:36] <yo61> <sigh> If I could only create zones... :)[23:08:07] <rmustacc> Well, there is always more you can do to debug, just depends how much you're willing to learn.[23:08:16] <yo61> Oh, I'm happy to learn[23:08:25] <yo61> I'm just a bit directionless[23:08:52] <rmustacc> So the questions to ask are 'why is smf in this current state?'[23:08:53] <yo61> I actually think you learn more about a system when you're trying to fix it when it's broken[23:09:03] <rmustacc> At least that's my first question.[23:09:10] <rmustacc> Basically, what exactly is this state.[23:09:21] <rmustacc> Why exactly did the EMI service fail?[23:09:25] *** chorrell has quit IRC[23:09:32] <rmustacc> Which will end up being a bunch of DTrace + mdb.[23:10:01] *** mgls has quit IRC[23:10:55] <yo61> What would I DTrace? Is there some process that creates the zone?[23:11:42] <rmustacc> First I would probably try to understand the current state from a created zone and then understand a bit more about how we got there.[23:12:05] *** benbange1t has joined #smartos[23:12:10] <rmustacc> The fact that EMI failed is very curious and the fact that you observed the other day that enabling it generated some of the same failures is useful.[23:12:13] *** W0rmDr1nk has quit IRC[23:13:44] *** W0rmDr1nk has joined #smartos[23:13:50] *** Alek has joined #smartos[23:14:12] *** benbangert has quit IRC[23:14:19] *** deirdres has quit IRC[23:14:32] *** roast__ has joined #smartos[23:15:13] *** roast_ has quit IRC[23:18:41] <jesse_> one thing you could try, create a zones pool on that disk you used to test re-install with the suspicious zfs options and see if it still works[23:19:06] <jesse_> less dtrace and mdb=)[23:19:09] *** altf2o has joined #smartos[23:20:02] <yo61> True[23:20:22] <yo61> How do I list what is run by the EMI service?[23:20:51] <rmustacc> You look at it's start method.[23:21:00] <yo61> exec=':true'[23:21:14] <yo61> EMI is actually run by svc.startd before any services are started. Thus, there is nothing for these methods to do.[23:21:28] <rmustacc> /lib/svc/method/manifest-import[23:22:02] <yo61> How would I find that?[23:22:40] <yo61> ie/. how did you know that EMI runs that?[23:24:12] <rmustacc> Well, experience with EMI.[23:24:17] <rmustacc> But let me go take a moment to get that fully.[23:24:32] *** szaydel has joined #smartos[23:27:23] <yo61> So this is the command that fails: /usr/sbin/svcadm _smf_repository_switch perm[23:27:31] *** andy_js has quit IRC[23:27:39] <rmustacc> And does that cause everything to bail out or?[23:27:49] *** andy_js has joined #smartos[23:28:19] <yo61> Well...[23:28:33] <yo61> If I run manifest-import[23:28:40] *** tonist has quit IRC[23:29:06] <yo61> I get the warning about WARNING: svccfg apply /etc/svc/profile/generic.xml failed[23:29:17] <yo61> Then svcadm: failed to switch repository: File operation error:[23:29:28] <yo61> Repository switch back operation failed, please check the system log for the[23:29:29] <yo61> possible fatal error messages.[23:29:42] <yo61> That comes from /usr/sbin/svcadm _smf_repository_switch perm[23:29:50] <yo61> The console log messages are:[23:29:57] <yo61> svc.configd: Fatal error: Backend copy failed: opening /etc/svc/volatile/fast_repository.db: No such file or directory[23:30:00] <yo61> svc.configd: Fatal error: Backend copy failed: remove /etc/svc/repository.db-XYague: Permission denied[23:30:04] <yo61> svc.configd: Fatal error: Backend copy failed: remove /etc/svc/volatile/fast_repository.db: No such file or directory[23:31:16] <rmustacc> So, I have no fast_repository.db in my images.[23:31:31] <yo61> Before that, svccfg apply /etc/svc/profile/generic.xml failed[23:32:14] <rmustacc> So in the gz edit /lib/svc/method/manifest-import and add a 'set -o xtrace' to it[23:32:31] <rmustacc> Run it and get the whole thing somewhere if you don't mind.[23:34:16] *** mamash has left #smartos[23:35:07] <yo61> https://gist.github.com/robinbowes/6556376[23:38:18] <rmustacc> What does pstack $(pgrep startd) show in that zone?[23:38:37] <yo61> This is the global zone, right?[23:38:46] <rmustacc> No, sorry, you want to run that command in the zone.[23:39:16] <yo61> OK[23:39:24] <yo61> But that xtrace was in the gz?[23:40:25] <yo61> https://gist.github.com/robinbowes/6556432[23:40:45] <rmustacc> Yeah, the xtrace was in the gz.[23:41:09] <rmustacc> Now in the zone.[23:41:21] <rmustacc> mdb -p $(pgrep startd) -e '807b235/s'[23:41:44] <yo61> 0x807b235: %s exited with status %d[23:41:53] <rmustacc> Aha.[23:42:14] <rmustacc> Okay, great.[23:42:18] <rmustacc> This is progress.[23:42:24] <yo61> W00t![23:42:31] <rmustacc> So the weird thing is that EMI failed this time.[23:42:38] <rmustacc> Err, failed at boot, but not when you ran it manually.[23:42:44] <rmustacc> So that says that we should reboot the zone.[23:42:59] <rmustacc> Because the files in /lib come from the GZ that modification will be there when you boot the zone.[23:43:03] <rmustacc> So we should get a better sense of that first boot.[23:43:14] <yo61> OK, so vmadm reboot?[23:44:21] <rmustacc> Yeah[23:45:20] <yo61> Not liking it. OK to stop it and start it?[23:46:00] <yo61> Oooh, more console output[23:46:04] <yo61> I was connected when it booted[23:46:36] <yo61> https://gist.github.com/robinbowes/6556494[23:51:33] *** vizant0 has joined #smartos[23:52:58] <yo61> Does that shed any further light?[23:53:18] <rmustacc> A bit.[23:53:24] <rmustacc> And the service log should now have the full set -o xtrace output.[23:56:12] *** vizant0 has quit IRC[23:56:20] <yo61> It does indeed: https://gist.github.com/robinbowes/6556601[23:56:20] *** mattconnolly has quit IRC[23:56:58] <rmustacc> Okay. So that basically points us back to that message on the console.[23:57:11] *** mattconnolly has joined #smartos[23:57:12] <rmustacc> The next thing to do is to look at the svcadm source and understand exactly what it's doing and what it's found that's wrong.[23:58:59] <yo61> Is that in https://github.com/joyent/smartos-live ?[23:59:39] <rmustacc> It's in illumos-joyent[23:59:43] <yo61> k