December 5, 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 5, 2012 [00:02:09] <yofuh> serial dowsnt work in magic ways, you have to set console settings as boot arguments[00:03:17] <chriss-> we stumbled across it once i guess[00:09:08] *** jdefelice has joined #smartos[00:13:40] *** dysinger has quit IRC[00:16:14] *** dysinger has joined #smartos[00:16:50] *** josephholsten has quit IRC[00:18:49] *** daleg has quit IRC[00:19:10] *** daleg has joined #smartos[00:29:17] *** neophenix has quit IRC[00:37:27] *** enmand has joined #smartos[00:39:22] <xenol> I would be interested in creating SmartOS package mirror. I emailed smartos at joyent dot com (I think it is deidres), but I didn't receive any response. Is there any rsync location avaiable? Wiki doesn't say anything.[00:40:02] <e^ipi_laptop> it's just http, you can wget --mirror it[00:41:14] <xenol> I am not happy to hear that :([00:45:31] <e^ipi_laptop> *shrug*[00:46:38] *** trentster has left #smartos[00:46:56] <ryancnelson> why? wget mirror won't fetch a thing a 2nd time if the datestamp hasn't changed, or the size.[00:47:35] <ryancnelson> rsync requires *CPU* on the server side. doesn't scale for 1000 mirror-ers[00:47:51] <ryancnelson> http caches great.[00:48:26] <ryancnelson> feel free to wget mirror it once, and check it into your 3TB github account[00:53:42] <xenol> ryancnelson: what is the estimated size of the whole repository?[01:30:59] *** ocherno has quit IRC[01:34:39] *** szaydel has quit IRC[01:44:54] *** jamesd has joined #smartos[01:48:37] <arekinath> hmm... I just pulled down the base-1.8.2 ds and tried to use it, and it's not picking up resolvers correctly... it's trying to use "mdata sdc:resolvers" to pick them up, but when I try this, on any of the zones, I get "Error getting metadata for key sdc:resolvers: FAILURE"[01:48:42] <arekinath> which doesn't seem very enlightening[01:48:48] <arekinath> are there logs somewhere on why it failed?[01:50:25] <arekinath> the zones all have the resolvers property set in their JSON. I can see it in vmadm get.[01:56:23] <arekinath> ah. I see in /usr/vm/lib/metadata/agent.js, it returns FAILURE for anything that is not a string, and resolvers is an array[01:56:26] <arekinath> hmm[01:58:01] *** ahaydock has quit IRC[01:59:20] *** kfr- has quit IRC[02:02:54] <arekinath> it's still like that in latest smartos in git, too. is SDC out of sync with the publically released code again? and the zoneinit for base-1.8.2 expects SDC?[02:05:35] *** jasonpincin has quit IRC[02:06:46] <arekinath> also, the zoneinit for 1.8.2 has "mdata-get $1 | grep -v '^No metadata'", but when mdata-get returns an error it has "Error getting metadata" at the start, not "No metadata"[02:15:22] *** ganbold has quit IRC[02:27:37] <xenol> I am having a dilemma if I should go with HW raid or zfs raid. Going with HW raid will break some features of ZFS, but everyone seems to be using HW raid these days[02:27:46] <xenol> what is the proposed setup under smartos?[02:28:21] <e^ipi_laptop> "everyone uses HW" because not everyone has zfs[02:28:48] *** bixu_ has quit IRC[02:29:00] *** bixu has joined #smartos[02:29:13] <xenol> e^ipi_laptop: simply put, I just create virtual disk on the controller and use zfs for raid, right?[02:32:05] <Triskelios> no, you should expose real disks[02:33:28] <xenol> ok, thanks Triskelios[02:35:56] <richlowe> e^ipi_laptop: the problem is not necessarily "everyone uses" but "some HBAs make it hard not to use"[02:36:08] <e^ipi_laptop> yeah, that's fair[02:36:23] *** llirkaz has quit IRC[02:36:48] *** ganbold has joined #smartos[02:37:55] <Triskelios> xenol: the sole useful aspect of a hardware RAID controller for ZFS is that you can disable cache flushes if it has a battery backup for the write cache. but you should not use the RAID firmware at all, if possible[02:39:58] <jaakkos> how does adding a separate ZIL compare to such a setup?[02:40:59] <jaakkos> an SSD would be slower but bigger than the RAM on the controller, but is there more to that?[02:41:24] <xenol> Triskelios: if I understand it correctly, building ZFS raid on top of HW raid kills some of ZFS features such as self-healing?[02:41:27] *** llirkaz has joined #smartos[02:43:51] <richlowe> and puts more faith in the raid firmware than it deserves, and ...[02:46:29] <Triskelios> xenol: it prevents repair from normal redundancy (from ditto blocks is still possible), and efficient resilvering[02:46:50] <xenol> well, I heard arguments for using HW raid controllers such as "it is dedicated piece of HW for raid and is much faster than SW raid"[02:46:58] <Triskelios> also, prevents normal responses to disk failures and replacements[02:47:14] <xenol> having thought about that it is bullshit as firmware is also a software[02:47:40] <xenol> Though I would still be interested in some benchmarks on this[02:48:42] <LeftWing> Meh, there's almost no point. Firmware-based RAID (as there is, of course, no Hardware RAID) is just stupid. Uninstrumentable, unfixable garbage. Good Riddens.[02:49:18] <LeftWing> Also, it's been a Firmware Shitshow of a week thus far, and I'm over it.[02:49:27] <Triskelios> xenol: modern CPUs are comparable in I/O performance to the ASIC-assisted embedded CPUs on most RAID cards. it's easy to get performance examples from the past decade, though[02:49:45] <Triskelios> (most high end RAID cards, that is. low end stuff is just shit)[02:51:59] *** Tekni has quit IRC[02:58:23] *** xinkeT has joined #smartos[03:01:36] *** wdent has joined #smartos[03:10:17] *** rc10 has joined #smartos[03:27:06] *** szaydel has joined #smartos[03:27:42] <richlowe> and the rest of it is a per-firmware revision gamble[03:27:51] <richlowe> as I presume the former fishworkers can expound upon at length.[03:34:37] *** kfr- has joined #smartos[03:49:09] <bixu> xenol: I can say from painful experience that a basic (non-RAID) SAS controller or the like is a lot more fun with ZFS than adding HW RAID.[03:49:22] <bixu> Many long nights spend wrestling shitty drivers/firmware.[03:49:50] <bixu> I've even had good luck the the really cheap Intel stuff (AHCI?).[03:54:12] *** ryancnelson has quit IRC[03:54:47] <rmustacc> You generally want to avoid ahci if you can for lack of hotplug.[03:54:57] <rmustacc> If that's the kind of thing you care about.[03:57:07] <bixu> That's a good point.[03:57:28] <bixu> For a big array, I used the LSI JBOD chipset (2808?).[03:59:49] *** ira has quit IRC[04:01:26] *** AlainODea has joined #smartos[04:01:37] <jamesd> rmustacc, ahci is better than eide, especialy the controller some bright engineer decided it was good to use in the u5 and u10[04:05:49] <chriss-> we have gone away from hw raid, zfs is much easier to deal with especially if you need/want to change things on the fly[04:06:04] <chriss-> and you deal with failures right in the os rather than behind the curtain[04:06:24] <chriss-> hp gear we have though wont do passthrough[04:06:28] <chriss-> sad[04:08:03] <chriss-> cant shuffle bigger drives into the mirror on the fly with hw raid[04:11:04] <rmustacc> jamesd: Sure, but I just kind of assume these days the question is AHCI vs. SATA/SAS controller.[04:12:08] *** jasonpincin has joined #smartos[04:19:52] *** wdent has quit IRC[04:22:02] *** codingstream has joined #smartos[04:23:04] *** codingstream has joined #smartos[04:47:01] *** marsell has quit IRC[05:09:48] *** rc10 has quit IRC[05:26:41] *** bixu_ has joined #smartos[05:30:21] *** wdent has joined #smartos[05:35:00] *** wdent has quit IRC[05:35:15] *** avrntsv has joined #smartos[05:36:52] *** PhoenixDraco has joined #smartos[05:43:28] *** abhishekvasisht has joined #smartos[05:50:41] *** sachinsharma has joined #smartos[05:51:47] *** PhoenixDraco has quit IRC[06:00:21] *** toebu has quit IRC[06:05:44] *** Webhostbudd has joined #smartos[06:07:15] *** chriss--- has joined #smartos[06:10:02] *** chriss- has quit IRC[06:10:19] *** sachinsharma has quit IRC[06:16:27] *** rc10 has joined #smartos[06:17:50] *** darjeeling has quit IRC[06:22:51] *** sachinsharma has joined #smartos[06:36:02] *** rc10 has quit IRC[06:41:34] *** abhishekvasisht has quit IRC[06:45:00] *** dap has quit IRC[06:48:12] *** dysinger has quit IRC[06:50:00] *** jesser has quit IRC[07:00:24] <Licenser> good morning :)[07:07:20] *** rc10 has joined #smartos[07:11:29] *** marsell has joined #smartos[07:17:15] *** Guest62621 has joined #smartos[07:17:24] <Guest62621> anybody here?[07:20:03] <LeftWing> Guest62621: Sure. What's up?[07:22:19] <Guest62621> can i install smartos on only one server?[07:23:07] <rmustacc> There's nothing in smartos that stops you from booting it onto multiple machines.[07:23:23] <rmustacc> But keep in mind that it is a live image and thus you will need boot media that isn't your internal hard drives to boot it.[07:23:34] <rmustacc> That medium can be the newtork, cdrom, usb, or something else.[07:23:39] <arekinath> there is also nothing that stops you from booting only one machine on its own, if your question was that, too[07:23:46] <arekinath> (it sounded ambiguous to me)[07:26:01] <Guest62621> i'm not good at english well,give me half sec to understand what u said[07:27:05] <arekinath> smartos is ok with any number of servers >= 1[07:27:31] <arekinath> or even >= 0 if you say VMs are 0.5 servers or something[07:27:33] <Licenser> Hmm is anyone running SmartOS on VMWare fusion here? In more then one instance at a time that is[07:27:59] <arekinath> >0 rather[07:28:29] <Licenser> arekinath I was about to say, unless smartOS allowes to instlall a negative number of VM's to increase server capacity you need at least one[07:28:55] <Licenser> which would be cool, I'd totally buy the cheepest system I can get and install -1 PB Ram VM on that to have a kick ass system[07:29:21] <arekinath> Licenser: smartos is pretty magic, but... probably not quite that magic. ;) the metaphysical operating system![07:29:38] <Licenser> would be cool[07:29:52] <Licenser> VMWare can do that I had a VM using 16EB of shared memory the other day[07:31:06] <Licenser> https://www.dropbox.com/s/nx9um36hyf3nrwx/Screen%20Shot%202012-12-04%20at%2014.05.51.png <- see[07:31:07] *** bixu_ has quit IRC[07:31:36] <arekinath> Licenser: r u some kind of wizard?[07:31:57] <Licenser> Yap, I wield erlang-magic![07:32:10] <arekinath> :O[07:32:36] *** jesser has joined #smartos[07:35:27] * Licenser shakes his fist at VMWare[07:35:31] <Licenser> now it killed my system :([07:42:21] *** jesser has quit IRC[07:43:03] *** sachinsharma has quit IRC[07:47:27] *** jesser has joined #smartos[07:50:17] <rmustacc> Licenser: What about Fusion?[07:51:25] <Licenser> rmustacc when I try to run 4 SmartOS VM's at the same time the system goes to death here (aka cpu goes to 100% on all cores mostly kernel time), it seems to have to do soething with the promiscuous mode the NIC needs to be set in[07:52:25] <rmustacc> Licenser: You fire up four instances of VMware Fusion?[07:52:43] <Licenser> no of cause not, just 1 instance of fusion running 4 VM's[07:53:08] <rmustacc> So you have four different VMs running in Fusion.[07:53:21] <rmustacc> And who's CPU is maxing out? The host or the guest?[07:53:38] <rmustacc> And how much dram are you assigning to each of the Fusion VMs?[07:53:46] <Licenser> host CPU, so my Mac's[07:54:05] <Licenser> each VM has 1G memroy, 16 total in the host system.[07:56:55] *** jamesd has quit IRC[08:00:23] <MerlinDMC> morning[08:02:01] <Licenser> morning merlin[08:03:37] <MerlinDMC> way too early -.-[08:04:03] <MerlinDMC> Licenser, is that vmware phenomenon fixed? ^^[08:04:11] <Licenser> No :([08:04:23] <MerlinDMC> i'll try to run some vms in promisc mode here[08:04:52] *** jesser has quit IRC[08:06:39] <Licenser> that'd be cool if you could try :)[08:07:06] <MerlinDMC> yeah ... you just want to see my pinging out ... i know ... but that will not happen![08:07:46] <MerlinDMC> how many vms do i need? ... 2, 3 or more?[08:08:15] <rmustacc> Also are you sure that the VMs aren't swapping?[08:08:47] <Licenser> rmustacc pretty much as I can run 4 of them in non promiscous mode[08:08:50] <Licenser> MerlinDMC 4[08:09:05] <Licenser> as soon as 1 is promiscous things go to hell for me[08:09:22] *** ryancnelson has joined #smartos[08:09:54] *** ryancnelson has left #smartos[08:11:02] *** sachinsharma has joined #smartos[08:14:31] <MerlinDMC> vm 3 + 4 booting ...[08:16:38] <Licenser> MerlinDMC you got the put in your password for promiscouse mode query?[08:16:39] <MerlinDMC> in NAT mode it's all fine ... i did switch then to bridged mode and ... i'm fine[08:16:49] <MerlinDMC> and all are promisc[08:18:03] <MerlinDMC> but now i don't know what i should do with those VMs ;)[08:18:10] <Licenser> darn I wonder what causes this :([08:18:16] <Licenser> MerlinDMC install FiFo :P[08:18:26] <MerlinDMC> do all of your vms have different mac addresses?[08:19:23] <Licenser> yup randomized them[08:19:55] <MerlinDMC> and you're on 10.8.2 as well? ... not 10.8.3?[08:20:42] <Licenser> 10.8.2 yap double checked[08:21:13] <MerlinDMC> k ... then it's simple ... vmware just hates you ... and safari6 hates me ;)[08:22:06] <Licenser> sucks well I'll keepo trying :)[08:27:29] *** Guest62621 is now known as daniel_wu[08:28:59] <arekinath> hmm... I wonder how hard localising the smartOS wiki would be... does confluence have a notion of translations of a page?[08:29:54] <arekinath> or do you have to set up a separate space for each language? :|[08:30:28] <MerlinDMC> isn't english enough? ^^[08:32:37] <arekinath> MerlinDMC: for those who can't read it very well, it can be pretty tricky. the English on some of those pages is pretty dense prose[08:34:41] *** jesser_ has joined #smartos[08:35:49] <MerlinDMC> arekinath, but you're probably producing more problems if not all changes in english pages get translated early enough ... languages will differ from each other and possibly point in different directions because of translation delays[08:36:18] <arekinath> MerlinDMC: yeah, that is true I guess[08:36:19] <arekinath> hmm[08:36:40] <arekinath> especially with how often we seem to like changing things. ;)[08:37:35] <MerlinDMC> yeah ... about changing things ... did mdata-get get an update lately?[08:38:07] <MerlinDMC> Alex posted in the mailing list about zoneinit problems on base 1.8.2[08:38:47] <arekinath> arekinath and Alex are one and the same, haha[08:38:52] <MerlinDMC> but that's not reproducible here with an liveimg 20121101 ... my mdata-get seems to give me other error strings back[08:39:04] <MerlinDMC> who can know that?[08:39:09] <arekinath> do you have a list in resolvers?[08:39:46] <MerlinDMC> i don't have resolvers set ... so the default of 8.8.8.8 8.8.4.4 gets set[08:39:57] <arekinath> yeah, it will be fine if you don't have it set[08:40:02] <arekinath> you have to have a resolvers list set in the zone json[08:40:04] <arekinath> for it to bug out[08:40:18] <MerlinDMC> hm ... let me set some[08:40:21] <taspts> paging konobi ..[08:40:37] <daniel_wu> basiclly ,the people in this chanal at this monent are aisan,is that right?[08:40:49] <MerlinDMC> daniel_wu, nope[08:41:11] <daniel_wu> others is sleeping[08:41:20] <arekinath> europe is awake about now[08:41:30] *** kschiess has joined #smartos[08:41:32] <arekinath> and people who stay up late[08:42:04] <daniel_wu> not healthy[08:42:53] <daniel_wu> AIX VS SOLARIS ,WHO WIN?[08:43:38] <daniel_wu> illumos vs solaris ,who win ?[08:43:54] <daniel_wu> illumos vs solaris 11's kernel[08:44:26] <rmustacc> daniel_wu: In what category?[08:44:31] *** daniel_wu has quit IRC[08:44:49] <MerlinDMC> arekinath, ah ... yeah ... and i know why it bugs there! ;)[08:44:52] <arekinath> what was that I don't even[08:45:03] *** ryancnelson has joined #smartos[08:45:17] *** Guest62621 has joined #smartos[08:45:29] <arekinath> MerlinDMC: because mdata / agent.js can't send lists? :P and it's somehow still a list in vmobj, even though it joins them with commas?[08:45:32] <MerlinDMC> arekinath, mdata-get can only read plain strings/integers/booleans ... no arrays or objects :P[08:45:38] *** abhishekvasisht has joined #smartos[08:45:47] <Guest62621> what did u say ,i lost connection to freenode[08:46:23] *** Guest62621 is now known as daniel_wu[08:46:25] <arekinath> MerlinDMC: I couldn't find where it splits it back up into a list though[08:46:26] <MerlinDMC> i did once play with that and was changing that behaviour to be able to send complete objects as json ... for getting json data to chef-solo[08:46:30] <arekinath> MerlinDMC: just where it joins them all up[08:46:44] <rmustacc> daniel_wu: I asked you to clarify your question.[08:47:30] <MerlinDMC> arekinath, maybe that change to mdata-get is missing in the latest smartos release ... mamash should know about that i guess[08:47:34] <daniel_wu> the perfomance and security between aix os and solaris[08:47:46] <arekinath> MerlinDMC: that's what I was thinking, something forgotten there[08:48:23] <MerlinDMC> but also ... if i would set resolvers those would be 8.8.8.8 and 8.8.4.4 ... so i'll just stick with the defaults :P[08:48:26] <rmustacc> That's a pretty broad question. To answer it correctly requires a lot of work.[08:48:35] <arekinath> MerlinDMC: but it should grep -v out the error message, too, not just "No metadata"[08:48:48] <arekinath> MerlinDMC: heh. outgoing dns to 8.8.8.8 doesn't work from all vlans here[08:49:05] <arekinath> firewall is all default deny and fairly tight, out and in.[08:49:26] <rmustacc> daniel_wu: But if you're stuck using AIX and are on power, not much anyone can do for oyu[08:49:29] <rmustacc> *you[08:49:41] <MerlinDMC> arekinath, it should inside mdata() check for the return code of mdata-get instead of greping a string there ... imo[08:50:47] * MerlinDMC back to daily work - i love it![08:51:05] <arekinath> is there a git repo for the zoneinit scripts somewhere?[08:53:24] <arekinath> it would be kinda nice to submit a pull request, but I can't find one, haha. it looks like they're just done internally and baked into base at each release[08:53:35] *** sachinsharma_ has joined #smartos[08:54:49] *** sachinsharma has quit IRC[08:54:49] *** sachinsharma_ is now known as sachinsharma[08:55:49] *** abhishekvasisht has quit IRC[08:56:01] *** abhishekvasisht has joined #smartos[09:03:11] *** _Jared_ has quit IRC[09:04:05] *** codingstream has quit IRC[09:04:11] *** alucardX has joined #smartos[09:08:50] *** danielwu has joined #smartos[09:09:14] *** daniel_wu has left #smartos[09:09:28] *** danielwu is now known as daniel_wu[09:10:38] *** jesser_ has quit IRC[09:11:13] <daniel_wu> quassel is not bad but the name for itself[09:11:36] *** bens1 has joined #smartos[09:11:39] <alucardX> morning[09:11:51] <daniel_wu> afternoon[09:13:11] *** Guest62621 has joined #smartos[09:15:44] <taspts> paging konobi ..[09:16:43] <konobi> mmm?[09:17:10] <taspts> Regarding smartos-discuss grub ipxe.lkrn chain to webserver, here I'm[09:17:44] <konobi> eh?[09:19:13] <konobi> ah yes[09:20:52] <taspts> Konobi: hit us up on the #smartos channel on freenode[09:20:52] <taspts> there's lots to talk/explain[09:21:14] *** darjeeling has joined #smartos[09:21:41] <konobi> taspts: have you tried it using undionly.kpxe?[09:22:43] *** texarcana has quit IRC[09:23:29] <taspts> Yes my first test, and you get a lot of if statements if you have an existing pxe env[09:23:57] *** texarcana has joined #smartos[09:25:14] <taspts> You have to update DHCP and PXE menu's[09:27:26] <taspts> Using ipxe.lkrn you only has to pu the files on a webserver and update menu.lst for that machine.[09:27:40] <taspts> /pu/put/[09:33:39] *** daleg has quit IRC[09:33:59] *** daleg has joined #smartos[09:34:44] <taspts> Konobi: What is your question and what needs to be explained ?[09:35:57] *** ryancnelson has quit IRC[09:37:48] <Guest62621> there is some guy familiar with building xen 4.2 on centos 6?[09:42:26] <arekinath> Guest62621: you might do better to ask on #xen or #centos[09:43:49] *** daniel_wu has quit IRC[09:45:23] <konobi> well, undionly will let you chain directly[09:45:47] <konobi> though i used it based on mac to chain to specific machines[09:46:28] <konobi> taspts: for whatever reason i mis-read your original email... it was late when i responded[09:46:37] *** danielwu has joined #smartos[09:47:34] <konobi> taspts: i used gpxe/ipxe a fair bit when i was doing sdc testing, so interesting to hear what fokls are using... especially with the new menu features[09:59:20] <MerlinDMC> how big should a slog device be? ... that would depend highly on count of synchronous writes right?[10:03:23] <Licenser> MerlinDMC I guess count and size?[10:03:50] <MerlinDMC> hmm ... what happens to sync write in the case slog device is full?[10:04:08] <Licenser> I guess they simply then act as if there were no slog?[10:04:33] <Licenser> from what I understand it's more like Application -> Slog -> disk[10:04:38] <Licenser> stuff is not supposed to stay in the slog[10:04:58] *** sachinsharma has quit IRC[10:05:30] <Licenser> so the moment your slog is full it would just work as if there were no slog, so you face the same latency as if you're writing to disk, which likely would be sucky in that situation since a filled up SLOG means your disk write is satturated to 100%[10:05:40] <Licenser> ^-My understanding of the topic, not facts[10:10:46] <konobi> you'd be hard pressed to fill a slog[10:11:55] *** Webhostbudd_ has joined #smartos[10:12:36] *** KermitTheFragger has joined #smartos[10:14:17] <MerlinDMC> I'm a really unlucky person ;)[10:14:19] <Licenser> konobi is that a challange? :P[10:15:03] * Licenser wonders[10:15:21] <Licenser> are there many applications left that depend on sync writes (I guess ACIDy thing but aside of that?)[10:15:33] <konobi> well, think about it... the slog slurps up writes and then writes those in one go[10:15:49] *** Webhostbudd has quit IRC[10:16:15] <Licenser> so you just have to request enough writes fast enough ;)[10:17:39] <konobi> Licenser: lemme know when you get it[10:17:41] <konobi> =0)[10:18:03] <Licenser> konobi once I've a system slog SLOG I will try :P[10:18:52] <konobi> wonder how the ocz server grade stuff is[10:19:01] <Licenser> but in theory you've just to request sync writes faster then the spinning disks can handle and keep that up long enough for the slog to slowly fill up[10:19:40] <Licenser> sentences like that with this niftly little 'just' in there are ment to go in the famouse last words category[10:22:38] <Licenser> okay I'm thrilled managed to get 3 VM's up :D now to test a distributed fifo![10:23:57] *** bixu_ has joined #smartos[10:27:46] *** mamash has joined #smartos[10:32:14] *** bixu_ has quit IRC[10:34:36] *** copec has quit IRC[10:36:50] *** copec has joined #smartos[10:41:04] <taspts> Kanobi: I my first try was using my existing pxeinstall (centos4 version) and smartos didn't work. next try was pxe chainloading ipxe (fedora 17 version). Next git version, that worked. Had to use if in dhcpd to keep the existing DHCP/PXE working. I wanted a working solution that can boot with no intervention from my part. Using httpd server is much faster that tftp and has more possiblities like I have several http server but only one DHCP active on a given[10:42:42] *** llirkaz has quit IRC[10:46:06] <taspts> Konobi: In my first try was using my existing pxeinstall (centos4 version) and smartos didn't work. next try was pxe chainloading ipxe (fedora 17 version). git version worked with httpd. Had to use IF in dhcpd to keep the existing DHCP/PXE working. I wanted a working solution that can boot with no intervention from my part. Using httpd server is much faster that tftp and has more possiblities like I have several http server but only one DHCP active on a give[10:47:35] *** llirkaz has joined #smartos[10:48:05] *** avrntsv has quit IRC[10:51:48] <konobi> yup... i was doing that along with some scripting of the ipxe scripts to boot in different ways[10:55:14] *** llirkaz has quit IRC[10:55:17] <konobi> taspts: no doubt the dhcp change was the if userclass stuff[10:58:09] *** avrntsv has joined #smartos[11:01:35] *** llirkaz has joined #smartos[11:02:05] *** Webhostbudd_ has quit IRC[11:09:07] *** Kireji has left #smartos[11:34:16] *** avrntsv has quit IRC[12:02:06] <konobi> yay ebay[12:03:15] *** avrntsv has joined #smartos[12:03:59] *** haydock_ has joined #smartos[12:06:38] *** haydock has quit IRC[12:07:12] <rc10> hi, any dtrace one liner to check zil activity ?[12:08:23] *** llirkaz has quit IRC[12:09:43] <arekinath> zpool iostat can give you a vague indication[12:09:54] <arekinath> otherwise you need one of the zilstat dtrace scripts or similar[12:10:38] <rc10> I have checked zilstat script but throws error - so wanted a one liner[12:11:02] <arekinath> try zpool iostat -v 2 if you just want to see if it's doing anything at all[12:14:05] *** llirkaz has joined #smartos[12:20:49] <konobi> which zilstat?[12:28:51] *** darjeeling has quit IRC[12:31:34] *** sachinsharma has joined #smartos[12:42:03] *** riceo has joined #smartos[12:59:58] *** abhishekvasisht has quit IRC[13:24:46] *** ira has joined #smartos[13:27:59] *** szaydel has joined #smartos[13:33:21] *** szaydel has quit IRC[13:39:03] *** sachinsharma has quit IRC[13:43:29] *** Azbruh_ has quit IRC[13:46:54] *** alucardX has quit IRC[14:05:50] *** mamash has left #smartos[14:12:59] *** mamash has joined #smartos[14:17:25] *** wdent has joined #smartos[14:22:27] *** avrntsv has quit IRC[14:25:15] *** darjeeling has joined #smartos[14:31:25] *** rc10 has quit IRC[14:34:45] *** Azbruh has joined #smartos[14:38:08] *** scanf has quit IRC[14:52:08] *** ahaydock has joined #smartos[15:10:22] *** alucardX has joined #smartos[15:12:43] *** alucardX has quit IRC[15:14:58] *** jim80net has joined #smartos[15:23:22] *** taspts has quit IRC[15:24:10] *** taspts has joined #smartos[15:34:55] *** apm1 has joined #smartos[15:36:58] *** _Jared_ has joined #smartos[15:38:46] *** jamesd has joined #smartos[15:39:36] *** apm1 has left #smartos[15:49:57] *** mamash has left #smartos[15:50:50] *** neophenix has joined #smartos[15:51:33] *** chorrell has joined #smartos[16:04:28] *** kfr- has quit IRC[16:35:24] *** codingstream has joined #smartos[16:39:20] *** dysinger has joined #smartos[16:40:46] *** CarlosC has joined #smartos[16:41:15] *** khushildep has joined #smartos[16:41:31] <khushildep> Hey folks[16:41:37] <khushildep> Trying to get NTP working but think I'm hitting http://mail.opensolaris.org/pipermail/indiana-discuss/2009-November/016997.html - is there a workaround for SmartOS?[16:41:48] <khushildep> I'm assuming yes but that my GoogleFU is not strong…[16:48:23] <jperkin> are you trying to enable ntp in a zone? that's not going to work.[16:50:16] <khushildep> Assuming it's do with the the adjtime call?[16:50:57] <jperkin> almost certainly, those syscalls are usually privileged[16:52:04] <jperkin> the default setup is to run a configured ntp in the GZ, so it should not be necessary[16:53:05] <khushildep> *nods* gotcha - thanks mate[16:54:25] *** jamesd has quit IRC[16:55:18] *** jamesd has joined #smartos[16:57:08] *** jamesd has quit IRC[16:58:38] *** jamesd has joined #smartos[16:58:43] *** rc10 has joined #smartos[17:00:07] *** jamesd has quit IRC[17:01:37] *** jamesd has joined #smartos[17:04:15] *** avrntsv has joined #smartos[17:09:15] *** ryancnelson has joined #smartos[17:10:37] *** mamash has joined #smartos[17:11:10] *** bens1 has quit IRC[17:20:46] *** rancor has joined #smartos[17:27:07] *** mamash has left #smartos[17:28:29] *** rc10 has quit IRC[17:30:32] *** khushildep has quit IRC[17:36:26] *** chorrell has quit IRC[17:38:01] *** defa has joined #smartos[17:42:16] <defa> Has anyone also experienced trouble booting from smartos-20121130T194746Z-USB.img.bz2? It has been pulled without notice from download.joyent.com … th Dec-03 USB image got stuck at a grub boot prompt.[17:43:36] <jperkin> defa: it was a broken image, the latest is now fixed.[17:44:27] <defa> jperkin: no, the fixed one is also broken...[17:45:12] <jperkin> ah right, I haven't tried that one yet.[17:45:31] *** KermitTheFragger has quit IRC[17:46:23] <defa> … smartos-discuss there was an announcement on dec 3rd. that the 2012-11-30 build was broken and has been replaced by 2012-12-03 but the USB image does not boot and has been pulled from the server… without notice :-([17:47:00] *** izaki_ has joined #smartos[17:48:43] <jperkin> defa: e^ipi_laptop should be awake soon and will know the deal.[17:49:00] <ryancnelson> what's your question?[17:49:15] <ryancnelson> you wanted us to leave the broken build available for download?[17:49:27] <ryancnelson> the usb image was bad, we took it down last night[17:50:31] <ryancnelson> the image was lacking a "menu.lst" in the grub menu.[17:50:51] <defa> Ryancnelson: no, that's okay and latest-usb has been updated to a working version, that's fine. I was just asking because it was pulled without notice and caused some trouble here.[17:50:57] <ryancnelson> if you have that image, and want to try using the menu.lst from the iso (or something), that may work.[17:51:02] <ryancnelson> there'll be a new build soon[17:51:48] <defa> is there a list besides smartos-discuss for announcements regarding new images?[17:52:37] <defa> shit happes… didn't want to complain in any way ;-)[18:00:39] *** jim80net has quit IRC[18:04:41] *** rancor has quit IRC[18:05:05] *** rancor has joined #smartos[18:08:11] *** codingstream has quit IRC[18:13:36] *** Webhostbudd has joined #smartos[18:27:15] *** dap has joined #smartos[18:31:48] *** mamash has joined #smartos[18:35:59] *** cncfanatics has joined #smartos[18:54:51] *** jim80net has joined #smartos[18:56:46] *** rancor has quit IRC[18:57:24] *** neophenix_ has joined #smartos[18:57:50] *** dap has quit IRC[18:59:03] *** neophenix has quit IRC[18:59:03] *** neophenix_ is now known as neophenix[18:59:57] *** ryancnelson has quit IRC[19:01:18] *** ryancnelson has joined #smartos[19:02:21] *** khushildep has joined #smartos[19:08:17] *** khushildep has quit IRC[19:14:10] *** LucidDreamZzZ has joined #smartos[19:14:55] *** tdowg1 has joined #smartos[19:17:24] *** tdowg1 has quit IRC[19:17:50] *** tdowg1 has joined #smartos[19:23:36] *** ryancnelson has quit IRC[19:31:34] <taspts> JMB363 SATA/IDE Controller have a disk that is not recognized under SmartOS (Same disk as I boot from), prtconf -v list the controller. Any idees ?[19:31:52] *** dap has joined #smartos[19:32:30] *** cncfanatics has quit IRC[19:36:38] *** mamash has left #smartos[19:40:32] *** mamash has joined #smartos[19:46:30] *** LucidDreamZzZ has quit IRC[19:50:08] *** microburn has joined #smartos[19:50:23] *** microburn has quit IRC[19:56:21] *** rancor has joined #smartos[19:59:39] *** LucidDreamZzZ has joined #smartos[20:00:26] *** SebasP has joined #smartos[20:00:35] <Licenser> for the record: deleting the VM you're working in is a BAD thing[20:07:14] <nahamu> Licenser: ouch![20:07:47] <Licenser> nahamu fortunately just a test system but still you sart to get confused when a service is ripped away under your feet :P[20:08:04] <jeffpc> defa: hrm, I got stuck at grub prompt with 20121203 as well[20:08:28] <jeffpc> ah[20:08:40] <jeffpc> I see it's no longer on the website[20:09:05] * jeffpc grabs 20121115[20:10:04] *** viridari has quit IRC[20:10:32] <defa> jeffpc: the boot menu file was missing...[20:10:46] <jeffpc> yep, I saw the scrollback after I spoke[20:17:42] *** mamash has left #smartos[20:25:49] *** ira has quit IRC[20:27:08] <ismell> does joyent provide private clouds?[20:27:13] <ismell> or managed private clouds[20:28:21] <nahamu> ismell: they sell the software to run your own private cloud.[20:28:23] <MerlinDMC> sdc is for sale afaik :)[20:28:33] <ismell> ah, well I don't want to manage it ;)[20:28:58] <MerlinDMC> you want a "private" part in their cloud? :P[20:29:13] <nahamu> talk to their sales people.[20:29:50] <ismell> maybe I will. We run a site that is COPPA compliant and there are a bunch of regulations around it[20:30:01] <elijah-mbp> ismell: we *do* do some managed installs for folks.[20:30:18] <elijah-mbp> it's generally very large engagements.[20:30:34] <ismell> our stuff is pretty small[20:30:42] <ismell> just a dife website we run[20:30:45] <ismell> just tired of realing with hardware[20:30:49] <ismell> *dealing[20:31:07] <elijah-mbp> is it the kind of thing where if we stuck you in a VLAN and you had VPN/firewall access you could get along?[20:31:25] <elijah-mbp> our network guys are awesome and helpful. ;-)[20:31:49] <ismell> I'll get some more clarification on what the law says[20:32:02] <elijah-mbp> anyway, one of the sales guys can probably help you navigate/interpret particular laws that you're concerned about.[20:32:29] <elijah-mbp> and help spell out places we might be out of spec and need to improve, for your sake and others. :-)[20:33:23] <elijah-mbp> (for making your stuff runnable in JPC / on someone else's SDC install)[20:34:08] <ismell> I'll craft an e-mail and see[20:36:42] *** avrntsv has quit IRC[20:37:06] <jasonpincin> I grabbed latest image, booting on new server and I'm getting dumped to login prompt instead of expected config questions[20:38:21] <jasonpincin> am i missing something obvious? i verified smartos=true in boot image grub config[20:42:36] *** ryancnelson has joined #smartos[20:49:04] *** cncfanatics has joined #smartos[20:49:16] <MerlinDMC> jasonpincin, iso or usb image?[20:50:35] <ryancnelson> you don't get the questions if you already have a zpool[20:50:43] <ryancnelson> zpool list[20:51:37] *** neophenix has quit IRC[20:55:01] <jasonpincin> do not have a zpool, breand new unconfigured server[20:55:10] <jasonpincin> i've tried both iso and usb[20:55:50] <jeffpc> cute... http://pastie.org/5485399[20:56:14] <MerlinDMC> jeffpc, that's old ... there is an issue for that[20:56:21] <jeffpc> ok[20:58:03] <jasonpincin> logged in at prompt verified no zpool[20:58:30] <MerlinDMC> but your disks are found? ... so the controller is supported?[20:58:31] <ryancnelson> what's zpool import say?[20:59:02] <jasonpincin> 1st time trying on this particular hardware... not seeing nic in ifconfig -a, if it doesn't support the nic would that prevent it?[20:59:04] <jasonpincin> yeah disks show[20:59:11] <ryancnelson> dladm show-phys[20:59:20] <ryancnelson> … ifconfig -a won't show nics until they're plumbed[20:59:50] <jasonpincin> dladm show-phys returns nothing[20:59:54] <jasonpincin> zpool import returns nothing[20:59:59] <ryancnelson> but, no nics shouldn't stop the config questions, i don't think[21:00:16] <ryancnelson> so, you have no nics, apparently.[21:00:19] <ryancnelson> what nics are there?[21:00:26] <jasonpincin> this is an IBM blade, i was concerned about it[21:03:11] *** neophenix has joined #smartos[21:03:19] <ryancnelson> what's uname -a say?[21:03:19] *** SebasP has quit IRC[21:03:28] <ryancnelson> are you using the dec 3 build, or the 11/15 build[21:04:17] <jasonpincin> I think this is the 11/15 because I didn't reflash the usb key, i did try the latest via an ISO boot (dec 3)[21:04:26] <ryancnelson> (shouldn't really matter much, if you can't see your nics, that didn't change. but there might be a setup issue… we tore down the dec 3 build for some bugs)[21:04:32] <jasonpincin> uname -a : 5.11 20121101T182015Z[21:04:35] <ryancnelson> usb, anyway[21:04:38] <rmustacc> I would start with prtconf -dD and look for your storage controller and make sure it actually has a driver attached.[21:04:40] <ryancnelson> ok, so 11/01[21:04:57] <ryancnelson> format < /dev/null[21:04:57] <ryancnelson> shows you some disks?[21:05:14] <jasonpincin> yeah, two of em[21:05:21] <jasonpincin> Which is what it has[21:05:51] <jasonpincin> blade chassis has cisco brand ethernet adapters, not sure exactly which model[21:06:00] <MerlinDMC> ryancnelson, configuration doesn't show up if there is no nic[21:06:05] <jeffpc> so, I installed using the nov 30 image, and now I'm trying to create a zone using the nov 15 image... it's still taking ages[21:06:06] <MerlinDMC> just checked that[21:06:07] <rmustacc> If dladm show-phys returns nothing then you're sol with nics.[21:06:14] <ryancnelson> ah. that's that, then[21:06:18] <jeffpc> are all systems created using the busted USB image bad?[21:06:38] <jasonpincin> OK, well that answers my issue, thanks guys[21:06:53] <jasonpincin> I have a 3650 M4 going to try next[21:07:31] <rmustacc> jasonpincin: You may be able to manually drop the cisco driver in there.[21:07:40] <ryancnelson> jason, i'm running on a 3650 M4 right now[21:07:41] <jasonpincin> rmustacc: yeah?[21:07:43] <ryancnelson> mazeltov[21:07:54] <jasonpincin> ryancnelson: good to hear[21:08:08] <jasonpincin> ryancnelson: did you flash the raid controller?[21:08:24] <rmustacc> We can't distribute it generally.[21:08:30] <rmustacc> But you may be able to get an s10 gldv3 to work.[21:08:41] <ryancnelson> did not.[21:08:42] <ryancnelson> [root@headnode (evdemo1:0) ~]# prtconf -v | grep -i raid[21:08:42] <ryancnelson> value='RAID controller'[21:08:42] <ryancnelson> value='ServeRAID M5110e'[21:08:42] <ryancnelson> value='scsiclass,00.vIBM.pServeRAID_M5110e.r3_19' + 'scsiclass,00.vIBM.pServeRAID_M5110e' + 'scsiclass,00' + 'scsiclass'[21:09:21] *** viridari has joined #smartos[21:09:21] *** viridari has joined #smartos[21:09:22] <jasonpincin> rmustacc: thanks, will try[21:10:24] *** codingstream has joined #smartos[21:10:25] <rmustacc> The other thing to keep in mind with respect to not seeing the prompt is around the default boot stream.[21:10:35] <rmustacc> We only put the prompt on the console which is determined by grub.[21:10:40] <rmustacc> So by default it's vga/text.[21:10:49] <ryancnelson> oh, and he's using blades[21:10:56] <rmustacc> If you're not on that and are instead on ttya/ttyb you won't see it.[21:11:16] <rmustacc> You should instead boot to ttya or ttyb, ttyb is more commonly the ipmi port for example.[21:11:20] <jasonpincin> yeah I'm java'd in w/ IBM's remote console stuff[21:11:29] <jeffpc> ok, vmadm create -f ... is hung[21:12:05] <jasonpincin> ok, good to know[21:12:14] <MerlinDMC> jeffpc, syntax should be vmadm create << file afaik[21:12:41] <jeffpc> MerlinDMC: running 'vmadm list' in a different terminal shows the new zone being created[21:12:56] *** tdowg1 has quit IRC[21:13:00] <jeffpc> MerlinDMC: smells like vmadm issue to me :/[21:13:29] <MerlinDMC> i can't remember clearly ... but there was smth mit that -f switch ... but could also be around vmadm receive ...[21:14:12] <ryancnelson> i always use -f[21:14:19] <ryancnelson> should still work, i think[21:14:19] *** rcamera has joined #smartos[21:14:28] <jeffpc> 'vmadm delete ....' takes minutes to complete[21:16:28] <jeffpc> yeah, it's vmadm in general[21:18:29] *** mamash has joined #smartos[21:26:06] *** rcamera has quit IRC[21:27:14] *** rcamera has joined #smartos[21:30:09] *** rcamera has quit IRC[21:38:25] *** viridari has quit IRC[21:55:29] *** kschiess has quit IRC[22:01:24] *** jamesd has quit IRC[22:08:46] *** bens1 has joined #smartos[22:13:54] *** codingstream has quit IRC[22:23:16] *** LucidDreamZzZ has quit IRC[22:25:16] *** LucidDreamZzZ has joined #smartos[22:27:19] *** enmand has quit IRC[22:31:54] *** wdent has quit IRC[22:34:58] *** rcamera has joined #smartos[22:41:11] *** jesser has joined #smartos[22:56:16] *** enmand has joined #smartos[23:04:29] *** deirdres has joined #smartos[23:04:54] *** bens1 has quit IRC[23:08:19] *** CarlosC has quit IRC[23:18:09] *** fsteinelX has joined #smartos[23:29:23] *** deirdres_ has joined #smartos[23:31:13] *** izaki_ has quit IRC[23:32:57] *** deirdres has quit IRC[23:32:58] *** deirdres_ is now known as deirdres[23:41:51] *** rcamera1 has joined #smartos[23:42:48] *** rcamera has quit IRC[23:43:22] *** viridari has joined #smartos[23:43:23] *** viridari has joined #smartos[23:44:59] *** haydock_ has quit IRC[23:47:47] *** defa has quit IRC[23:51:14] <jeffpc> thoughts? http://pastie.org/pastes/5486115/text[23:51:28] <jeffpc> vmadm create takes agest[23:51:31] <jeffpc> ages[23:53:26] <richlowe> didn't someone mention that being a silly bug, yesterday?[23:53:36] <richlowe> or was that something else that, also, took ages?[23:54:21] <jeffpc> I thought that was all badness with the recent release[23:54:26] <jeffpc> (this is nov 15)[23:54:41] <jeffpc> (fresh install)