September 30, 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 30, 2013 [00:00:50] *** axonpoet has joined #smartos[00:02:29] *** creativefeather has quit IRC[00:02:52] *** creativefeather has joined #smartos[00:03:29] *** lloydde has joined #smartos[00:10:46] *** jaimef has joined #smartos[00:18:13] *** szaydel has joined #smartos[00:30:20] *** andy_js has quit IRC[00:30:45] <Alasdairrr> On SmartOS in the GZ it doesn't show vnics, on OpenIndiana it does[00:31:29] <Alasdairrr> I have to go to bed now though[00:31:31] * Alasdairrr disappears[00:31:33] *** jaimef has quit IRC[00:47:05] *** axonpoet has quit IRC[00:50:40] *** jesser has quit IRC[01:07:17] *** axonpoet has joined #smartos[01:07:44] *** jaimef has joined #smartos[01:13:14] *** szaydel has quit IRC[01:28:43] *** szaydel has joined #smartos[01:32:35] *** pringlescan1 has joined #smartos[01:34:11] *** pringlescan has quit IRC[01:39:44] *** szaydel has quit IRC[01:40:49] *** pringlescan1 has quit IRC[01:54:51] *** lloydde has quit IRC[01:55:18] *** lloydde has joined #smartos[01:59:26] *** lloydde has quit IRC[02:01:26] *** pringlescan has joined #smartos[02:04:26] *** sjorge has quit IRC[02:09:58] *** sjorge has joined #smartos[02:14:15] *** kimc has quit IRC[02:32:53] *** szaydel has joined #smartos[02:44:06] *** szaydel has quit IRC[02:48:08] *** creativefeather has quit IRC[02:49:42] *** jsm has joined #smartos[02:51:55] *** szaydel has joined #smartos[03:25:51] *** lloydde has joined #smartos[03:30:15] *** lloydde has quit IRC[03:35:46] *** pringlescan has quit IRC[03:44:47] *** lloydde has joined #smartos[03:48:53] *** lloydde has quit IRC[04:00:47] *** lloydde has joined #smartos[04:04:45] *** lloydde_ has joined #smartos[04:05:58] *** lloydde has quit IRC[04:06:34] *** szaydel has quit IRC[04:07:09] *** pringlescan has joined #smartos[05:02:12] *** lloydde_ has quit IRC[05:02:39] *** lloydde has joined #smartos[05:04:16] *** pringlescan has quit IRC[05:05:46] *** lloydde_ has joined #smartos[05:06:26] *** lloydde has quit IRC[05:08:22] *** axonpoet has quit IRC[05:14:58] *** mary5030 has joined #smartos[05:39:23] *** szaydel has joined #smartos[06:07:51] *** mary5030 has quit IRC[06:13:20] *** ImTheBitch has quit IRC[06:17:58] *** zerobsd has quit IRC[06:37:05] *** nefilim has quit IRC[06:40:03] *** wiedi has joined #smartos[06:40:57] *** szaydel has quit IRC[06:43:47] *** noeyesan1icw has joined #smartos[06:43:58] *** ivan\_ has joined #smartos[06:44:35] *** whyz_ has joined #smartos[06:44:35] *** tat_ has joined #smartos[06:44:46] *** disharmo1y has joined #smartos[06:44:47] *** apsonge has joined #smartos[06:44:58] *** tat_ is now known as Guest81353[06:49:40] *** wiedi- has quit IRC[06:49:50] *** ivan\ has quit IRC[06:49:55] *** disharmony has quit IRC[06:49:57] *** tat- has quit IRC[06:49:58] *** whyz has quit IRC[06:49:59] *** asonge has quit IRC[06:49:59] *** noeyesanticw has quit IRC[06:50:00] *** whyz_ is now known as whyz[06:50:01] *** ivan\_ is now known as ivan\[06:53:15] *** mary5030 has joined #smartos[07:06:47] *** lloydde has joined #smartos[07:07:41] *** lloydde_ has quit IRC[07:18:19] *** lloydde has quit IRC[07:18:46] *** lloydde has joined #smartos[07:22:13] *** kmojo has joined #smartos[07:23:08] *** lloydde has quit IRC[07:23:47] <kmojo> Hey guys, I'm looking to restore a clonezilla image (ubuntu server) into a zone... What'd be the best approach? Tried making the ubuntu zone/vm, then booting that from the clonezilla iso. I'm able to get to the 10gb image from an nfs share but when I go to restore it clonezilla asks me to create a partition[07:35:53] *** jsm has quit IRC[07:37:36] *** mary5030 has quit IRC[07:40:49] *** nikolam has quit IRC[07:42:45] *** lloydde has joined #smartos[07:46:59] *** nikolam has joined #smartos[07:47:12] *** lloydde has quit IRC[07:48:32] <nikolam> kmojo, first of all, there is no linux zone alive in illumos (there were linux branded zones that could execute lixux distributiions/binaries without linux kernel, but up to linux kernel calls for version 2.4)[07:48:35] <nikolam> You are actually using KVM VM, that is supposed to be also under the solaris zone.[07:49:05] <nikolam> Then it is the wuestion what you want to do. Do you want to migrate working linux machine to VM or you want to transfer data[07:54:48] <kmojo> I'd like to migrate the linux machine to the VM[07:55:13] <nikolam> One can open zfs sparse volume to write to it as a raw block device (ls -l /devices/pseudo/zfs*raw)[07:55:14] <kmojo> The only artifact of that linux machine I have is a clonezilla partition image[07:55:34] <nikolam> that is for setting permissions[07:56:12] <nikolam> and raw volume is /dev/zvol/rdsk/rpool/path_to_volume/volume[07:56:15] <nikolam> Now, you first have to figure out where is the volume that your new KVM use is stored[07:56:33] <kmojo> aaah okay[07:57:12] *** chakatz has joined #smartos[08:00:49] *** nikolam has quit IRC[08:03:18] *** nikolam has joined #smartos[08:24:25] *** mamash has joined #smartos[08:24:38] <LnxTx> how can i check what use my ram? no kvms and: "Memory: 64G phys mem, 5128M free mem"[08:26:55] <MerlinDMC> maybe the ARC?[08:27:28] *** chakatz has quit IRC[08:29:08] <MerlinDMC> arcstat should tell you the ARC size ... everything else might go into mdb but I'm not sure[08:29:15] *** tonist has joined #smartos[08:30:36] <LnxTx> hmm, echo ::memstat | mdb -k and: "ZFS File Data 12559202 49059 75%"[08:30:47] <LnxTx> huh[08:42:51] *** lloydde has joined #smartos[08:43:46] <LnxTx> "set zfs:zfs_arc_max=4294967296" at startup should work?[08:47:03] *** lloydde has quit IRC[08:47:17] <MerlinDMC> why would you limit arc if you don't use the memory? .. arc shrinks itself if you have memory pressure[08:50:02] <LnxTx> i can't create kvm with 20GB, vmadm said not enough memory[08:50:23] <LnxTx> that's problem[08:51:49] <MerlinDMC> and your platform version is not too old? ... there was some problem that should already be fixed in the latest releases[08:52:09] <LnxTx> build: 20130629T040542Z[08:54:05] <MerlinDMC> LnxTx, this one afaik https://github.com/joyent/illumos-joyent/commit/cf746768a898264fa491791a8252865e5338dc67[08:56:19] <MerlinDMC> LnxTx, the perl command helps as well so you don't need to upgrade: https://www.illumos.org/issues/1974[08:57:54] *** kmojo has quit IRC[08:59:03] *** tonist has quit IRC[09:00:45] <LnxTx> i can upgrade, its pre-production environment[09:02:40] <LnxTx> thanks for help[09:03:12] *** alcir has joined #smartos[09:16:26] *** hadret has joined #smartos[09:20:01] *** texarcana has quit IRC[09:22:06] *** texarcana has joined #smartos[09:26:48] *** alcir has quit IRC[09:27:00] *** tonist has joined #smartos[09:27:11] *** oddsignals has quit IRC[09:29:07] *** chakatz has joined #smartos[09:35:47] *** sparq_ has joined #smartos[09:43:56] *** Ymarg has quit IRC[09:44:27] *** Ymarg has joined #smartos[09:47:09] *** sa5bke has quit IRC[10:00:16] *** oddsignals has joined #smartos[10:18:01] *** andy_js has joined #smartos[10:18:15] *** alcir has joined #smartos[10:20:03] *** leecallen35 has joined #smartos[10:23:02] *** leecallen has quit IRC[10:25:50] *** yruss972 has joined #smartos[10:27:42] *** yruss972 has quit IRC[10:32:55] *** yruss972 has joined #smartos[10:37:55] *** sparq_ has quit IRC[10:41:58] *** andy_js_ has joined #smartos[10:45:00] *** andy_js has quit IRC[10:45:04] *** andy_js_ is now known as andy_js[10:45:04] <mburns> Anyone run an ids like aide or tripwire?[10:53:12] <MerlinDMC> jperkin, does pkgsrc something special if you build ruby gems? gems that are build using `gem install` or bundler seem to be platform version dependent[10:56:21] <jperkin> MerlinDMC: in what way are they dependent?[10:57:50] <MerlinDMC> I did build up an image with mibe including gitlab and some gems get linked against /usr/lib - If I provision that image on an older platform (201305somewhat) those gems fail to load[10:58:23] <jperkin> depends which libraries, there are some known cases of pulling in libxml/libbz2 etc from the platform which may be version-dependent[10:58:30] <MerlinDMC> on 3 different versions of the platform from last month and this month it works fine tho[10:59:04] <jperkin> those are bugs in pkgsrc where we aren't pulling in the pkgsrc versions correctly, but we certainly don't (or it'd be a big bug if we do) specifically build against platform things which are version-specific.[10:59:37] <MerlinDMC> kgio (binding to libsocket.so.1) was one that was failing - can't remember the rest and I did kill the vmware machine but that can be reproduced[11:00:05] <jperkin> well, libsocket is fine[11:00:55] <jperkin> you're not going to get away from linking against core libraries, it's the third-party ones in the platform which are the problems.[11:01:12] <jperkin> anyway, happy to take a look at a specific bug report[11:01:19] <MerlinDMC> i'll spin up vmware during lunch to reproduce this and give you the actual output[11:01:23] <jperkin> k[11:02:45] <MerlinDMC> was hoping for some voodoo in pkgsrc that would help me there ;)[11:04:28] <jperkin> there is some gem handling but I don't know how extensive it is[11:04:39] <jperkin> just copy/paste some existing packages[11:05:51] *** chakatz has quit IRC[11:07:04] *** khushildep has joined #smartos[11:12:03] *** yruss972 has quit IRC[11:12:06] *** jihyun has quit IRC[11:15:11] *** KermitTheFragger has joined #smartos[11:16:28] <khushildep> guys is there a list of what dtrace probes and providers work inside NGZ and only inside GZ please?[11:18:21] <MerlinDMC> dtrace -l and make a diff?[11:20:36] *** jihyun has joined #smartos[11:31:13] *** nikolam has quit IRC[11:43:53] *** lloydde has joined #smartos[11:45:33] *** alcir has quit IRC[11:48:42] *** lloydde has quit IRC[11:56:12] <MerlinDMC> jperkin, https://gist.github.com/MerlinDMC/bed69704a5439c891494[12:00:44] <mamash> merlindmc: all those libs linked in are valid[12:01:12] <mamash> those are standard SunOS libs typically needed[12:04:47] <MerlinDMC> hm ... the "version not found" thing is irrelevant then?[12:05:15] <MerlinDMC> I wonder why that thing does run on newer platform like expected without any problem -.-[12:07:56] *** marsell has joined #smartos[12:07:59] <mamash> what kind of platform difference are we talking about here?[12:08:41] <jperkin> MerlinDMC: ok, odd, I'll need to dig in a little deeper, or wait for robert or keith to wake up[12:09:15] <jperkin> it's possible it got updated, but it would be very annoying if it did[12:09:40] <jperkin> sun decided to bump libnsl during the sol10 cycle which forced us to build on a particular update release to ensure generic binaries[12:09:41] <MerlinDMC> the image was build on smartos 20130613T201213Z ... does run well on 201309xx does fail on the 20130405[12:10:20] *** yruss972 has joined #smartos[12:10:49] <jperkin> there are occasionally ways to work around it too, i.e. by avoiding symbols which require a particular version[12:11:58] *** bens1 has joined #smartos[12:12:10] <MerlinDMC> your ruby193-kgio package seems to work fine on the old box ... tho my assumption that pkgsrc may do smth different[12:12:36] <jperkin> it's likely that there's some socket interface that is auto-detected by the build, and as we build on older platforms we don't build-in support in our packages[12:13:08] <MerlinDMC> but would also be annoying to have some ruby app running for months and after a platform upgrade everything breaks ;)[12:13:19] <MerlinDMC> what is your current build platform version?[12:13:41] <jperkin> it's SDC so it doesn't map exactly, but it's based on SDC 6.5 so is pretty old at the moment[12:13:59] <jperkin> something like 1yr+ behind on illumos trunk, plus any important patches[12:15:35] <MerlinDMC> hm ... i'll rebuild the image on some really old platform and see how that changes things ... mibe should wok there so it should be pretty simple[12:15:44] <jperkin> my guess is that it's this: https://github.com/joyent/illumos-joyent/commit/5dbfd19ad5fcc2b779f40f80fa05c1bd28fd0b4e[12:16:01] <jperkin> accept4() requires ILLUMOS_0.1[12:16:26] <jperkin> which is what your binaries are missing, so try building with a flag which explicitly disables accept4 support from the OS[12:16:34] <jperkin> the app will likely have some compat workaround[12:16:49] <MerlinDMC> hm ... that seems to be a good point[12:17:07] <MerlinDMC> i'll look into that - thx[12:17:18] <Alasdairrr> I hit this the other day, building stuff on a newer SmartOS doesn't run on an older one[12:17:25] <Alasdairrr> if it uses something new[12:18:16] *** axonpoet has joined #smartos[12:20:01] *** CaTtleyA has quit IRC[12:20:42] <jperkin> yeh, I'm not really a fan of this way of doing things, would rather new interfaces were only introduced in a new major lib version - but I'm certainly no expert on these things and there are likely many issues with that way too[12:21:53] *** CaTtleyA has joined #smartos[12:22:58] *** tonist has quit IRC[12:39:31] *** aortmann has joined #smartos[12:40:14] *** alcir has joined #smartos[12:41:47] <aortmann> hi, anyone can say me how to update the resolver for a local zone? I did it in the local zone in /etc/resolv.conf but after a restart of the local zone the old resolver is in the config file.[12:42:24] <jzu_> you need to update it to JSON I guess[12:43:09] <MerlinDMC> jperkin, seems to really be the accept4() ... and it's only kgio what's failing hopefully you'll maintain the needed version in pkgsrc for long enough :) ... that's another pizza/beer I owe you[12:43:16] <Alasdairrr> jperkin: is there a way to just mask all the ILLUMOS_0.1 symbols?[12:43:59] *** Fuzai has joined #smartos[12:44:11] <jperkin> Alasdairrr: I don't know, it's certainly be interesting if that's possible[12:44:14] *** lloydde has joined #smartos[12:44:18] *** yruss972 has quit IRC[12:45:08] *** alcir has quit IRC[12:46:34] *** Fuzai has quit IRC[12:48:14] *** lloydde has quit IRC[12:48:25] <aortmann> I created the local zone out of a json file but where is the json saved after I created it?[12:50:24] *** alcir has joined #smartos[12:50:31] <oddsignals> aortmann: you can get the JSON object for a zone with 'vmadm get <uuid>'[12:51:43] <oddsignals> aortmann: I don't think the exact JSON you used to create it is stored anywhere unless you save it yourself - I keep mine in /opt/vmspecs[12:53:37] <MerlinDMC> aortmann, is that a kvm zone or a os zone?[12:54:55] <oddsignals> jperkin: it took a while, but I've finally got a self-hosting 64-bit GHC[12:55:30] <oddsignals> unfortunately it has a lot more failing tests than the 32-bit one[12:55:59] <jperkin> oddsignals: great! I'll take it ;)[12:58:24] *** alcir has quit IRC[12:58:41] <aortmann> its a os zone[12:58:58] <aortmann> I will try now to delete the resolver attribute with zonecfg and after add a new one[13:00:45] <MerlinDMC> aortmann, it should be updateable via vmadm update ... but I always get exeptions here - so I do smth. wrong or it's not working as expected[13:01:18] <aortmann> I dont geht exceptions with update but when I after get the informations from the zone its still the old resolver[13:01:39] <oddsignals> jperkin: a lot of the failing tests seem to be debugger-related - hopefully something minor[13:01:48] <MerlinDMC> aortmann, even after a reboot?[13:02:17] <aortmann> ehh we will see one sec[13:02:47] <aortmann> yeah still old settings[13:15:22] *** yruss972 has joined #smartos[13:21:02] *** pringlescan has joined #smartos[13:22:49] *** pringlescan has quit IRC[13:23:39] *** yruss972 has quit IRC[13:26:41] *** pringlescan has joined #smartos[13:30:17] <khushildep> is three an iosnoop.d which can run in an NGZ at all?[13:32:35] <opeth__> aortmann: what is your method? I'd do 'vmadm update <uuid> resolvers=["foo","bar"]' and afterwards check what I've done with 'vmadm get <uuid> | json resolvers'[13:34:03] *** szaydel has joined #smartos[13:36:05] *** yruss972 has joined #smartos[13:37:45] <opeth__> aortmann: it breaks that way, rather do:[13:37:47] <opeth__> [root@global ~]# cat /tmp/res.json[13:38:01] <opeth__> {"resolvers": [ "foo", "bar" ]}[13:38:25] <opeth__> put that in a file and then do vmadm update <uuid> < /path/to/your/file[13:38:46] <opeth__> the verification method I wrote is still valid[13:39:37] <opeth__> I've just verified this and it's working[13:44:35] *** lloydde has joined #smartos[13:48:53] *** lloydde has quit IRC[13:51:02] *** pinja has quit IRC[14:00:31] *** Guest81353 is now known as tat-[14:05:48] *** WormDrink has joined #smartos[14:08:26] *** tonist has joined #smartos[14:33:09] *** hadret has quit IRC[14:33:37] *** pringlescan has quit IRC[14:35:07] *** pringlescan has joined #smartos[14:39:47] *** ira has joined #smartos[14:43:18] *** jsm has joined #smartos[14:44:56] *** lloydde has joined #smartos[14:45:57] *** yruss972 has quit IRC[14:49:11] *** lloydde has quit IRC[14:51:46] *** yruss972 has joined #smartos[14:55:36] *** nikolam has joined #smartos[14:56:29] *** alcir has joined #smartos[15:03:05] *** nikolam has quit IRC[15:06:22] *** robinsmidsrod has quit IRC[15:07:37] *** robinsmidsrod has joined #smartos[15:15:31] *** jwk404 has quit IRC[15:23:37] *** volci has joined #smartos[15:23:56] *** potatosalad has joined #smartos[15:25:01] *** yruss972 has quit IRC[15:27:17] *** nikolam has joined #smartos[15:33:19] *** aortmann has quit IRC[15:37:19] *** jwk404 has joined #smartos[15:37:26] *** nikolam has joined #smartos[15:45:17] *** lloydde has joined #smartos[15:46:57] *** zerobsd has joined #smartos[15:49:18] *** lloydde has quit IRC[15:53:40] *** alcir has quit IRC[15:55:31] *** neophenix has joined #smartos[15:56:04] *** yruss972 has joined #smartos[15:57:45] *** pringlescan has quit IRC[16:04:23] *** yruss972 has quit IRC[16:07:23] *** ira has quit IRC[16:08:52] *** mary5030 has joined #smartos[16:12:16] *** alcir has joined #smartos[16:16:33] *** axonpoet has quit IRC[16:18:23] *** pringlescan has joined #smartos[16:21:52] *** pringlescan1 has joined #smartos[16:22:37] *** pringlescan has quit IRC[16:22:37] *** aortmann has joined #smartos[16:23:00] *** alcir has quit IRC[16:27:28] *** yruss972 has joined #smartos[16:27:42] *** nefilim has joined #smartos[16:28:50] *** pringlescan1 has quit IRC[16:28:55] *** daleg has joined #smartos[16:33:31] *** nikolam has quit IRC[16:35:58] *** trentster has quit IRC[16:40:33] *** axonpoet has joined #smartos[16:42:35] *** lloydde has joined #smartos[16:47:55] *** lloydde has quit IRC[16:57:29] *** orangeroo has joined #smartos[16:58:10] <rmustacc> jperkin, MerlinDMC, Alasdairrr: Basically most applications assume that the host you're building on is what you want to deploy on. So you should really be building software for distribution on the oldest release that you want to support.[16:58:44] <rmustacc> jperkin: We are actually versioning those libraries. If you just changed the soname every time you added new symbols it'd be a compatability nightmare. Instead what we're doing is making sure that every public symbol is in a version. Those versions can never change.[16:58:51] *** demona has left #smartos[16:58:53] <rmustacc> Either a library has that version or it doesn't so they'll fail hard and fast.[16:59:13] <xenol> does joyent stores somewhere old releases of smartos? I am running older version in production and I can't find it on the mirror.[16:59:17] <jperkin> yeh, thinking it through a bit further I came to that conclusion[16:59:38] <jperkin> however, it'd appear to be a useful feature to be able to build against a specific version somehow[17:00:16] <jperkin> xenol: https://download.joyent.com/pub/iso/[17:00:46] <xenol> jperkin: nice, thanks. I was looking manta links and didn't find that. Thanks once afain[17:01:04] *** ImTheBitch has joined #smartos[17:01:13] *** chorrell has joined #smartos[17:02:01] <rmustacc> jperkin: The battle there is more often the build system.[17:02:06] *** aortmann has quit IRC[17:02:27] <rmustacc> Because it's not always just symbols in libraries but also header values which many apps use as a way of saying I should use this.[17:05:04] <rmustacc> e^ipi: Would you mind syncing all the old releases into manta?[17:05:18] *** lloydde has joined #smartos[17:07:07] *** tonist has quit IRC[17:18:09] *** pringlescan has joined #smartos[17:24:17] <MerlinDMC> rmustacc, thx for the info ... i'll use a VMware box for building future images maybe ... or i'll just add an requirement of version x to the manifest[17:25:03] *** mamash has left #smartos[17:25:11] <rmustacc> We try to ensure that all software built in the past works in the future.[17:25:15] <rmustacc> It's a bit hard to go the other way around.[17:25:26] <rmustacc> If I had a time machine, I wouldn't be using it for SmartOS. ;)[17:25:57] <MerlinDMC> ye ... but the upgrade problem is non existent then (probably) and that's more important[17:26:16] <rmustacc> At least as far as the platform is concerned, yes.[17:30:46] *** pringlescan has quit IRC[17:33:22] *** nefilim has quit IRC[17:34:32] *** wiedi has quit IRC[17:34:34] *** pringlescan has joined #smartos[17:42:20] <e^ipi> rmustacc: yes. modulo there's no more msync[17:43:07] <rmustacc> Yeah, it sh[17:43:13] <rmustacc> *Yeah I know. It should just be a one time thing though.[17:43:32] *** pringlescan has quit IRC[17:43:42] *** bens1 has quit IRC[17:46:06] *** lloydde has quit IRC[17:49:33] *** tonist has joined #smartos[17:54:13] *** WormDrink has quit IRC[17:59:33] *** sjorge has quit IRC[18:00:01] *** CaTtleyA has quit IRC[18:00:23] *** sjorge has joined #smartos[18:00:32] *** CaTtleyA has joined #smartos[18:02:24] *** sjorge has quit IRC[18:07:54] *** WormDrink has joined #smartos[18:12:48] *** mamash has joined #smartos[18:15:25] *** yruss972 has quit IRC[18:16:00] *** yruss972 has joined #smartos[18:31:55] *** yruss972 has quit IRC[18:33:24] *** yruss972 has joined #smartos[18:33:27] *** alfredodeza has joined #smartos[18:34:06] <alfredodeza> I am looking to change the MTU of one of the interfaces, would /usbkey/config be the right place to do that?[18:35:16] *** ryancnelson has joined #smartos[18:35:59] <rmustacc> alfredodeza: Changing the MTU isn't currently supported, but if it was, that'd be the right place.[18:36:20] <alfredodeza> thanks rmustacc[18:36:43] <orangeroo> I'be brought up the topic before. I haven't found a way to do it.[18:36:55] <alfredodeza> I am trying to make some network changes as I am seeing poor performance on a host[18:37:30] <alfredodeza> went from 100MB/s reads to just about 40MB/s[18:37:34] <orangeroo> In Solaris you had to edit the nic's driver config.[18:37:38] <alfredodeza> right[18:38:15] <ryancnelson> changing mtu on a global zone's nic, or a vm's??[18:38:32] <orangeroo> For me, the global zone.[18:38:38] <alfredodeza> ryancnelson: global[18:39:08] <alfredodeza> would changing it in an OS make any difference if the global doesn't change?[18:39:34] <alfredodeza> for example, if the global is at 1500, wouldn't that restrict regardless of what a VM sets?[18:39:42] <ryancnelson> if you make a custom SMF service that makes your adjustment, you can run that shortly after boot[18:39:50] <ryancnelson> using the /opt/custom/smf mechanism[18:39:52] <ryancnelson> http://www.psychicfriends.net/blog/archives/2012/03/21/smartosorg_run_things_at_boot.html[18:40:32] <ryancnelson> just change my example that copies in ssh key to one that sets your mtu to 1460 (or whatever)[18:40:54] <alfredodeza> got it[18:41:34] <ryancnelson> i've seen people who had to do exactly that because of mtu crazyness on bnx drivers specifically[18:41:37] <orangeroo> But how can you edit something like /kernel/drv/igb.conf and get that to stick? I don't think smf will hellp there.[18:41:55] <ryancnelson> no, you can't, without building the platform[18:42:09] <orangeroo> Yeah.[18:42:10] <ryancnelson> smf will let you run[18:42:10] <ryancnelson> ifconfig igb0 mtu 1460[18:42:10] <alfredodeza> oh, I think that /kernel/ stuff requires you to boot[18:42:37] *** lloydde has joined #smartos[18:43:04] <orangeroo> Yup[18:43:24] <orangeroo> Or reinit the device if possible[18:44:11] <orangeroo> It's possible, but not trivial to enable jumbo frames.[18:45:09] <alfredodeza> heh, I am finding that myself :([18:45:10] <alfredodeza> ifconfig: setifmtu: SIOCSLIFMTU: e1000g0: Invalid argument[18:45:26] <orangeroo> I would think it would make sense to enable jumbo frames in the platform and then configure it down to MTU 1500 in SMF on boot for the masses.[18:45:27] *** WormDrink has quit IRC[18:46:15] *** ryancnelson has left #smartos[18:46:18] *** sjorge has joined #smartos[19:04:11] *** andoriyu has joined #smartos[19:05:05] *** khushildep_ has joined #smartos[19:07:08] *** wolstena has joined #smartos[19:13:44] *** khushildep_ has quit IRC[19:18:54] *** tmiller has joined #smartos[19:19:46] <leecallen35> nahamu: you asked me to remind you - please review http://wiki.smartos.org/pages/viewpage.action?pageId=2687032[19:20:10] <leecallen35> nahamu: (and please tell me how I can exchange that ugly URL for a nice one)[19:21:20] *** mamash has left #smartos[19:27:19] *** khushildep has quit IRC[19:29:35] *** Alek has joined #smartos[19:30:39] *** yruss972 has quit IRC[19:31:28] *** orangeroo has left #smartos[19:39:58] <nahamu> leecallen35: thanks for the reminder. I think a simpler page title will allow for a friendly URL. the '?' might be the culprit.[19:40:51] <nahamu> leecallen35: is it all right if I just do something of a brain dump onto the bottom of the page for you to rearrange?[19:50:29] *** potatosalad has quit IRC[20:00:05] <leecallen35> sure![20:05:28] *** nikolam has joined #smartos[20:08:38] *** lloydde has quit IRC[20:09:12] *** lloydde has joined #smartos[20:11:13] *** lloydde_ has joined #smartos[20:11:23] *** lloydde has quit IRC[20:11:42] *** gotwf has joined #smartos[20:16:28] <gotwf> Hey folks, is there any SmartOS/IllumOS hardware compatibility info available other this what is here http://illumos.org/hcl/[20:16:58] <wesolows> what else would you like? There's a "works for me" list of BOMs on smartos.org.[20:16:59] <gotwf> I am curious specifically about E5-2600v2 Xeons[20:17:11] <wesolows> There's Joyent's BOMs at https://github.com/joyent/manufacturing[20:17:19] <wesolows> But those aren't HCLs, they're just specific configurations.[20:17:35] <wesolows> As for Ivy Bridge, we have one running in our lab.[20:18:06] <gotwf> Cool[20:19:17] <gotwf> I have a supermicro box with an X9DAi board[20:19:27] *** bens1 has joined #smartos[20:20:06] <gotwf> that I never got around to fleshing out. Now the 2600v2's are upon us, I just wanted to make sure that there weren't any issues.[20:21:10] <wesolows> I'd expect them to work.[20:21:20] <gotwf> perchance anyone have experience with this board?[20:21:34] <wesolows> we don't use the DAi; we use DRD.[20:21:42] <gotwf> http://www.supermicro.com/products/motherboard/Xeon/C600/X9DAi.cfm[20:22:23] <wesolows> No support for the SCU. Otherwise probably ok.[20:25:55] *** bens1 has quit IRC[20:26:14] *** ImTheBitch has quit IRC[20:27:10] *** ImTheBitch has joined #smartos[20:28:37] *** sparq_ has joined #smartos[20:29:53] *** potatosalad has joined #smartos[20:31:08] <gotwf> wesolows: Thanks.[20:35:43] *** sjorge has quit IRC[20:51:41] *** pringlescan has joined #smartos[20:51:46] *** ryancnelson has joined #smartos[20:53:26] <yo61> Anyone know the default root password for the CentOS image? Or know how to find out what it is?[20:54:18] <rmustacc> I guess it's not in metadata?[20:55:18] <ryancnelson> there isn't a root password[20:56:44] <ryancnelson> you'd set one, via metadata, or log into the vnc console, where you can log in with no password (for most images)[20:57:24] <yo61> Ah, OK[20:57:42] <ryancnelson> see "passing ssh keys to the vm" here:[20:57:42] <ryancnelson> http://wiki.smartos.org/display/DOC/How+to+create+a+zone+(+OS+virtualized+machine+)+in+SmartOS[20:58:03] <ryancnelson> sorry, actually, here: http://wiki.smartos.org/display/DOC/How+to+create+a+KVM+VM+%28+Hypervisor+virtualized+machine+%29+in+SmartOS[20:58:31] <ryancnelson> ### comment: add this to the faq[21:00:15] *** mamash has joined #smartos[21:01:21] <yo61> sshkey worked, thanks[21:01:54] <ryancnelson> woot![21:02:08] <yo61> Interesting - you deploy a 3.9.x kernel on the CentOS Image[21:02:27] <ryancnelson> technically, *you* did[21:02:32] <yo61> :p[21:03:04] <yo61> Actually I deployed the image - somonese *else* deployed the later kernel and took a snapshot[21:03:30] <yo61> :)[21:08:12] *** lloydde_ has quit IRC[21:08:47] *** lloydde has joined #smartos[21:13:03] *** lloydde has quit IRC[21:14:24] *** papasmurf has joined #smartos[21:14:31] *** volci has quit IRC[21:18:08] <yo61> Oooh, there's magic going on[21:18:34] <yo61> I added a new disk to the CentOS vm and it was already partitioned, and mounted as /data[21:18:44] *** nikolam has quit IRC[21:20:01] <yo61> That's presumably something in the image start up scripts?[21:22:33] *** pringlescan has quit IRC[21:27:59] <yo61> Ugh, no dewvice mapper in the CentOS machine image?[21:28:41] <yo61> uname -a[21:28:43] <yo61> Oops[21:29:30] <yo61> Who provides these images?[21:35:30] *** mamash has left #smartos[21:36:49] <leecallen35> yo61: are you aware you can create an 'empty' KVM VM and install CentOS into it?[21:36:56] <yo61> I am, yes[21:37:01] <yo61> I usually do[21:37:06] <leecallen35> just making sure[21:37:12] <yo61> I thought - no, I'll use an image, it's quicker[21:39:14] <yo61> Looks like there's all sorts of skull-duggery going on in the smartos image[21:39:37] <yo61> I shall roll with it for now[21:50:43] <alfredodeza> I need to re-do the ZFS filesystem with brand new disks and wanted to know if there is a guide on how to backup (and later import back) virtual machines into SmartOS[21:50:59] <alfredodeza> is it just a matter of copying the directories ?[21:57:46] <ryancnelson> yo61: the "magic" is in /etc/rc.local, at the end[21:57:55] <yo61> Yes - I found it[21:57:56] <ryancnelson> that calls some stuff that's in /lib/smartdc , i believe[21:58:00] <ryancnelson> formats the disk[21:58:09] <ryancnelson> pulls in the keys, if there's metadata set[21:59:18] *** julianduque has joined #smartos[22:00:20] <yo61> Can I make it mount in a different place by default?[22:01:00] <ryancnelson> make what mount who?[22:01:09] <yo61> 2nd disk, mount to /opt instead of /data ?[22:01:37] <ryancnelson> you'd need to modify the image, then[22:01:45] <ryancnelson> you can make it do whatever you want[22:02:02] <yo61> Yeah, I know. Was wondering if there was any metadata that might do the trick[22:02:04] *** theCzar has joined #smartos[22:02:08] <ryancnelson> there isn't[22:02:56] <ryancnelson> other than "user-script" metadata[22:03:06] *** nikolam has joined #smartos[22:03:13] <ryancnelson> which is straight-up exectutable script[22:03:49] <ryancnelson> so write a script that sleeps 120 seconds, then umounts /data and re-mounts it at /opt[22:04:18] <ryancnelson> example stuff you can do with user-script in a vm:[22:04:21] <ryancnelson> metadata user-script="/usr/bin/apt-get -y update ; /usr/bin/apt-get -y install git-core"[22:05:19] <yo61> An, nice.[22:05:21] *** WormDrink has joined #smartos[22:05:22] <yo61> *Ah[22:08:23] <ryancnelson> the user-script only works because of other in-image "magic", though. just like there's something in the image that looks for a secondary disk and formats it, there's something that pulls your ssh-keys from metadata, and there's something that pulls your "user-script" from metadata and executes it.[22:08:52] <ryancnelson> if you make a custom image that looks for your own specially-named metadata keys and does something with it, then you've added a feature[22:09:10] <ryancnelson> ... on the back of the very-flexible metadata mechanism[22:11:00] <opeth__> I know that's a fairly stupid question but lacking any comparisons could someone tell me how long it usually takes an average linux kvm guest to boot, from 'the cold' to runlevel 3?[22:11:09] <opeth__> on your average boxen naturally[22:14:05] *** lloydde has joined #smartos[22:14:06] <ryancnelson> there's no such thing as an average linux box[22:14:54] *** mamash has joined #smartos[22:15:02] <rmustacc> opeth__: Generally I see it go fairly fast. So assuming not a lot of extra stuff, I'd expect < 1m[22:15:02] <ryancnelson> somewhere between 7 seconds and 3 minutes?[22:15:25] <rmustacc> But it can depend on a fair amount of factors.[22:15:53] <xenol> I have openvpn SMF manifest, which I am able to import on OpenIndiana, but it does not get imported in SmartOS zone. However, svccfg -v import states that it was imported successfully.[22:16:08] <xenol> manifest is working and can be found here https://raw.github.com/OpenIndiana/oi-userland/oi/hipster/components/openvpn/files/openvpn.xml[22:16:20] <xenol> (I modified it a bit)[22:16:41] <xenol> All I changed was paths to binaries.[22:18:23] <rmustacc> xenol: Nothing related to it in svcs -a then?[22:19:51] <xenol> rmustacc: http://pastebin.com/auwCQxZw nope[22:20:42] *** leecallen has joined #smartos[22:20:42] *** leecallen35 has quit IRC[22:20:44] <ryancnelson> grep -i ?[22:21:14] <ryancnelson> <common_name>[22:21:14] <ryancnelson> <loctext xml:lang="C">OpenVPN</loctext>[22:21:14] <ryancnelson> </common_name>[22:21:22] <ryancnelson> capitols matter[22:21:48] <xenol> FMRI is not capitalized, so that doesn't matter[22:22:39] <xenol> it doesn't seem to work in the GZ either. Not sure if it is related to manifest or not.[22:22:51] *** bens1 has joined #smartos[22:23:10] <rmustacc> I would suspect part of the problem here is that it appears you have no instances for that service.[22:24:35] <xenol> doh, right.[22:25:15] <opeth__> rmustacc, ryancnelson: thank you; in that case my rig does a fair job of booting my gentoo guest in seven seconds total[22:25:24] <xenol> but weird, it worked on OI. I guess I had a previosly imported manifest[22:25:27] <xenol> anyways, thanks rmustacc.[22:27:28] <rmustacc> xenol: I bet if you actually go into svccfg you'll find that fmri.[22:27:53] <xenol> rmustacc: yup[22:28:10] <xenol> though it's a bug, there needs to be a default instance[22:33:10] <alfredodeza> I am about to re-do this host and would love if someone can point me to how to restore these VM's after getting new disks for the global zone[22:33:15] <alfredodeza> is it a matter of just copying everything back?[22:33:33] <alfredodeza> e.g. manifests and the UUID directories containing the virtual machines?[22:33:53] <alfredodeza> I could not find a section in the wiki that talked about how to re-add virtual machines[22:36:03] <ryancnelson> your vm's are in zfs filesystems, not files[22:36:44] <ryancnelson> if you zfs snapshot (recursively) the zones zpool, and save it someplace, and then restore it, you'll have restored the entire system.[22:36:58] <ryancnelson> that's exactly why live-boot (from usb) is awesome.[22:37:11] <alfredodeza> so I would need to backup the snapshot[22:37:21] <ryancnelson> with zfs send[22:37:24] <alfredodeza> right[22:37:58] <alfredodeza> is that the only way of doing it?[22:38:02] <ryancnelson> or you can add your new disks to the zpool, as mirrors, let it mirror, then remove the old disks, and not even have to reboot[22:38:34] <ryancnelson> if you have the disk-slots available[22:38:50] <alfredodeza> I do have the slots available[22:39:34] <alfredodeza> do you mean to extend the current mirror or as a separate one? I currently have 2 mirrors with 2 drives each[22:39:56] <ryancnelson> your vms' data isn't stored in any file anywhere. it's a zfs zvol. you'll need to back up that "device", which is done with zfs snapshot , send, and receive, so you can't just copy files around, no.[22:40:10] <alfredodeza> roger[22:40:16] <ryancnelson> i mean add a device to the existing mirror[22:40:24] *** KermitTheFragger has quit IRC[22:40:33] <ryancnelson> so, for you, a 3rd mirror[22:40:57] <ryancnelson> i'd still back it up, for safety, though[22:41:02] <alfredodeza> sure[22:41:25] <alfredodeza> I would love to have a way to go from 2 mirrors (x2 disks each) to 1 mirror and 4 disks[22:41:39] <alfredodeza> but I don't see any other way to do this other than redoing the whole filesystem[22:56:50] *** whyz has quit IRC[23:00:49] *** bens1 has quit IRC[23:03:30] *** whyz has joined #smartos[23:07:32] *** deirdres_ has joined #smartos[23:08:04] *** mamash has left #smartos[23:08:29] <deirdres_> hey folks, in case you've managed to miss my PR on multiple other channels, there's training coming up from Joyent that may be of interest: http://www.joyent.com/developers/training-services/upcoming-courses[23:09:01] *** maxb` has joined #smartos[23:09:02] *** papasmurf has quit IRC[23:20:23] *** sjorge has joined #smartos[23:20:48] *** bsdguru has joined #smartos[23:23:01] *** tonist has quit IRC[23:23:57] *** volci has joined #smartos[23:25:18] <nikolam> deirdres_, uh, expensive. unless company pays for it ;p[23:25:37] <deirdres_> that's generally what happens, yes[23:25:50] <deirdres_> this is 3 days (or 5) in the classroom with Max or Brendan[23:25:59] <deirdres_> very intensive, maximum 12 students per class[23:26:10] <yo61> Anyone help me out getting a kvm instance to pxe boot?[23:26:12] <nikolam> I like your videos to much, to avoid seeing your post ;p[23:29:29] <deirdres_> thanks! always nice to know they're useful / popular[23:30:09] <nikolam> deirdres_, My favorite is video with advices, how to actually film community videos ;p[23:31:12] *** ira has joined #smartos[23:31:18] <deirdres_> thanks! a bunch of those are linked from here: http://www.beginningwithi.com/2010/03/09/the-videoblogging-manual/[23:32:12] <yo61> Oh ho ho ho, I see - order=cn,once=n doesn't seem to work with vmadm create but it does with vmadm start[23:32:59] <alfredodeza> maybe I am misunderstanding mirroring here, but what is the difference between 3 2-way mirrors and 1 6-way mirrors?[23:33:38] <alfredodeza> I thought that 1 6-way mirrors was faster and could lose up to 3 drives[23:34:04] <alfredodeza> but would be in trouble if I lost an entire mirror (2 of their drives going bad)[23:36:14] <alfredodeza> the docs do mention this: " Going to a 4-way (or greater) mirror may offer only marginal improvements in data protection. Concentrate on other methods of data protection if a 3-way mirror is insufficient. "[23:36:37] <alfredodeza> so maybe my assumption that having 6 drives in a single mirror is better than 3 mirrors with 2 drives is wrong?[23:37:49] <rmustacc> alfredodeza: A 6-way mirror is putting the same data on all 6 drives.[23:38:02] <rmustacc> If you want any three drives to fail you want RAIDz3[23:39:47] <alfredodeza> I thought that for RAIDz3 you wanted at least 9 drives[23:40:22] * alfredodeza could very well be wrong[23:42:04] <alfredodeza> rmustacc: is my assumption correct, that in a 3 2-way mirror setup, if one of the mirrors fails the filesystem is unrecoverable ?[23:44:04] <rmustacc> Yes.[23:44:14] *** creativefeather has joined #smartos[23:44:22] <rmustacc> You were discussing your failure criteria, that's why I mentioned RAIDz3.[23:44:32] <rmustacc> If you want any three drives to fail and have your data be intact, that is the only way.[23:44:45] <alfredodeza> thank you rmustacc[23:46:16] *** creativefeather has quit IRC[23:46:18] *** creativefeather1 has joined #smartos[23:46:51] *** axonpoet has quit IRC[23:48:40] *** theCzar has quit IRC[23:54:02] *** theCzar has joined #smartos[23:54:15] *** jwk404 has quit IRC