Switch to DuckDuckGo Search
   March 18, 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 | 31 | >

bottom
[00:03:42] *** xmerlin has quit IRC
[00:05:55] *** szaydel has joined #smartos
[00:17:58] *** xmerlin has joined #smartos
[00:22:55] *** bluezenix has joined #smartos
[00:26:41] *** wolfeidau has quit IRC
[00:26:58] *** wolfeidau has joined #smartos
[00:34:25] *** darjeeling has quit IRC
[00:40:04] *** Cpt_Oblivious has quit IRC
[00:40:39] *** dysinger has quit IRC
[00:40:58] *** dysinger has joined #smartos
[00:49:37] *** des2 has quit IRC
[00:54:01] *** alpharender has quit IRC
[00:54:53] *** bluezenix has quit IRC
[00:57:56] *** des2 has joined #smartos
[01:11:20] *** Lumb has quit IRC
[01:15:28] *** Lumb has joined #smartos
[01:27:15] *** AlainODea has joined #smartos
[01:27:18] *** denizr has joined #smartos
[01:28:01] <AlainODea> Has anyone tried the Joyent Ops Fat Client (Chef) on a recent build of SmartOS?
[01:34:08] *** axonpoet has quit IRC
[01:53:22] *** szaydel has quit IRC
[01:59:00] <LeftWing> AlainODea: I've only used it inside JPC zones.  Are you experiencing difficulty?
[02:03:18] *** vsomes has quit IRC
[02:07:26] <AlainODea> LeftWing: I'm not able to pull down http://8.12.35.49/smartos/Chef-fatclient-SmartOS-10.14.2.tar.bz2 which is Ben's client.  I can run the bootstrap, but it dies with a bzip2 error within gtar after curl fails to download that
[02:08:06] <AlainODea> I'm wondering if it's part of the chef-server image.
[02:08:35] <LeftWing> That looks like a JPC zone IP.  I'll see if it's still alive. :P
[02:10:32] *** tonyarkles has joined #smartos
[02:12:15] <LeftWing> AlainODea: Yeah, I have no idea what that used to be.  I would ping benr about it.
[02:12:46] <AlainODea> LeftWing: I'll do that.  Thank you :)
[02:24:18] <_lb_> Guys... How do I get a full copy of a vm? I am trying zfs snapshot but the result is too small to be the full image...
[02:24:24] <_lb_> I know a snapshot is a differential...
[02:24:38] <_lb_> But is it possible to make a full copy I can move around for a dataset?
[02:26:18] <jesse_> use the snapshot?
[02:26:29] <jesse_> it only uses the differential of space
[02:26:42] <jesse_> but it is the full contents
[02:27:00] <_lb_> Then I don't know what I am doing wrong...
[02:27:12] <jesse_> neither do I=)
[02:27:13] <_lb_> I take the snapshot, and of course itself is smaller than the image...
[02:27:34] <_lb_> scpd to my headnode... and created the manifest... etc...
[02:27:37] <_lb_> won't boot
[02:28:59] * jesse_ wonders if it's possible to read a snapshot
[02:29:19] <jesse_> what do you scp?
[02:29:29] <_lb_> a zfs.gz file
[02:29:33] <jesse_> I'd probable zfs send the snapshot
[02:30:01] <_lb_> zfs send zones/bd9649bb-1b6b-489c-b669-d9922deeed84-disk0@prototype03082013  | gzip -c > /var/tmp/win2008r2.prototype001.zfs.gz
[02:30:04] <_lb_> like that?
[02:33:34] <jesse_> did you import it with a manifest?
[02:34:13] <jesse_> and (I'm not sure about this) I think the image needs to be bzipped
[02:37:30] <jesse_> also, "headnode" sounds like sdc=)
[02:39:20] <_lb_> definitely sdc... but creating a dataset should be independent
[02:39:54] <_lb_> I created a manifest for the snapshot
[02:40:08] <_lb_> The root image was also created via snapshot
[02:40:17] <_lb_> But I can't replicate what we did :(
[02:41:15] <Licenser> jesse_ actually gzip works for imgamd :P I faced exactly that issue today
[02:44:21] *** xmerlin has quit IRC
[02:45:04] <jesse_> Licenser, good to know=)
[02:45:40] <Licenser> yea I dind't knew gz datasets existed, I hunted the bug 4h or something untill I noticed the dataset that wasn't inporting was gz-packed
[02:46:16] <jesse_> that reminds me
[02:46:49] <jesse_> I compiled up-to-date java on multiarch/trunk image
[02:47:22] <jesse_> anybody want/need it?
[02:48:05] <jesse_> (1.7.0_17-b02)
[02:48:37] <Licenser> no need for java, sorry :) but cool that you build it!
[02:48:40] <jesse_> it'll give you the nice feeling of a guinea pig when you run it!
[02:49:07] <jesse_> I compiled the first openjdk7 some two months ago
[02:49:34] <jesse_> but oracle kept hitting new 0-day bugs and taking weeks to get the source out and stamped on openjdk...
[02:50:18] <jesse_> I'm still trying to figure out what would be the easiest way of keeping the illumos/smartos patch alive
[02:50:35] *** szaydel has joined #smartos
[02:50:40] <Licenser> jesse_ I'm running my pre test fifo stuff :P I'm sure java is more stable then that
[02:51:00] <jesse_> Licenser, ah, but java is a platform
[02:51:07] <jesse_> it can fuxor anything you do=)
[02:51:34] <Licenser> heh
[02:51:44] <Licenser> I am not sold, sorry mate
[02:51:45] <jesse_> anyway, the older version did go through most of the tests without a problem
[02:51:55] <jesse_> only some awt/swing tests failed
[02:51:57] <Licenser> if you had a R16 erlang you might get me interested :P
[02:52:33] <jesse_> some, probably, because I had to test against remote X running on osx as X11 doesn't compile even Xvfb on smartos=)
[02:52:59] <Licenser> jesse_ there is a slight posibility yes
[02:53:05] <jesse_> My guess is, R16 erlang _can't_ be much worse than jdk=)
[02:53:25] <Licenser> I'm pretty sure it is very easy
[02:53:46] <Licenser> aside fromt he syntax that drives everyone but me nuts
[02:54:04] <Licenser> then again I can't spell for my life so the syntax looks quite usual to me P
[02:54:06] <jesse_> then again, it might be 2 months of waddling through newly found bugs in everything to get it to work=)
[02:56:17] *** xmerlin has joined #smartos
[03:03:46] *** Teknix has quit IRC
[03:07:08] *** AlainODea has quit IRC
[03:12:15] *** axonpoet has joined #smartos
[03:12:53] *** xmerlin has quit IRC
[03:16:35] *** alpharender has joined #smartos
[03:16:55] *** trist4n has joined #smartos
[03:17:07] <trist4n> hi! anyone got any thoughts about combining smartOS with drive encryption
[03:22:16] *** szaydel has quit IRC
[03:37:51] *** axonpoet has quit IRC
[03:39:19] *** miine_ has joined #smartos
[03:39:44] *** miine has quit IRC
[03:39:44] *** miine_ is now known as miine
[04:03:08] *** tonyarkles has quit IRC
[04:04:32] *** dysinger has quit IRC
[04:04:54] *** miine has quit IRC
[04:10:50] *** alpharender has quit IRC
[04:13:38] *** alpharender has joined #smartos
[04:14:20] <LeftWing> trist4n: Yes.  Specifically: what is the threat model that you are attempting to defend against?
[04:16:45] <jelmd> someone is beaming the hdds out of its servers ;-)
[04:17:35] <jesse_> or taking the server out entirely and booting somewhere else
[04:17:49] <jesse_> or the disks are on an unreliable san
[04:18:05] <LeftWing> Unreliable meaning what, exactly?
[04:18:08] *** szaydel has joined #smartos
[04:18:18] <jesse_> LeftWing, 'managed by not-you'?=)
[04:18:42] <LeftWing> Well, I wouldn't generally use a SAN, even if I *did* manage it.
[04:18:43] <jesse_> unreliable is not the right here, though, I think
[04:18:48] <jesse_> untrustworthy?
[04:19:07] <jesse_> LeftWing, that's your pro/con=)
[04:19:21] <jesse_> apparently, there are people who do
[04:19:25] <LeftWing> There are.  I was one of them!
[04:19:33] <jelmd> if you really care, put it into a bunker ;-)
[04:19:37] <LeftWing> There are meetings you can go to, while recovering.
[04:19:58] <jesse_> then you got taste of the smartos mushrooms and everything went rainbow?=)
[04:20:12] <jelmd> lol
[04:20:32] <jesse_> anyway, I wouldn't say no to zfs encryption
[04:20:48] <jesse_> I have pools that don't need constant access, and less access is better...
[04:20:59] <LeftWing> ZFS encryption sounds like an excellent way to replace a probably non-existant threat with a very real, very operator error-prone key management problem.
[04:21:13] <jesse_> true
[04:21:22] <LeftWing> *existent
[04:21:42] <jesse_> but the key management problem is infinidesimal compared to the cheap-sata-disks -problem=)
[04:22:04] <LeftWing> Sure, but encryption is nothing to do with cheap SATA disks.
[04:22:06] <jesse_> anyway, I'd be nice to be able to give a disk to the local dealer 'this broke, replace it'
[04:22:17] <jesse_> it helps with that
[04:22:32] <jesse_> now it's 'can I have more of that breaking crap, please?'
[04:22:43] <LeftWing> If your data is sufficiently sensitive, then you should not give the drive back to the vendor.
[04:22:59] <LeftWing> At least, not without erasing it somehow.
[04:23:17] <jesse_> ...like with not writing the data on it in the first place?
[04:23:25] <jesse_> ...like when you encrypt it?
[04:25:04] <LeftWing> In theory, sure.  In practice, nobody cares about whatever weird pornography you have on your disks.  If you have PCI (or other regulatory compliance issues) then you ought to be dealing with those in your application.
[04:25:24] <jesse_> or, in the fs...
[04:25:43] <jesse_> you know, if you need to compresss something, then you ought to be dealing with those in your application
[04:26:14] <jesse_> hell, even my laptop disk is encrypted
[04:26:15] <LeftWing> Doing compression in the filesystem has tangible benefits -- you tend to reduce the level of I/O done all around.
[04:26:27] <jesse_> and the disk images on it with customer data are encrypted
[04:26:42] <jesse_> (I'm counting on the keys not cancelling each other here...=))
[04:27:15] <jelmd>  http://www.supermagnetman.net/product_info.php?cPath=31&products_id=220
[04:27:19] <jesse_> LeftWing, I've done compression on the app level to get more disk space
[04:27:29] <LeftWing> This is all an Engineering problem -- that is: a trade-off between dimensions like "effort to implement", "cost", "actual requirement", etc.
[04:27:32] <jesse_> ending up getting better performance for free
[04:27:49] <LeftWing> There is a clear existence proof for ZFS-level Encryption, so it can be done.
[04:28:09] <jesse_> the specs for it are crappy at best, though=)
[04:28:11] *** avrntsv has quit IRC
[04:28:12] *** vsomes has joined #smartos
[04:28:21] <jesse_> 'we'll reuse this bit and that bit for xxxx'
[04:28:34] <LeftWing> I would argue, however, that everybody is better off doing encryption in a way that makes sense for their application.  It's complicated to do correctly, entirely generally and usefully in the filesystem.  Once it goes in there, you can't really change it.
[04:28:48] <LeftWing> You can do whatever the damn hell you like in your application, though, as I can in mine.
[04:28:54] <jesse_> well, if your application is 'nas'
[04:28:59] <jesse_> then you're fubar
[04:29:12] <jesse_> because it would require the kernel to do it for nfs?
[04:29:24] <jesse_> and yes, disks get stolen
[04:29:40] <LeftWing> Not generally from secure DCs
[04:29:51] <jesse_> but from offices
[04:29:52] <LeftWing> It's
[04:30:01] <jesse_> think of your print server
[04:30:13] <LeftWing> much more likely that you have not correctly secured the OS, or the OS is buggy and cannot be secured, and thus you are broken into over the Internet or some other network.
[04:30:30] <LeftWing> Or, frankly, it's more likely that some shithead in your company steals your data.
[04:30:40] <LeftWing> My print server does not require encrypted disks.
[04:30:42] <jesse_> they'd still need the keys
[04:30:54] <LeftWing> No, I mean: some shithead in the company *with* *access* to the *NAS*.
[04:31:00] <LeftWing> i.e. somebody you thought you could trust.
[04:31:19] <jelmd> But usually those guys stealing stuff don't even know, what ZFS is ;-)
[04:31:31] <jesse_> jelmd, that's true=)
[04:31:35] <LeftWing> They don't need to -- they're just copying files from their Windows desktop to the USB stick.
[04:31:38] <LeftWing> Stop thinking about stealing the disks.
[04:31:40] <jesse_> but that's just obscurity
[04:31:42] <jelmd> security by obscurity ;-)
[04:31:49] <jelmd> :)
[04:32:16] <jesse_> anyway, it's not unheard of disks going missing
[04:32:18] <LeftWing> I'm trying to point out that there are about a thousand things more likely, which you probably haven't protected against sufficiently, before you get *anywhere* *near* worrying about encrypting data at rest on disks.
[04:32:32] <jelmd> yepp
[04:32:51] <jesse_> LeftWing, yes, but *after* you've taken care of all that stuff, you'll still need the encryption
[04:33:06] <LeftWing> And that for applications that *require* for regulatory compliance that you encrypt data at rest, that is best done in your application in a flexible enough way to be useful and compliant.
[04:33:46] <jesse_> LeftWing, except, that the app can be a windows app, running in KVM, saying 'use disk encryption'
[04:33:52] <jesse_> oooops
[04:34:22] <jesse_> (not that I've seen it. The windows that is, I haven't gotten that far yet, even=))
[04:34:31] <LeftWing> If you think that doing "security sensitive" things on a Windows server is production grade, then I direct you back up to "about a thousand things more likely"
[04:34:57] <jesse_> most likely, someone steals the whole box and boots it up
[04:35:12] <LeftWing> All of this boils down to: nobody who stands to make money from having on-disk encryption has, thus far, been willing to pay for it to be implemented.
[04:35:23] <LeftWing> Which generally means it's a problem that doesn't actually require a solution at this time.
[04:35:37] <jesse_> LeftWing, on the same news, we have weekly leaks of passwords
[04:35:38] <jesse_> 0)
[04:35:39] <jesse_> =)
[04:35:53] <jesse_> (yes, I know...)
[04:36:33] <jesse_> LeftWing, eventhough it's a problem that doesn't require a solution, technically, may be a requirement nonetheless
[04:36:39] <LeftWing> Insofar as that's in any way related to encrypting data at rest, I would direct you back to "that is best done in your application" -- namely, a password vault with encryption that makes sense for small chunks of data.
[04:36:42] <jesse_> (case in point: TSA)
[04:37:17] <LeftWing> I can't fix the TSA.
[04:37:26] <jesse_> re-re-re-re-implementing encryption to every application is waste of resources
[04:38:16] <jesse_> (ofcourse, it to be equivivalent to application encryption, one would need something besides totally transparent fs encryption)
[04:38:22] <LeftWing> You probably store your data in a database.  If you're wise, that database is probably PostgreSQL.  It does encryption pretty transparently.
[04:38:35] <jesse_> LeftWing, but you _can_ fix the fs!=)
[04:39:05] <jesse_> actually, I'm wiser than psql, but let's not get into that
[04:39:19] <jesse_> (and no, it doesn't do encryption yes)
[04:39:33] *** dysinger has joined #smartos
[04:39:35] <jesse_> yet
[04:39:43] <LeftWing> I would be keen to hear what database is wiser than PostgreSQL, but which does not yet do encryption?
[04:41:01] <jesse_> now you're assuming you're heard of all software
[04:41:30] <jesse_> I wish I had=)
[04:42:14] <LeftWing> No, I just can't stand nonsense answers that don't mean anything.
[04:42:31] <LeftWing> If it's software I can read about, even briefly, I would like to hear what you are using.
[04:42:34] <LeftWing> It's not like I can't Google it.
[04:49:12] *** denizr has quit IRC
[04:56:54] *** enmand has quit IRC
[05:07:35] <trist4n> LeftWing: nothing super important, its personal files on a nas that i've always encrypted in the past.
[05:07:51] <trist4n> i figure if im worried enough to have contents insurance and air gapped backups its worth encrypting to me.
[05:08:02] <trist4n> but its proving a gigantic pain
[05:08:23] <LeftWing> Contents insurance and backups are to prevent having to recreate the data -- encryption is to prevent somebody else from seeing it.
[05:08:49] <LeftWing> (well, s/data/worth/ for insurance)
[05:08:58] <trist4n> yeah right, im saying its my personal stuff, and i dowant want something opening it up if my desktop were to be stolen
[05:09:03] <trist4n> (which it has been before)
[05:09:18] <trist4n> i was attempting to state its relatively important stuff
[05:26:27] *** darjeeling has joined #smartos
[05:32:27] *** whil1eq1_away has joined #smartos
[05:32:31] <whil1eq1_away> anyone up?
[05:34:53] *** d[^_^]b has joined #smartos
[05:42:13] *** alpharender has quit IRC
[05:52:43] *** avrntsv has joined #smartos
[05:56:50] *** sachinsharma has joined #smartos
[05:57:57] *** theCzar has joined #smartos
[06:00:05] *** theCzar has quit IRC
[06:07:11] *** leecallen has quit IRC
[06:07:38] *** leecallen has joined #smartos
[06:14:04] *** matthewpucc_ has quit IRC
[06:18:53] *** szaydel has quit IRC
[06:38:37] *** Webhostbudd has joined #smartos
[06:45:57] *** nefilim has quit IRC
[06:51:56] *** while1eq1_ has joined #smartos
[06:54:12] *** while1eq1 has quit IRC
[06:58:34] *** while1eq1_ is now known as while1eq1
[06:58:54] *** while1eq1 has left #smartos
[07:02:51] *** while1eq1 has joined #smartos
[07:04:42] <trentster1> btw if a smartos server experienced hardware failure and you had to move the zones disk sets to another server and pxe-booted smartos on that server, wouldn't you need to change the mac addresses for the new nics in order for the machine to operate correctly again?
[07:08:32] *** wolfeidau has quit IRC
[07:13:53] *** porkbelt has quit IRC
[07:14:46] *** porkbelt has joined #smartos
[07:23:41] *** sachinsharma has quit IRC
[07:25:52] *** sachinsharma has joined #smartos
[07:40:22] *** whil1eq1_away has quit IRC
[07:57:50] <jelmd> I would always assign a fixed MAC for each vnic ...
[08:07:00] *** Meths has quit IRC
[08:07:40] *** Meths has joined #smartos
[08:14:59] *** alucardX has joined #smartos
[08:15:59] <alucardX> morning
[08:17:05] <MerlinDMC> morning / hello
[08:20:56] *** texarcana has quit IRC
[08:22:15] *** texarcana has joined #smartos
[08:26:06] *** bluezenix has joined #smartos
[08:38:08] *** Meths has quit IRC
[08:38:28] *** Meths has joined #smartos
[09:00:09] *** mikl has quit IRC
[09:16:31] *** alcir has joined #smartos
[09:23:18] *** Meths has quit IRC
[09:23:33] *** Meths has joined #smartos
[09:23:50] *** bluezenix has joined #smartos
[09:23:56] *** bluezenix1 has joined #smartos
[09:26:08] *** Meths has quit IRC
[09:28:33] *** Meths has joined #smartos
[09:31:18] *** siezer has quit IRC
[09:35:47] *** bens1 has joined #smartos
[09:35:47] *** Meths has quit IRC
[09:35:58] *** Meths has joined #smartos
[09:49:02] *** chh has joined #smartos
[09:58:43] *** KermitTheFragger has joined #smartos
[10:02:27] *** Meths has quit IRC
[10:02:44] *** Meths has joined #smartos
[10:04:01] *** Meths has quit IRC
[10:07:44] *** Meths has joined #smartos
[10:12:45] *** Meths has joined #smartos
[10:19:23] *** Meths has quit IRC
[10:20:11] *** Meths has joined #smartos
[10:23:01] *** Meths has quit IRC
[10:31:13] *** chh has quit IRC
[10:31:59] *** chh has joined #smartos
[10:41:51] *** alpharender has joined #smartos
[10:48:50] *** darjeeling has quit IRC
[11:20:13] *** chh has quit IRC
[11:22:04] <linuxprof> problems with the latest ubuntu dataset? i get "Error: zfs receive exited non zero: 1" when trying to install it
[11:23:03] <linuxprof> maybe i forgot update. hrm.
[11:25:32] <MerlinDMC> linuxprof, zfs list -t all | grep 1327ed36-5130-11e2-95a8-9b5a153adf3e ... maybe there is an partially transferred version that you should delete before importing
[11:25:52] <linuxprof> check that, there isnt
[11:27:00] <MerlinDMC> hmm ... I don't know if there is any better log for the zfs execution
[11:27:21] *** siezer has joined #smartos
[11:31:40] *** chh has joined #smartos
[11:32:33] <linuxprof> is it because im not running the latest smartos? since imgadm was rewritten
[11:33:46] <MerlinDMC> it should fetch + install without problems
[11:33:54] <MerlinDMC> the old dsapi service did not change
[11:34:02] <linuxprof> should and does are two different things ;)
[11:42:22] <linuxprof> anyone has bind running in a zone?
[11:43:44] *** wolfeidau has joined #smartos
[11:45:59] <linuxprof> nvm, human error as usual.
[11:49:16] <alcir> hi
[11:49:18] <alcir> Missing required properties: disks.0.image_size
[11:49:26] <alcir> I'm creating a KVM vm
[11:49:40] <alcir> why I must specify image_size?
[11:49:52] <MerlinDMC> alcir, using an provided image or an empty one to for example install windows?
[11:53:36] <alcir>  provided image
[11:54:08] <alcir> centos-6
[11:54:54] <MerlinDMC> then you need to specify iamge_size with the exact value that is in the manifest so that vmadm can create the zvol correctly
[11:55:04] <MerlinDMC> *image_size
[11:55:35] <MerlinDMC> or at least it should be exactly what the manifest tells you ... it can also be larger
[11:59:15] <alcir> but it is a new thing
[11:59:16] <alcir> ?
[11:59:18] *** darjeeling has joined #smartos
[11:59:25] <Eimann> VLANs on aggr-interfaces are support in smartos, right?
[11:59:36] <Eimann> *supported
[12:00:19] <MerlinDMC> alcir, not new ... but was not forced on much older smartos platforms afaik
[12:06:56] <alcir> MerlinDMC what is the right way to get such info?
[12:08:40] <MerlinDMC> alcir, imgadm info <uuid> | json manifest.image_size
[12:11:00] <alcir> and
[12:11:11] <alcir> it is different from size
[12:11:14] <alcir> ?
[12:11:45] <MerlinDMC> you should use the size property only on disks that don't use image_uuid
[12:24:39] <alcir> and if I want to create a vm with more space?
[12:25:54] <MerlinDMC> add a second disk for your data
[12:25:55] *** avrntsv1 has joined #smartos
[12:26:27] <MerlinDMC> even if you specify a larger size for the image_size property hte root disk is not resized automatically
[12:26:54] <MerlinDMC> if you need space t store stuff you should add a disk for your data (that is automatically mounted at /data)
[12:28:04] *** avrntsv has quit IRC
[12:29:21] <alcir> thanks
[12:29:24] <alcir> great
[12:31:05] <Eimann> admin_vlan_id doesn't seem to work on LACP/aggr-interfaces :-/
[12:35:10] <MerlinDMC> Eimann, afaik the admin interface can't get a vlan ... but I'm not sure because I never use them
[12:35:52] <Eimann> https://github.com/joyent/smartos-overlay/pull/5
[12:35:59] <Eimann> I probably need to look at the code :)
[12:40:55] *** szaydel has joined #smartos
[12:48:11] *** alpharender has quit IRC
[12:50:02] *** sachinsharma has quit IRC
[12:54:11] *** ira has joined #smartos
[12:59:12] *** avrntsv1 has quit IRC
[13:02:41] *** bens1_ has joined #smartos
[13:04:39] *** bens1 has quit IRC
[13:04:39] *** bens1_ is now known as bens1
[13:05:08] <Eimann> apparently the admin_vlan_id code is not in the 20130307 release ...
[13:05:43] *** avrntsv has joined #smartos
[13:06:47] *** enmand has joined #smartos
[13:10:18] *** avrntsv has quit IRC
[13:10:25] *** alucardX has quit IRC
[13:11:00] *** avrntsv has joined #smartos
[13:13:18] *** enmand has quit IRC
[13:28:07] *** alucardX has joined #smartos
[13:28:43] *** alcir has quit IRC
[13:29:39] *** enmand has joined #smartos
[13:31:24] *** avrntsv has quit IRC
[13:31:25] *** darjeeling has quit IRC
[13:32:40] *** bluezenix1 has quit IRC
[13:36:29] *** miine has joined #smartos
[13:37:58] *** darjeeling has joined #smartos
[13:40:47] *** avrntsv has joined #smartos
[13:42:24] *** enmand has quit IRC
[13:42:38] *** enmand has joined #smartos
[13:46:25] *** avrntsv has quit IRC
[13:46:43] <linuxprof> is there a good way to make imgadm tell me why it cant import a dataset?
[13:49:24] <MerlinDMC> linuxprof, is that imgadm v1 or the newer one?
[13:49:40] <linuxprof> v1
[13:50:15] *** AlainODea has joined #smartos
[13:50:44] *** orangeroo has joined #smartos
[13:51:34] <MerlinDMC> AlainODea, hey ... I did look into that imgadm + proxy stuff but don't have the code ready - can I send you a repo link as soon as I have something to test?
[13:51:56] <AlainODea> MerlinDMC: definitely!  Thanks :)
[13:52:24] <AlainODea> I'm alain DOT odea at gmail
[13:52:40] <AlainODea> and AlainODea on Github
[13:52:40] <MerlinDMC> I think I got your mail address here somewhere
[13:52:46] <AlainODea> Cool :)
[13:53:31] <MerlinDMC> I also saw your other request on restify + proxys on ... don't know where that was
[13:53:32] <MerlinDMC> ;)
[13:54:07] <MerlinDMC> stackoverflow i guess
[14:02:20] *** Hazel|artemis has joined #smartos
[14:02:28] <Hazel|artemis> btw, typo on http://joyent.com/products/smartdatacenter/
[14:02:38] <Hazel|artemis> it says "Options to run Linux or Window operating systems"
[14:02:42] <Hazel|artemis> s/Window/Windows/ I presume?
[14:06:56] *** Teknix has joined #smartos
[14:11:14] <nahamu> Hazel|artemis: probably a typo, yes.
[14:12:53] <Hazel|artemis> nahamu: yes, I was wondering it this was something I should report to them, and if so, whether mentioning it here will get their attention, or if I should email them?
[14:13:20] <jperkin> email, I don't know that anyone who works on the website hangs out here
[14:13:28] <Hazel|artemis> tbf, it's not like I'm going to buy SmartDataCentre for my personal use, and we don't use Solaris at work, so it was a friendly "heads up" rather than anything else
[14:13:32] <Hazel|artemis> jperkin: okiedokie
[14:13:55] <jperkin> I would hope that webmaster@ would get to them, but it seems fashionable these days to ignore RFCs so who knows ;)
[14:17:16] <Hazel|artemis> well, RFC822 says that "postmaster@" is special
[14:17:31] <Hazel|artemis> but I wasn't aware that "webmaster@" was required by RFC?
[14:17:43] <Hazel|artemis> I thought it was merely convention
[14:18:31] <jperkin> rfc2142
[14:20:40] <Hazel|artemis> jperkin: ah, fair point. I saw that and skipped over it (whilst browsing in links) as it seemed to be talking about NNTP, whoops.
[14:23:01] *** vsomes has quit IRC
[14:24:43] <Hazel|artemis> jperkin: you win two points...
[14:24:44] <Hazel|artemis> Delivery to the following recipient failed permanently:
[14:24:45] <Hazel|artemis> webmaster at joyent dot com
[14:24:53] <Hazel|artemis> Google tried to deliver your message, but it was rejected by the server for the recipient domain joyent.com by aspmx.l.google.com
[14:25:00] <Hazel|artemis>  The email account that you tried to reach does not exist
[14:25:02] <Hazel|artemis> yada yada
[14:25:02] <Hazel|artemis> heh
[14:26:42] <Hazel|artemis> forwarded to sales@, the only address I can find on the homepage
[14:26:53] <Hazel|artemis> now to stop wittering about a 1 letter typo on IRC
[14:26:54] <Hazel|artemis> XD
[14:27:21] *** darjeeling has quit IRC
[14:28:36] *** bluezenix has joined #smartos
[14:35:45] *** mikl has joined #smartos
[14:39:09] <jperkin> yeh, I'm not even surprised anymore ;)
[14:41:38] *** denizr has joined #smartos
[14:41:55] *** neophenix has joined #smartos
[14:48:22] <yofuh> am i lucky and someone has a smartos image with sunstudio and apache STL installed?
[14:49:15] <yofuh> i use it for testing a certain C++ application before releases, and my option now is booting up an openindiana somewhere
[14:50:22] <ira> yofuh: If you have an image server… making up your own is very viable.
[14:51:37] *** ktk has joined #smartos
[14:51:44] <ktk> hi everyone
[14:51:51] <ira> (Or if the zone is persistent… I know building SmartOS they install Sun's compiler…)
[14:52:32] <ktk> short q: one of my kvm vms needs IP aliases, I could configure it in the guest OS but can't ping/access it. I suppose I have to make this known to the smartos vm config as well but couldn't find hints
[14:52:47] <ktk> anyone got some pointers to doc?
[14:53:00] <MerlinDMC> ktk, currently one ip per interface ... so you could add nics
[14:53:11] <ktk> ah ok
[14:53:46] <yofuh> ira: makeing an image doesn't require an image server, but installing apache stl from scratch requires more work than i'm willing to spend on that
[14:54:06] <ktk> MerlinDMC: will that work from a routing POV when the IP is in the same subnet? won't add a second default gw then
[14:54:12] <ira> Does pkgin have it?
[14:54:32] <yofuh> ira: and no, the sun compiler is not the issue here, it is the apache stl that sun folks made work with sun studio and is not bundled with the compiler
[14:54:43] <ira> ah!
[14:54:52] <MerlinDMC> ktk, you can only make one nic primary .. so you'll end up having only one default gateway
[14:55:13] <ktk> right so the routing table should work
[14:55:14] <ktk> ok
[14:55:21] <ktk> MerlinDMC: tnx will try that
[14:55:30] <MerlinDMC> if unsure ... you can probably just leave the gateway line out of the json as well
[14:55:35] <ktk> yeah
[14:55:56] <MerlinDMC> or set netmask to 255.255.255.255 ;)
[14:56:22] <ktk> yeah that's what I do in the alias config on fbsd so far
[14:57:27] *** axonpoet has joined #smartos
[15:05:59] *** Teknix has quit IRC
[15:07:28] *** Teknix has joined #smartos
[15:08:35] *** ipalreadytaken has joined #smartos
[15:10:50] <leecallen> Good morning.  Joining smartos to an AD domain...
[15:11:04] <leecallen> it cannot be done in a non-GZ, so I joined the GZ...
[15:11:14] <leecallen> but how to make it persistent?
[15:11:24] <leecallen> do I need to make the necessary config changes and 'smbadm join' on each boot?
[15:12:28] <MerlinDMC> leecallen, create a smf service that gets executed every boot ... there was some little doc about that
[15:13:00] *** Teknix has quit IRC
[15:13:04] <leecallen> and that service would set the configs and do the smbadm join?
[15:13:53] <MerlinDMC> leecallen, http://wiki.smartos.org/display/DOC/Administering+the+Global+Zone#AdministeringtheGlobalZone-CreatingPersistentServicesUsingSMF
[15:14:30] *** Teknix has joined #smartos
[15:15:48] *** elijah-away has quit IRC
[15:15:54] <leecallen> MerlinDMC: Thanks.  I have set up such an SMF already - launches a script that performs my on-the-fly customization.
[15:16:17] <leecallen> MerlinDMC: I know I can extend that to do this too, just needed to hear that's the way to do it.
[15:22:54] *** Teknix has quit IRC
[15:24:19] *** Teknix has joined #smartos
[15:24:54] <MerlinDMC> leecallen, would be better to have smth like chef or puppet do the configuration but yeah ... the script should do as well
[15:27:07] *** daleg has joined #smartos
[15:27:25] <leecallen> okay thanks.
[15:35:32] *** chorrell has joined #smartos
[15:37:39] <linuxprof> I'm using the IPv6 guide from the wiki http://wiki.smartos.org/display/DOC/Setting+up+IPv6+in+a+Zone but i get permission denied when running the ipadm commands
[15:37:44] <linuxprof> what have i missed? =)
[15:38:40] <nahamu> leecallen: do you have to provide credentials when you do the "smbadm join"?
[15:43:15] <leecallen> nahamu: yes
[15:43:43] <linuxprof> I didn't miss anything except restarting the zone. sigh.
[15:44:18] <nahamu> leecallen: yeah, that's probably going to be a bit annoying to automate.
[15:44:59] <leecallen> maybe.  I am having an offline conversation with Ira, he pointed out I will have to put the AD admin username & password in a script
[15:45:10] <leecallen> not such a lovely thing to do
[15:45:13] <nahamu> right.
[15:45:20] <ira> Or just an account with join permissions.
[15:45:34] <leecallen> yes, better
[15:45:48] <ira> On a small OU… not the end of the world.. but still ugly. ;)
[15:45:58] <windows> hello
[15:46:08] <nahamu> in theory if after running the join the system drops config files, might there be some way of saving those off somewhere, then on a reboot pushing them back into place and bouncing the service?
[15:46:09] <leecallen> thanks all, I have some good direction here
[15:46:11] <ira> (I thnk you'll need remove and join?)  But I'm not an AD guy.
[15:46:43] <leecallen> nahamu: I am going to look into that - does smbadm join create a file I can save and restore on reboot?
[15:46:55] <nahamu> leecallen: I don't know. I've never used it.
[15:46:57] <leecallen> I didnt mean to ask that, thats the question I will investigate
[15:47:15] <nahamu> I'd certainly be curious to hear about what you find.
[15:47:29] <leecallen> okay I will report back here
[15:49:02] *** tonyarkles has joined #smartos
[15:49:45] <windows> I'm running SmartOS at home without SDC, i want to host cifs/nfs on the network, is it "better" to do this in a SmartMachine or in the global zone?
[15:52:47] <jperkin> you can't serve NFS from a zone, but samba works fine
[15:53:02] *** Teknix has quit IRC
[15:54:29] *** Teknix has joined #smartos
[15:56:53] <windows> jperkin: so if I want to export nfs, I have to host it in the global zone?
[15:58:08] <jperkin> yes
[15:58:52] *** khushildep has joined #smartos
[15:59:09] <windows> maybe OmniOS is better fit for my needs.
[15:59:53] <rmustacc> Could very well be.
[16:00:17] <windows> I don't want to run more than one server at home, but want to have ZFS backed nfs and cifs.
[16:00:31] <jperkin> if you only care about NAS features then yes.
[16:00:36] <windows> plus zones and ability to run KVM guests
[16:01:00] <windows> NAS is seconday to being a VM lab
[16:02:05] *** ipalreadytaken has quit IRC
[16:02:41] *** nefilim has joined #smartos
[16:02:57] <ira> If you want VMs/Zones, SmartOS is great… NAS.. well.. it is… a painful climb :)
[16:03:49] <windows> whats in the middle? OmniOS?
[16:04:27] <ira> OmniOS is a great system.  See if it fits.
[16:05:21] *** darjeeling has joined #smartos
[16:05:48] <windows> I'm one of the Linux folks learning this realm, it's been fun thus far.
[16:06:46] <rmustacc> Yeah, OmniOS will make the NAS stuff a lot simpler. It just doesn't, iirc, have much tooling around vms. You get qemu and that's it.
[16:08:24] <rmustacc> Not that you can't get it working on SmartOS right now, it's just painful.
[16:08:52] <rmustacc> More so than we'd really like it to be, but it's also not what we're trying to solve with SmartOS
[16:09:12] <MerlinDMC> windows, kschiess has management tools for zones/vms on omnios
[16:10:26] *** axonpoet has quit IRC
[16:10:27] <MerlinDMC> https://bitbucket.org/kschiess/qemu-toolkit afaik
[16:13:09] <windows> ira, rmustacc, MerlinDMC, jperkin : thanks for the info
[16:13:15] *** chorrell has quit IRC
[16:14:42] *** sebasp has joined #smartos
[16:15:12] *** khushildep has left #smartos
[16:17:52] *** tonyarkles has quit IRC
[16:19:27] *** tonyarkles has joined #smartos
[16:19:41] *** denizr has quit IRC
[16:20:01] *** denizr has joined #smartos
[16:24:26] *** kfr- has joined #smartos
[16:25:49] *** dap has joined #smartos
[16:33:53] *** Teknix has quit IRC
[16:35:09] *** dn2k has left #smartos
[16:35:19] *** Teknix has joined #smartos
[16:35:21] *** CarlosC has joined #smartos
[16:43:03] *** Teknix has quit IRC
[16:44:27] *** Teknix has joined #smartos
[16:54:23] *** sebasp has quit IRC
[16:54:38] *** sebasp has joined #smartos
[16:54:47] *** ryancnelson has joined #smartos
[16:56:08] *** chh has quit IRC
[16:58:26] <ktk> MerlinDMC: btw related to my other question, what happens when I want an IPv6 address as well, not possible on the same interface right now?
[17:00:18] *** alucardX has quit IRC
[17:00:48] <nahamu> has anyone else been having trouble creating their own template images (for KVM) on the latest version of the platform? I'm getting "Invalid value(s) for: disks.*.image_uuid,disks.0.image_uuid" errors...
[17:12:39] *** ipalreadytaken has joined #smartos
[17:15:41] *** ipalreadytaken has quit IRC
[17:17:31] <leecallen> nahamu: back to 'smbadm join' - it looks like the AD join state info is stored in memory only, not in files
[17:17:59] <nahamu> leecallen: that sounds odd... how does it persist across reboots on e.g. OmniOS?
[17:19:06] <leecallen> I don't know.  I think I drilled all the way down through the (Illumos) source code and I see it stored in memory only...
[17:19:42] <leecallen> also I took filesystem snapshots before & after joining (smartos) and I don't find any differences -- but I looked only at /etc  /usr  and  /var
[17:20:55] <leecallen> (off to lunch now, I will check back later)
[17:30:01] *** bluezenix has quit IRC
[17:40:18] *** tonyarkles has quit IRC
[17:45:11] *** dalethion has quit IRC
[17:45:24] <yofuh> is 2012Q2 still the latest repository for smartos packages with smf integration? is rhe rails there up2date (without remote exploits)?
[17:46:14] <jperkin> yofuh: there are 2012Q4 packages for base1.9.x, and also trunk (though SMF only for packages which we've migrated)
[17:47:12] <yofuh> jperkin: did you change the baseurs or is it just not done (only can find http://pkgsrc.joyent.com/sdc6/ )
[17:47:37] <jperkin> yofuh: right, we moved to http://pkgsrc.joyent.com/packages/SmartOS/
[17:48:12] <yofuh> ok
[17:48:12] <jperkin> to avoid confusion for sdc >6 (sdc5 and sdc6 were very different and so necessitated the distinction)
[17:48:58] <yofuh> well, i did wonder what sdc6 is about, but for me i'm happy if i get updates
[17:49:22] <yofuh> let's see hoe good a binary upgrade from 2012Q1 to Q4 will work...
[17:49:33] *** tonyarkles has joined #smartos
[17:49:43] <jperkin> assume it will blow up, then you won't be disappointed ;)
[17:51:11] <yofuh> i'd be disappointet if everything else than that happens, i had to install some ruby additional gems, so the system is f*cked for sure
[17:51:59] <yofuh> gems, cpan and whatever uses a similar system played never good witch package management
[17:52:30] <yofuh> half the packages are to hard to build into packages and the other half will bitch arround when you update
[17:53:15] <yofuh> for some reasons developer just don't stop using that stuff, because everything seems so easy on their xcode
[17:57:56] <opeth__> jperkin: is there somewhere a mysql-enabled postfix for the last base64 image in pkgsrc?
[17:58:33] <jperkin> opeth__: no, you'd need to build it yourself
[17:59:16] <jperkin> opeth__: you should be able to follow http://www.perkin.org.uk/posts/pkgsrc-on-smartos-zone-creation-and-basic-builds.html mostly to the letter, and just set PKG_OPTIONS.postfix+= mysql or so.
[18:00:37] <opeth__> yeah, I've tried doing so and realized it failed because I did it on a base64 zone
[18:00:58] <opeth__> but I thought I'd wanted a 64-bit build which you seem to have been able to provide
[18:01:18] <jperkin> opeth__: it shouldn't matter whether it's base or base64
[18:01:38] <opeth__> the base64 bmake dies trying to acquire gcc-3.3.6 because of nbpatch
[18:01:54] <opeth__> because it finds no binary and tries to build it but fails
[18:01:59] <jperkin> opeth__: ah, remove GCCBASE from /opt/local/etc/mk.conf
[18:02:44] <opeth__> it's not in there
[18:02:59] <jperkin> did you pkgin in gcc47?
[18:03:13] <opeth__> yes of course
[18:03:24] <opeth__> could it have been the bad src uri?
[18:03:25] <opeth__> I've used...
[18:03:28] <jperkin> is there no /opt/local/bin/cc symlink?
[18:03:33] <opeth__> BINPKG_SITES=           http://pkgsrc.joyent.com/sdc6/2012Q2/x86_64/
[18:03:43] <jperkin> if not, add 'GCCBASE=/opt/local/gcc47' to mk.conf
[18:04:06] <jperkin> that looks ok
[18:04:34] <jperkin> though if you are using base1.9.x you'd need to use http://pkgsrc.joyent.com/packages/SmartOS/2012Q4/x86_64/
[18:04:49] <opeth__> noted, correcting
[18:05:07] <opeth__> do I need a new zone or could I upgrade to these packages?
[18:05:27] <jperkin> if you haven't installed much then an upgrade should be ok
[18:06:23] <opeth__> hm
[18:06:24] <opeth__> SmartMachine base64 1.8.4
[18:06:37] <opeth__> I thought it was the last base64 I saw
[18:06:56] *** bluezenix has joined #smartos
[18:06:56] <jperkin> 1.9.x was only released a week or so ago
[18:07:47] *** dysinger has quit IRC
[18:07:47] *** leecallen has quit IRC
[18:08:15] *** leecallen has joined #smartos
[18:08:45] *** dysinger has joined #smartos
[18:10:10] <opeth__> *** Configuration x86_64-sun-solaris2.11 not supported
[18:10:13] <opeth__> that's what it says
[18:10:34] <opeth__> the symlink didn't exist, so I added the GCCBASE variable to mk.conf
[18:10:41] <jperkin> gist/pastebin the full log somewhere, along with your mk.conf ?
[18:10:49] <opeth__> then CC would have had to be set still so I added /opt/local/gcc47/bin to $PATH
[18:10:53] <opeth__> then I end up here.
[18:11:01] <opeth__> that's as far as I've ever gotten
[18:11:08] <jperkin> you don't need to do that, GCCBASE ensures it gets set
[18:11:27] <opeth__> maybe mk.conf isn't sourced then?
[18:11:58] <jperkin> how are you determining that CC isn't set?
[18:12:01] <opeth__> but anyway, that's the least of the pain, the path
[18:12:15] <opeth__> by configure raising an error for it
[18:12:26] <jperkin> possibly, it should be mostly superfluous but can occasionally be harmful
[18:12:54] <jperkin> make sure you do a 'bmake clean clean-depends' before retrying after modifying mk.conf
[18:12:56] <opeth__> but once gcc is found in $PATH, the error I get is that evil not-supported one
[18:18:36] <yofuh> jperkin: any hint how to uninstall or upgrade bootstrap-mk-files? it's keep telling me that it is maked as not for deletion even if i unkeep it
[18:18:58] *** powerosx has joined #smartos
[18:19:09] <powerosx> hi there
[18:19:47] <powerosx> I had a super idea...
[18:20:16] <jperkin> yofuh: easiest is probably to just unpack the newer bootstrap over the top, e.g. http://pkgsrc.joyent.com/packages/SmartOS/bootstrap/bootstrap-2012Q4-x86_64.tar.gz
[18:20:36] <jperkin> then do a pkg_admin rebuild
[18:20:49] <powerosx> is it possible to create a zone but… with openindiana inside???
[18:21:11] <ryancnelson> yes.  On openindiana.
[18:21:11] *** szaydel has quit IRC
[18:21:23] <powerosx> noooooo
[18:21:41] *** denizr has quit IRC
[18:21:45] <powerosx> please…. I want to create with smartos
[18:21:59] *** denizr has joined #smartos
[18:22:15] <ryancnelson> sorry.  no.
[18:22:34] <powerosx> I-m trying to create in openindiana, then export the zone.. and import in smartOS ...
[18:22:38] <ryancnelson> you can run openindiana as a kvm-guest
[18:22:43] <ryancnelson> zones don't work that way.
[18:23:39] <powerosx> but it's always a zfs base… illumines kernel… and same package
[18:23:41] <ryancnelson> a zone is an OS-partition of the parent kernel.  the parent kernel on smartOS is, by definition, smartos
[18:23:57] <ryancnelson> begging won't make it happen.
[18:23:57] <rmustacc> The problem is that openindiana does not use a sparse image.
[18:24:25] <rmustacc> So you could make it work, but you'd have to insure that you were starting with the target smartos set of kernel versions, libraries, etc.
[18:24:29] <powerosx> what is sparse image?
[18:24:35] <rmustacc> Then every time you update the gz you have to update everything else.
[18:24:59] <rmustacc> In an ipkg zone like you have with OI and OmniOS, you have full root zones.
[18:25:02] <opeth__> jperkin: I wish I knew what made it work but it's compiling postfix now :) thank you very much! I'll try to grok it.
[18:25:09] <rmustacc> That means every zone has a copy of libc, etc.
[18:25:22] <nahamu> powerosx: what features of openindiana are you looking for?
[18:25:23] <rmustacc> The version of libraries such as libc must match the kernel.
[18:25:39] <rmustacc> SmartOS takes an alternate approach.
[18:25:54] <rmustacc> Every zone is sparse in that it shares a copy of /usr from the GZ. The side effect of this is that you can't install into the GZ.
[18:26:23] <nahamu> if all you care about is being able to "pkg install stuff" and you're aiming to do server-y things, perhaps http://wiki.smartos.org/display/DOC/3rd+Party+Software+Repos is a good answer.
[18:26:26] *** bdha has quit IRC
[18:26:27] <powerosx> I know.. know.. but smartos doesn 't have a good support for a storage server
[18:26:40] <nahamu> oh, you want to make a storage server...
[18:26:48] <nahamu> yeah, OI or OmniOS.
[18:26:56] <powerosx> I know
[18:26:56] <nahamu> (or learn to do it the SmartOS way)
[18:27:24] <rmustacc> Basically we don't support pkg(5) full root zones on SmartOS.
[18:27:25] <yofuh> jperkin: to late, i voted for removal of the +PRESERVE file and uninstall the old before pkgin the new
[18:27:41] <powerosx> but smarts is a very smart approach
[18:27:49] <jperkin> opeth__: without the 'clean' it was probably still using the wrong dependencies from when GCCBASE was set incorrectly
[18:28:20] <jperkin> opeth__: ..which is mostly my fault, as it wasn't removed prior to creating those images - will be fixed for future.
[18:28:32] *** mamash has joined #smartos
[18:28:33] <ira> powerosx: What feature of smartos is driving you to use smartos?
[18:28:55] <powerosx> I just thinking to start my server with usb boot on smartos
[18:29:20] <opeth__> jperkin: thank you. any ideas perhaps on 'xsasl_cyrus_server.c:77:18: fatal error: sasl.h: No such file or directory
[18:29:24] <opeth__> '?
[18:29:26] <powerosx> then with a zone with inside openindiana a good storage server
[18:29:44] <opeth__> could it be configuring itself with a wrong parameter or is a dependency not pulled in?
[18:29:51] <ira> Use ports and a normal smartos zone.
[18:30:01] <jperkin> opeth__: something isn't including cyrus correctly, which package is failing?
[18:30:02] <ira> When on SmartOS… do things the SmartOS way.
[18:30:07] <ira> Or suffer… :)
[18:30:17] <opeth__> jperkin: postfix-2.8.13
[18:30:50] <powerosx> why smartos does not have a vm designed for storage???
[18:31:00] <jperkin> opeth__: does 'bmake show-options' include sasl?
[18:31:03] *** szaydel has joined #smartos
[18:31:03] <ryancnelson> because zones are not vm's
[18:31:06] <ktk> I'm still trying to assign more than one IP from the same subnet to a kvm guest, I tried the IP and the correct netmask and the ip and 255.255.255.255 as netmask. result is that both ways the ip is not reachable outside the vm
[18:31:13] <opeth__> jperkin: /usr/include/sasl/sasl.h exists
[18:31:15] <opeth__> letme see
[18:31:18] <ktk> anyone knows what would be the right way doing that?
[18:31:39] <opeth__> jperkin: yes it does, I've added it to the configuration myself manually before trying the build the first time
[18:32:01] <jperkin> opeth__: pkgsrc has a system called 'buildlink' which creates a specific environment for compiling, so even if certain incldues/libraries exist on the system, buildlink may still be hiding them from the build, so it is necessary to ensure that the correct infrastructure is included to enable them.
[18:32:21] <ira> powerosx: SmartOS is not designed for storage as far as being a NFS/CIFS server.  To turn it into one.. you are going against the grain.
[18:32:38] <ira> They won't stop you… but the "architecture" isn't going to help you.
[18:33:32] <opeth__> jperkin: am I to hack at a buildlink.mk?
[18:33:41] <ira> If someone from Joyent thinks I'm off on my opinion, please feel free to correct me ;)
[18:33:42] <ryancnelson> zones are not vm's… your misunderstanding of that concept undermines the notion that you prefer the way that smartos or OI do things… you don't boot a full OS inside a zone.
[18:33:57] <jperkin> opeth__: no, it's odd that you're hitting this as our postfix package includes sasl by default
[18:34:06] <opeth__> options.mk has BUILDLINK_INCDIRS.cyrus-sasl= include/sasl
[18:34:13] <jperkin> unless by adding mysql support something is screwing with it
[18:34:15] <ira> ryancnelson, has it right there also.  Zones are zones… :)
[18:34:23] <jperkin> I'll try to reproduce, though it may need to wait until tomorrow
[18:34:28] <powerosx> I really enthusiastic of the smartos features
[18:34:46] <powerosx> it's a very very good work…
[18:34:49] <opeth__> jperkin: I'd be grateful even for only a mysql-enabled 64-bit postfix package itself
[18:34:50] <powerosx> but
[18:35:06] <opeth__> but in the long run of course it'd be grand to be able to roll my own stuff
[18:35:07] <powerosx> in the real world… with simple clients
[18:35:34] <powerosx> need a storage vm
[18:36:09] <powerosx> and maybe some features of virtualbox....
[18:36:13] <powerosx> hihi
[18:36:32] <powerosx> yessss.. I confess
[18:36:46] <nahamu> powerosx: so at scale, most of us would run the storage server separately.
[18:37:15] <powerosx> I want to build a openindiana zone coz I want a good storage and some feautures
[18:37:22] <powerosx> of virtualbox
[18:37:51] <jperkin> run OmniOS
[18:37:58] <nahamu> powerosx: so your choices are to run a separate OI storage server, or run OI under KVM and pay the performance penalty.
[18:38:10] <ryancnelson> seriously, your questions suggest that you don't quite understand what the "features of smartos" really are.
[18:38:15] <nahamu> (or run OmniOS and manage your virtualization the OmniOS way)
[18:38:16] <ira> jperkin: powerosx needs to understand the system he's using.  He's not making sense.
[18:38:35] <jperkin> well, that's fine, the #omnios guys can help instead ;)
[18:38:42] <powerosx> nooooo… I want to boot with usb drive and Running Without a ZFS Root Pool !!!  :D
[18:38:50] *** axonpoet has joined #smartos
[18:38:53] <powerosx> it's very faigo
[18:39:18] <nahamu> powerosx: if you do that, how do you expect to persist the configuration?
[18:39:31] <e^ipi> nahamu: zfs properties. I've done it
[18:39:41] <nahamu> e^ipi: you're not helping me. ;)
[18:39:44] <e^ipi> smartos works fine as a storage / NAS server.
[18:39:48] <e^ipi> no, i'll do that
[18:40:09] <nahamu> e^ipi: do you have the CIFS component joined to a domain?
[18:40:21] <nahamu> If so, leecallen would love to know how you persist that across reboots.
[18:40:26] <nahamu> (as would I)
[18:40:32] <e^ipi> no, i don't feel any need for microsoft protocols. my home is all mac & UNIX
[18:41:09] <nahamu> right, as a relatively vanilla NFS server, SmartOS seems like it would be fine.
[18:41:09] <ira> e^ipi: As an NFS server… SmartOS works well.. I'd agree there.
[18:41:31] <powerosx> smartos was skimmed of all components of a storage server
[18:41:52] <nahamu> powerosx: which components? NFS is there CIFS is there.
[18:41:55] <ira> Nah… it has the NFS bits… and the CIFS bits I think.
[18:42:01] <powerosx> just build a openindiana inside a zone
[18:42:19] <powerosx> and that's it
[18:42:26] <powerosx> haha
[18:42:35] <nahamu> http://local-static1.forum-files.fobby.net/forum_attachments/0020/3336/You_keep_using_that_word_large.jpg
[18:42:51] <powerosx> guyyyys .. I know it is possible
[18:42:59] <ryancnelson> you are wrong.
[18:43:15] <nahamu> powerosx: then *do* it, and then write up how to do it and show us.
[18:43:29] <ryancnelson> i'm done here.
[18:43:31] *** ryancnelson has left #smartos
[18:43:53] <opeth__> :D
[18:43:55] <nahamu> but expecting help from people who are saying "it can't be done" is a bit silly.
[18:43:59] <powerosx> I can't…. I am not a good programmer like youuuuu
[18:44:07] <opeth__> it's not easy to handle people like that.
[18:44:39] <nahamu> powerosx: I have neither the time nor the desire to work on that, nor do any of the actual developers.
[18:44:47] <powerosx> I just ask if someone helps meeee
[18:45:56] <nahamu> from the department of using SmartOS as intended...
[18:45:58] <rmustacc> It is technically possible to get an ipkg zone working, however that doesn't solve the NAS issues because neither the NFS server or CIFS server run in a zone.
[18:46:09] <nahamu> has anyone else been having trouble creating their own template images (for KVM) on the latest version of the platform? I'm getting "Invalid value(s) for: disks.*.image_uuid,disks.0.image_uuid" errors...
[18:46:52] <szaydel> rmustacc: It really would be awesome to have access to NFS server in a zone!
[18:46:59] <nahamu> the only odd thing I did was instead of simply zfs send / receive, I created the target zones/uuid zvol in advance and then set compression=lz4 before doing a zfs receive -F
[18:47:44] <nahamu> szaydel: all of illumos agrees, yet no one is paying a developer to do it... :)
[18:47:52] <powerosx> really????? nfs and cifs not run in a zone??
[18:48:11] <powerosx> :(
[18:48:20] <szaydel> nahamu: I am going to convince my management to do it.
[18:48:31] <nahamu> szaydel: !!!
[18:48:36] <jperkin> samba runs in a zone, just not the native cifs server
[18:48:37] <rmustacc> powerosx: The server cannot.
[18:48:49] <rmustacc> sazydel: Yes, it'd be nice.
[18:48:51] <szaydel> Samba is samba...
[18:49:09] <szaydel> Not that I am a fan of CIFS, but native CIFS would also be awesome in a zone.
[18:49:29] <jperkin> sure, get working on it ;)
[18:49:45] <szaydel> Alright, on it already.
[18:49:48] <jesse_> while you're at it, add native afs, too?=)
[18:50:04] <opeth__> I'm also missing native nfs
[18:50:08] <jesse_> (apple- not andrew-)
[18:50:09] <szaydel> Skip AFS, but maybe gluster?
[18:50:10] <opeth__> migrating off s11 sure isn't easy damnit
[18:50:16] <powerosx> so with samba.. there is no problem?
[18:50:29] <szaydel> Samba will compile and run.
[18:50:34] <powerosx> good
[18:50:44] <szaydel> OR, "should" compile and run.
[18:51:02] *** denizr has quit IRC
[18:51:10] <rmustacc> It does, ira can atest to that.
[18:51:16] <jperkin> http://www.perkin.org.uk/posts/setting-up-samba-on-smartos.html
[18:51:19] <jperkin> is what I use @ home
[18:51:22] <ira> It does.
[18:51:23] <jesse_> hmmm, no samba in pkgsrc?
[18:51:49] <ira> At least our most recent deployment says it does.
[18:51:54] <jesse_> (no, I don't need it, just checked out of curiosity)
[18:51:55] <powerosx> yes yessss… it's possible
[18:52:13] <nahamu> jperkin: did you ever get a chance to play with those patches from noahmehl for Samba 4 for pkgsrc?
[18:52:16] <jperkin> jesse_: yes, but standard images didn't (still don't? not sure) include it
[18:52:21] <powerosx> so … a mystic openindiana zone… it's possible
[18:52:21] <jesse_> ira, I'm looking at the trunk
[18:52:39] <rmustacc> powerosx: You need to really clarify your terminology here.
[18:52:50] <jesse_> keeps jperkin happier when I always first do pkgin -y update before asking stupid questions=)
[18:52:53] <rmustacc> Do you meant that you want an ipkg branded zone?
[18:52:56] <jperkin> nahamu: no, though a netbsd dev is working on it too
[18:52:59] <powerosx> sorry for my english
[18:53:08] <powerosx> I'm just learning
[18:53:11] <powerosx> hihi
[18:53:17] <rmustacc> It's fine, just trying to make sure you're talking about the right thing.
[18:53:36] <rmustacc> There is no way to take your current OI system and magically turn it into a zone. The only way for that is through hardware virtualization.
[18:53:58] <rmustacc> Given that you don't want to write code and no one in here wants to write it or keep it up, an ipkg branded zone is really unlikely to happen.
[18:54:03] <ira> jperkin: Is someone talking to Samba Team about upstreaming whatever's wrong?
[18:54:26] <jperkin> ira: I haven't looked.
[18:55:00] <ira> jperkin: Thanks.  If there's an issue, please at least inform me… I'm not too hard to find ;)
[18:55:10] <jperkin> sure
[18:55:18] <powerosx> is it really impossible to do a openindiana's zone  ?? with smartos
[18:56:33] <jesse_> powerosx, there are two different virtualizations in smartos, 'zones', which are virtual smartos instances, and kvm, which is full hardware virtualization
[18:56:54] <nahamu> szaydel: is RackTop's stuff also built on illumos?
[18:57:08] <powerosx> ok, sorry… while I translated my question… you wrote the answer
[18:57:15] <jesse_> powerosx, running openindiana in a 'zone' doesn't make sense, because it's always running on smartos kernel
[18:57:37] <szaydel> nahamu: It depends, some is still Nexenta based, new stuff we are doing is Illumos.
[18:57:51] <jesse_> powerosx, but you can run a whole virtualized hardware with kvm and install OI there - you'll just get a big performance hit
[18:57:56] <rmustacc> It's not technically impossible to get an ipkg zone which is the kind of zone opendindiana uses, but it'd be a pain.
[18:58:05] <jesse_> powerosx, and some limitations
[18:58:07] <rmustacc> Most people who started going down that path stopped.
[18:58:12] <ira> rmustacc: It'd be easier to just install the needed ports. :)
[18:58:16] <powerosx> but smartos kernel is not Illumos kernel??
[18:58:20] <jesse_> rmustacc, and keeping the kernel/ipkgs in sync would be pain, like someone said
[18:58:37] <rmustacc> powerosx: Well, smartos is based on illumos, yes.
[18:58:37] <jesse_> powerosx, it is
[18:58:43] <rmustacc> But the bits we ship and the bits that OI ships are different.
[18:58:54] <nahamu> I thought I saw some feature overlap w/Nexenta. Interesting.
[18:59:07] *** Lumb has quit IRC
[19:00:26] <szaydel> nahamu: For older systems we augment Nexentastor with Ecosystem specific stuff, build our own API, etc.
[19:00:29] <powerosx> **curse** … But mine was a good idea ..hihi
[19:00:33] *** Lumb has joined #smartos
[19:00:45] *** bluezenix has quit IRC
[19:02:20] <powerosx> do you think that I have good performance with openindiana inside kvm
[19:02:24] <powerosx> ?
[19:02:38] <jesse_> bad, compared to running samba in a zone
[19:02:47] <jesse_> or nfs sharing from the global zone
[19:03:30] <powerosx> what do you mean nfs sharing from the global zone???
[19:04:27] <rmustacc> The only way that the nfs server works today is through the global zone.
[19:04:56] *** Trixboxer has joined #smartos
[19:04:57] <jesse_> configuring it is a bit tricky, as the global zone is not persistent
[19:05:02] *** Lumb has quit IRC
[19:05:20] <ira> (nod) It is a bit of a hack… but, if you're going to do this… you're going to do worse ;)
[19:06:06] <ira> You'd almost think there wasn't a few companies out there selling ZFS based NAS :)
[19:06:23] <szaydel> Yeah, like one or two. :)
[19:06:46] <Melinko> Anyone had any luck on getting a multi-display working for KVM vm's?
[19:07:33] <nahamu> Melinko: I haven't really, so I doubt anyone else has.
[19:07:44] <nahamu> (I assume you're talking multi-head spice)
[19:07:50] <Melinko> ;)
[19:08:05] <nahamu> in theory you should only need to add an extra "-device qxl" to the qemu_extra_args
[19:08:12] <Melinko> I tried -qxl 4 and adding "qxl": "4", the -qxl wouldn't make a bootable VM, "qxl": "4" does
[19:08:16] <nahamu> but I'm not sure how well it works in practice.
[19:08:20] <yofuh> jperkin: the install script of nginx is currently broken, it tries to create user and group www with uid/gid 847
[19:08:36] <nahamu> oh!
[19:09:07] <nahamu> yeah, try appending an extra "-device qxl" to your existing qemu_extra_opts and let me know if that works.
[19:09:07] <Melinko> However even as bootable
[19:09:17] <Melinko> Ok
[19:09:27] *** Lumb has joined #smartos
[19:09:31] <jperkin> yofuh: ah yes, with 2012Q4 we moved to statically-assigned uid/gid
[19:09:33] <Melinko> let me see how that works, ill make a Json for it right now
[19:09:37] <jperkin> if you already have users set up it will break
[19:10:25] <jperkin> pkgsrc now starts at uid=999 and works down, thus useradd will start at 1000 and we avoid conflicts.
[19:11:48] <Melinko> Creating
[19:12:05] <ira> jperkin: Is there a file we can provide pkgsrc to say "Don't use this UID.  Please…" We have a legacy UID in a really dumb place :/
[19:12:18] <yofuh> jperkin: www has uid/gid 80 for ages and it is still true in smartos zones
[19:12:19] <nahamu> Melinko: at work, we're currently just running the fedora VMs at 3840x1200 resolution and use something we found called fakexinerama to make the window managers treat it as two separate displays.
[19:12:31] <jperkin> ira: no, you'd need to build your own
[19:12:47] <ira> Or just do what we do now.. and suffer. ;)
[19:13:03] <jperkin> yofuh: right, that's why the recommended thing is always to provision with the matching image for the package set, as that uid is removed in e.g. base1.9x
[19:13:03] <jesse_> jperkin, are the uids decided statically at build time and not dynamically when installing the package?
[19:13:09] <nahamu> and the windows vms just seem to max out at 1920x1200, which I just accept since there are so few Windows VMs.
[19:13:57] <jperkin> jesse_: right
[19:14:05] <yofuh> jperkin: you might agree or not agree, but reinstalling the full zone everytime you want to update, may sound silly to some peaople
[19:14:22] <jesse_> jperkin, oh, I thought there was a script that iterated down from 999 and used the first free one=)
[19:14:28] <jesse_> when installing
[19:15:23] <jperkin> yofuh: sure, in an ideal world it wouldn't be necessary, but with the nature of pkgsrc it's the only sensible one - plus it has similar advantages to the GZ design, i.e. it forces you into proper config management
[19:16:01] <nahamu> Melinko: did that work do add an additional display?
[19:16:05] *** opeth__ has quit IRC
[19:16:20] <nahamu> (windows or linux guest?)
[19:16:38] <jperkin> jesse_: no, the point is to ensure that the same uid/gid is used across all your zones and not randomly allocated.
[19:16:49] <nahamu> (I think the windows driver handles multiple qxl devices much better than the linux drivers do)
[19:17:11] <Trixboxer> hi, is http://wiki.smartos.org/display/DOC/Installing+pkgin applicable ?  my smartos does not have pkg_admin command
[19:17:13] <ira> Actually, instantiating the zone on install to upgrade is great… it makes sure people don't screw up :)
[19:17:24] <jesse_> jperkin, kind of makes sense. But then you need to keep track of the uids somewhere so that new sets always get the same uids
[19:17:27] <yofuh> jperkin: if you ask me, failing to continue to use existing users on package upgrades, is much more far away from a perfect worls, then necessary
[19:17:39] <jesse_> I don't really care either way
[19:17:41] <jperkin> jesse_: yes, we have a static list
[19:17:53] <yofuh> what is the problem just to skip useradd in the case the user is already there?
[19:17:58] <jesse_> just explains an empty id 997 or something in my zone=)
[19:17:58] <rmustacc> This is just like the Debian upgrade the most recent Debian upgrade.
[19:18:00] <jperkin> and each time a package is added to pkgsrc which needs to create users we add it to the list
[19:18:00] <rmustacc> It's unfortunate.
[19:18:06] <rmustacc> but you have to bite the bullet at some point.
[19:18:50] <jperkin> yofuh: I'll agree that it shouldn't fail quite as hard as it does right now, I need to figure out which bits actually need the hardcoded id's rather than using the names
[19:18:51] <rmustacc> I can type sentences that parse, I promise.
[19:18:55] <mamash> yofuh: we had to make this decision once. we've seen customers hit the other side of the coin as they tried to maintain static UID/GID mapping for their real users across systems, and were constantly bitten by the random package users created by pkgsrc
[19:18:57] <ira> If you upgrade your stuff in a separate zone… you can always run them side by side. :)
[19:19:17] <ira> (or do your setup/test, while the previous version is still running.)
[19:19:21] *** ahaydock has joined #smartos
[19:20:56] <jperkin> yofuh: as when using completely dynamic uid/gid I don't believe there are problems with pre-creating the users.
[19:24:05] *** Lumb has quit IRC
[19:24:08] <yofuh> it will likely to be problematic to let the uid randomly choosen if systems share a nfs for example, so from pkgsrc perspective, ther should be an option to let the user set their uid instead of forcing specific or unspecifig uids
[19:24:27] <yofuh> for the moment a manual pkg_add -I did the job but still...
[19:25:20] <yofuh> and btw, i miss nginx-passenger in Q4, but that might be on purpose
[19:28:58] *** Lumb has joined #smartos
[19:30:38] *** bradleymeck has joined #smartos
[19:31:04] *** denizr has joined #smartos
[19:32:33] *** AlainODea has quit IRC
[19:33:53] *** AlainODea has joined #smartos
[19:37:23] *** bradleymeck has quit IRC
[19:38:11] *** bradleymeck has joined #smartos
[19:39:47] *** opeth__ has joined #smartos
[19:47:10] *** avrntsv has joined #smartos
[19:48:38] *** mamash has left #smartos
[19:52:10] *** mamash has joined #smartos
[19:58:30] *** powerosx has left #smartos
[20:00:23] *** bradleymeck has quit IRC
[20:01:16] *** ipalreadytaken has joined #smartos
[20:05:09] *** bradleymeck has joined #smartos
[20:08:52] *** richardshaw has quit IRC
[20:17:33] *** AlainODea has quit IRC
[20:19:21] *** tat- has quit IRC
[20:25:49] *** AlainODea has joined #smartos
[20:26:48] *** ryancnelson has joined #smartos
[20:26:49] *** tonyarkles has quit IRC
[20:28:03] *** Webhostbudd has quit IRC
[20:30:24] *** bens1 has quit IRC
[20:31:24] *** ipalreadytaken has quit IRC
[20:35:17] *** ipalreadytaken has joined #smartos
[20:48:48] *** AlainODea has quit IRC
[20:49:44] <Trixboxer> Where can I check VM boot errors ?  A  VM is stuck while booting up. VNC shows "Booting from Hard Disk"
[20:54:58] <EMH_Mark3> /zones/zoneuuid/root/tmp/vm.log
[20:55:50] *** Meths has joined #smartos
[20:57:34] <Trixboxer> it says "Booting from 0000:7c00"
[20:57:44] <Trixboxer> EMH_Mark3: and its stuck there
[20:58:06] <EMH_Mark3> unfortunately that's what my working VM says... so can't really help past that.
[20:59:08] <rmustacc> What kind of guest are you booting?
[20:59:22] <Trixboxer> EMH_Mark3: I started other machine ( windows ) which is working
[20:59:27] <Trixboxer> rmustacc: its centos
[20:59:31] <Trixboxer> centos5
[21:01:37] <ryancnelson> sounds like your vm is running just fine, but your OS is messed up
[21:02:52] <Trixboxer> it was fine, and nothing has been done to make it not boot
[21:03:03] <rmustacc> So did you actually get to grub?
[21:03:18] <rmustacc> And/or did grub start booting from the disk image?
[21:03:39] <rmustacc> Though I do know that centos 5 virtualization has had occasional issues due to the rather old Linux kerel in use.
[21:05:34] <Trixboxer> rmustacc: I dont get grub
[21:05:57] *** ryancnelson has quit IRC
[21:07:27] <rmustacc> Really? That almost sounds like your disk image is messed up then.
[21:07:48] <rmustacc> Generally people just have really low timeouts with grub and by the time they get attached to the serial console, it's selected a default boot option.
[21:12:45] <opeth__> it's absolutely blasphemous to run this solaris 11 image on smartos that I'm replicating a zfs dataset with a zone into
[21:12:56] <opeth__> I'd even enjoy it if only it knew virtio
[21:13:20] *** ryancnelson has joined #smartos
[21:25:58] *** ipalreadytaken has quit IRC
[21:28:49] <jesse_> doesn't smartos have virtio drivers?
[21:29:06] <jesse_> (they might work in sol11=))
[21:32:23] *** ryancnelson has left #smartos
[21:38:46] *** KermitTheFragger has quit IRC
[21:42:01] <Trixboxer> rmustacc: checking more :(
[21:42:18] <Trixboxer> rmustacc: fortunately I had a clone of it of 7 days back
[21:42:26] <Trixboxer> which is working fine
[21:42:44] <Trixboxer> rmustacc: sorry, the OS was centos6
[21:50:59] *** Forced has joined #smartos
[21:59:42] *** bradleymeck has quit IRC
[22:08:30] *** siezer has quit IRC
[22:09:45] *** enmand has quit IRC
[22:15:53] *** kaladis has joined #smartos
[22:17:03] *** denizr has quit IRC
[22:19:15] *** orangeroo has left #smartos
[22:21:02] <kaladis> What do I have to do, so I can create a zpool within a zoned smartos instance? I get a "permission denied" error
[22:21:51] <rmustacc> You generally can't create a zpool inside of a zone.
[22:22:41] *** bens1 has joined #smartos
[22:26:17] *** enmand has joined #smartos
[22:27:53] <jesse_> kaladis, you can delegate a pool/zfs to a zone
[22:28:16] <jesse_> zpool commands won't still work on it, though
[22:28:52] <kaladis> Okay, good to know
[22:28:55] <jesse_> and vmadm doesn't do delegation after the zone has been created
[22:28:55] <kaladis> What I
[22:29:21] <kaladis> What I'm trying is to create encrypted files via lofiadm and then creating a zpool out of that for encrypted storage
[22:29:23] <jesse_> what I've done, is to use zonecfg to delegate a pool to the zone
[22:29:29] <jesse_> seems to work for me, ymmv=)
[22:31:10] <jesse_> who was it yesterday who was arguing that zfs encryption is not needed?=)
[22:31:14] <wesolows> you might be able to create the encrypted loopback device, pool, and a filesystem within it in the GZ and then pass the filesystem into the zone as a lofs mountpoint.
[22:31:54] <wesolows> jesse_: it wasn't me, but I've said many times that no one has made a business case for doing it.
[22:32:03] <wesolows> not needed != no business case
[22:32:20] <kaladis> Thats what I'll be trying to do then
[22:32:34] <jesse_> I'd say no-one is willing to pay it to be implemented
[22:32:41] <jesse_> there are business cases, too=)
[22:32:43] *** olafm has joined #smartos
[22:32:59] <olafm> Hi.
[22:33:09] <wesolows> kaladis: yeah, at least all of those steps are known to work individually.
[22:33:42] <jesse_> lofs mounts have same problem as the delegation - you can't do it with vmadm after creating the zone
[22:33:53] <jesse_> but zonecfg works with that, too=)
[22:34:29] <wesolows> I was more thinking that delegation doesn't allow you to specify the dataset, but if you're using zonecfg anyway then that presumably works.
[22:34:44] <wesolows> at least the lofs mount works with vmadm *at all*.
[22:34:46] *** tonyarkles has joined #smartos
[22:35:17] <kaladis> Alright ^^
[22:35:43] *** wolfeidau has quit IRC
[22:35:51] <olafm> So installing any software into the global zone seems to be a frowned upon. Then how should I monitor the health of my machine? From inside a dedicated monitoring VM?
[22:36:34] *** ipalreadytaken has joined #smartos
[22:36:36] <wesolows> that's an approach, sure.
[22:37:43] <jesse_> olafm, I'd just 'install' monitoring software to gz:/opt
[22:37:56] <jesse_> with matching autoloaded smf manifest
[22:38:05] <wesolows> the question is what you're trying to "monitor"
[22:38:32] <wesolows> virtually all failures are software or config bugs, meaning they'll be affecting things specific to a given zone
[22:38:38] <olafm> wesolows: Hardware utilization and the running zones.
[22:38:44] <wesolows> which means you want to monitor those things like you would any other machine
[22:39:09] <olafm> Yeah, I have monitoring scripts running inside each zone as well.
[22:39:45] <olafm> But I would like to track the usage of my /zones pool for example.
[22:40:07] <linuxprof> what monitoring system are you using?
[22:40:48] <olafm> In the zones I am running collectd to collect statistics and thewn have custom scripts that check the collected data.
[22:40:57] *** ipalreadytaken has quit IRC
[22:41:14] <linuxprof> okok
[22:41:22] <linuxprof> im running NRPE in the GZ
[22:41:29] <olafm> All very low-tech but I only have one machine and about 7 zones.
[22:42:52] <olafm> Thanks for the input. I'll play around with installing collectd and some scripts in /opt with appropriate SMF manifests.
[22:44:48] *** bens1 has quit IRC
[22:46:30] *** vsomes has joined #smartos
[22:46:33] <Trixboxer> Hi, I'm unable to boot my centos6
[22:46:40] <Trixboxer> there is an error in "messages"
[22:46:40] <Trixboxer> unhandled wrmsr: 0x81024ab0 data fffffd7fffdfe810
[22:46:43] *** enmand has quit IRC
[22:47:01] <Trixboxer> kvm: [ID 391722 kern.info] unhandled wrmsr: 0x0 data 0
[22:53:38] *** bradleymeck has joined #smartos
[22:56:42] <yofuh> hm, any advise how to use a second ip adress on one interface in the gz? just creating a second nic tag on the same mac adress or will that fail?
[22:57:14] *** wolfeidau has joined #smartos
[22:57:36] <jesse_> 'second nic on the same nic_tag'
[22:57:43] <jesse_> and it should work
[22:58:18] *** Cpt-Oblivious has joined #smartos
[22:58:19] <yofuh> what do you mean with 'second nic on the same nic_tag'
[22:58:54] <jesse_> ah, 'in the gz', missed that part
[23:00:11] *** neophenix has quit IRC
[23:00:17] <yofuh> wiki shows something like <nic_tag><instance#>_ip so could it be admin_ip and admin_2_ip?
[23:03:14] <rmustacc> admin_ip is special.
[23:03:25] <rmustacc> But generally you want external0, external1.
[23:04:41] <yofuh> that server does only have one physical nic and some special requirements (so called root-server which means a rented dedicated server with no physical access)
[23:05:27] <yofuh> si i go for admin1_ip and hope that this will work...
[23:05:56] *** daleg has quit IRC
[23:22:08] *** denizr has joined #smartos
[23:29:17] *** bluezenix has joined #smartos
[23:37:44] *** szaydel has quit IRC
[23:37:50] *** ktk is now known as ktkNA
[23:39:58] *** andoriyu_ has joined #smartos
[23:40:05] <andoriyu_> hey
[23:40:09] *** andoriyu__ has joined #smartos
[23:40:10] *** axonpoet has quit IRC
[23:40:37] <andoriyu_> how do i update smartos that is running inside guest?
[23:41:00] <e^ipi> what do you mean by that+
[23:41:27] <e^ipi> pkgsrc, or the base system
[23:42:05] <andoriyu_> I have zone that was installed long long time ago
[23:42:17] <andoriyu_> so. I guess I need to update base system and pkgsrcf
[23:42:38] <e^ipi> the base of the system is the same as the global zone
[23:42:49] <e^ipi> pkgsrc, is more difficult. your best bet is to just recreate the zone
[23:43:19] <andoriyu_> hmm
[23:43:32] *** Licenser has quit IRC
[23:43:36] <e^ipi> i mean, you *can*, but it's a total pain
[23:43:54] <e^ipi> not worth it. it'll be less work to migrate your data to a new zone with whatever the new dataset you want to use is
[23:43:57] <andoriyu_> more pain than configuring eveything again?
[23:44:01] <e^ipi> yes
[23:44:19] <jperkin> yes, pkgsrc does not guarantee any compatability between branches
[23:44:22] <e^ipi> but ask yourself why you want to move to the new pkgsrc in the first place
[23:44:36] <jperkin> it's effectively a major release every quarter
[23:44:38] <e^ipi> is your application working? yes? then don't mess with it.
[23:44:40] <andoriyu_> emm, I need new versions of packages
[23:45:14] <e^ipi> no you don't.
[23:45:24] <e^ipi> you may want specific features and bugfixes, but newer != better
[23:45:39] <andoriyu_> that is what i mean by newer version
[23:45:45] <e^ipi> a lot of times newer just means you get new bugs that you might actually be hitting, as opposed to old ones that don't affect you
[23:45:52] <andoriyu_> Server: nginx/1.0.14 is not cool at all
[23:46:22] <jperkin> best is always to keep track of what you apply on top of a base OS and then have a procedure where you can re-deploy it on top of a new image and test everything still works
[23:47:03] <andoriyu_> I know. Wasn't following that pattern.
[23:47:21] <andoriyu_> mostly because chef-client wasn't working on smartos
[23:47:21] <e^ipi> ugh
[23:47:25] *** andoriyu__ has left #smartos
[23:47:26] <e^ipi> i hate the word "pattern"
[23:47:45] <e^ipi> it reeks of clueless project manager
[23:47:48] *** hungryblank has quit IRC
[23:48:06] <e^ipi> not a criticism of you, it's a visceral reaction
[23:48:18] <andoriyu_> I have the same opinion
[23:48:38] <andoriyu_> maybe it is better to rebuild whole system.
[23:49:07] <jesse_> ...and document it this time=)
[23:49:16] <andoriyu_> everything is documented :)
[23:49:38] <andoriyu_> That server that is a problem is just http router.
[23:50:14] <andoriyu_> probably better to make new one with kvm guest.
[23:50:49] <jesse_> if you now have a zone, no
[23:51:00] <jesse_> zone is way faster
[23:52:23] *** bluezenix has quit IRC
[23:55:44] <andoriyu_> 9T19:43:00Z
[23:55:44] <andoriyu_> 03d515f2-872c-11e2-8314-8387b5044570  base           1.9.0    smartos  2013-03-07T15:11:00Z
[23:55:44] <andoriyu_> is the latest one, ight?
[23:58:03] *** trentster has joined #smartos
top

   March 18, 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 | 31 | >