Switch to DuckDuckGo Search
   March 21, 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
[00:00:01] <miine> ira: the computer can't turn itself on too...
[00:00:39] <miine> olafm: so you have the rights to read "secure paper x". so your web browser can too...
[00:00:50] *** deirdres has quit IRC
[00:00:50] *** deirdres_ is now known as deirdres
[00:00:51] <ira> miine: Depends on how you set it up.  Some setup their servers to turn on, when power is removed ;)
[00:00:58] <ira> and restored...
[00:01:05] <ira> So… In fact, the machine will come up.
[00:01:40] <miine> there IS already a linux based solution which does exactly that using virtualized linux instances... .
[00:01:41] <olafm> miine: No. ssh gets to read ~/.ssh/id_rsa but firefox does not get access.
[00:01:58] <olafm> Just as an example.
[00:02:51] <miine> olafm: than you're at that famous sandbox model. as apple. works great there too :-(
[00:04:09] <olafm> miine: MAC is not a sandbox. Its just a much finer grained form of access control. Your zone system on the other hand sounds very much like a sandbox.
[00:05:58] <miine> and how to you tell - lets say that bash can't access ~/.ssh/id_rsa sometimes and sometimes it can? depending on the task you've to do?
[00:07:20] <miine> you would have to fork it from a "task/security" process. thats whats zones are for... .
[00:08:56] <olafm> I don't understand the last sentence.
[00:10:32] <miine> my understanding is that MACs are assigned to processes/forks or binaries. so you can tell (as in your example) that firefox can't access some files.
[00:11:03] <miine> if a tool needs sometime access to data, but it sometimes shouldn't - how is that solved?
[00:11:38] <olafm> So you want to temporarily elevate the program/processes priviledges?
[00:11:58] <miine> like firefox: it should be able to access internal files if connected to the internal ERP, but it shouldn't if connected to the world...
[00:12:41] <miine> olafm: if the use is able to click "grant rights" - "OK" - what will happen??
[00:12:47] <miine> use -> user
[00:13:42] <miine> the idea of the trusted extensions stuff was right. thats why others did implement it (albeit on a different level) again...
[00:13:48] <olafm> miine: I am _not_ a MAC expert, but I believe for your firefox/ERP scenario you would want to firefox "instances". One that can access local files and the local network/ERP system and one that only has access to internet.
[00:14:39] <olafm> Solaris Trusted Extensions / Trusted Solaris is a MAC solution if I am not mistaken (I have only dealt with SELinux).
[00:14:54] <miine> olafm: so you would have two firefox binaries.... . it is much easier to have them in zones and show the user (window color) which zone the process is in...
[00:15:22] <olafm> miine: Not necessarily two binaries. Just two running processes with differen labels.
[00:16:15] <miine> olafm: so either you've to write a wrapper to start the "right" one, or you've to patch firefox do to it "automagically" depending on the URL...
[00:16:21] *** andoriyu has quit IRC
[00:16:48] <olafm> miine: Yes.
[00:17:18] <miine> and even then: the user can't distinguish which window is in which security model...
[00:17:56] *** axonpoet has joined #smartos
[00:18:55] <miine> it;s also easy to limit access (say to web pages) on zones. so it would be able to block online banking everywhere but a high security zone...
[00:20:29] <miine> IF Apple would be able to do zones - they would. but they can't because of their display model which can only show processes of one user&root...
[00:20:48] <olafm> miine: I'm not saying zones cannot provide this. I'm only suggesting that it would likely end up replicate what is in the Trusted Extensions. Granted, there are no TX in Illumos.
[00:22:11] <miine> some of the TX technology is needed (X11, monitoring/disabling cut/copy/paste, etc.). but the rest would be easier to be implemented in zones. and to maintain too...
[00:25:53] *** deirdres has quit IRC
[00:26:10] *** deirdres has joined #smartos
[00:28:34] *** szaydel has joined #smartos
[00:29:44] *** Melinko has quit IRC
[00:30:32] *** denizr has joined #smartos
[00:41:54] *** khushildep has quit IRC
[01:00:20] *** ni291187 has joined #smartos
[01:02:56] *** codingstream1 has joined #smartos
[01:04:10] *** codingstream has quit IRC
[01:10:45] *** ni291187 has quit IRC
[01:14:21] *** szaydel has quit IRC
[01:28:14] *** emocakes has joined #smartos
[01:37:04] *** tonyarkles has quit IRC
[01:49:59] *** denizr has quit IRC
[01:50:20] *** CarlosC has quit IRC
[01:56:03] *** tonyarkles has joined #smartos
[01:58:39] *** ryancnelson has quit IRC
[02:01:25] *** denizr has joined #smartos
[02:01:56] *** codingstream1 has quit IRC
[02:02:35] *** codingstream has joined #smartos
[02:04:31] *** dap has quit IRC
[02:09:19] *** danielwu has joined #smartos
[02:13:00] *** porkbelt has quit IRC
[02:15:01] *** porkbelt has joined #smartos
[02:15:24] *** danielwu has quit IRC
[02:20:17] *** codingstream has quit IRC
[02:20:40] *** codingstream has joined #smartos
[02:26:20] *** szaydel has joined #smartos
[02:27:14] *** enmand_ has joined #smartos
[02:30:12] *** enmand has quit IRC
[02:36:12] *** emocakes has quit IRC
[02:40:22] *** deirdres has quit IRC
[02:49:45] *** dysinger has quit IRC
[02:59:48] *** jim80net has quit IRC
[03:01:02] *** ira has quit IRC
[03:02:21] *** emocakes has joined #smartos
[03:10:20] *** szaydel has quit IRC
[03:13:06] *** darjeeling has joined #smartos
[03:26:48] *** denizr has quit IRC
[03:31:01] *** dysinger has joined #smartos
[03:34:37] *** darjeeling has quit IRC
[03:36:51] *** miine_ has joined #smartos
[03:38:08] *** miine_ has quit IRC
[03:38:26] *** miine_ has joined #smartos
[03:39:18] *** miine has quit IRC
[03:39:18] *** miine_ is now known as miine
[03:46:08] *** arai_ has joined #smartos
[03:47:29] <arai_> Hey, I'm getting the following error when I attempt to deploy a new zone on smartos 20130307T214308Z: Command failed: Brand error: missing parent ZFS dataset for /zones
[03:48:10] <arai_> second line: cannot open '/bad2face-8738-11e2-ac72-0378d02f84de': invalid dataset name
[03:48:32] *** deirdres has joined #smartos
[04:01:53] <arai_> Restarted the machine, the problem is gone :-(
[04:16:36] *** ipalreadytaken has joined #smartos
[04:23:09] <richlowe> I think that would happen if either zones didn't import, or (maybe) more likely, the root dataset of the zones pool didn't mount.
[04:23:31] <richlowe> the latter is probably the kind of thing that'd be subtle enough to have it happen without it being immediately clear why, at least.
[04:35:32] *** tonyarkles has quit IRC
[04:39:20] <arai_> Thanks.  Also, I'm curious if it was a design decision to break ZFS zone delegation?
[04:39:53] *** deirdres has quit IRC
[04:39:57] <arai_> or do I need to read more?
[04:41:47] *** enmand_ has quit IRC
[04:45:08] *** axonpoet has quit IRC
[04:46:55] *** xinkeT has quit IRC
[04:48:57] <jesse_> 'works for me'
[04:49:38] *** AlainODea has quit IRC
[04:51:07] *** axonpoet has joined #smartos
[04:54:36] <arai_> jesse: you're getting something other than "no datasets available" when using zfs within a zone?
[04:55:46] *** ChrisPartridge has joined #smartos
[04:56:21] <arai_> bbl
[04:56:25] *** arai_ has left #smartos
[04:57:38] <rmustacc> arai_: By default the dataset isn't delegated.
[04:57:46] <rmustacc> Oh, he's gone.
[04:58:11] <ChrisPartridge> I've installed the postgresql92-server-9.2.2nb1 package from pkgsrc in a zone, however can't seem to find a manifest for svcadm - can anyone point me in the right direction?
[05:02:23] *** sachinsharma has joined #smartos
[05:13:24] <LeftWing> ChrisPartridge: svcs -a | grep postgres ?
[05:14:35] <LeftWing> I would expect it to be something like "svc:/pkgsrc/postgresql:default"
[05:14:56] *** codingstream has quit IRC
[05:16:37] <ChrisPartridge> LeftWing: Yeah, that was my first guess as well - nothing there for postgres though
[05:18:34] <LeftWing> ChrisPartridge: What image is this?
[05:19:30] <ChrisPartridge> LeftWing: bad2face-8738-11e2-ac72-0378d02f84de smartos 2013-03-07 sdc:sdc:base64:1.9.0
[05:21:59] * LeftWing spins up a zon
[05:22:04] <LeftWing> ..e
[05:22:44] <LeftWing> Ah, it does, in fact, seem to be missing.
[05:22:54] *** arai has joined #smartos
[05:23:40] *** axonpoet has quit IRC
[05:23:53] <ChrisPartridge> phew, so I'm not going crazy
[05:24:09] <LeftWing> I didn't say that. :)
[05:24:39] <LeftWing> jperkin: Sir?
[05:24:43] <arai> That's usually the last thing I think before I do something crazy
[05:26:41] <ChrisPartridge> pkgin provides postgresql92-server-9.2.2nb1 - seems to only list 1 file, weird
[05:27:16] <rmustacc> arai: You were asking about dataset delegation, right?
[05:27:31] <LeftWing> ChrisPartridge: Well, I would have a word with jperkin when he's about.
[05:27:56] <LeftWing> I would also try sdc:sdc:base:1.9.0 instead of base64, if you're in a hurry.
[05:27:59] <arai> rmustacc: I was, yes
[05:28:17] <rmustacc> arai: The dataset that is used to form the base of your vm is never delegated.
[05:28:30] <ChrisPartridge> LeftWing: Will do - thanks for the help :)
[05:28:32] <rmustacc> However you can use vmamd to delegate additional ZFS filesystems.
[05:28:39] <LeftWing> ChrisPartridge: You're welcome!
[05:29:44] <rmustacc> So hopefully that clarifies what's going on.
[05:29:47] *** denizr has joined #smartos
[05:30:30] <arai> rmustacc: this seems to address several major issues I've been having with SmartOS so far, thank you
[05:34:20] <rmustacc> arai: See the delegate_datset option in vmadm.
[05:34:44] *** axonpoet has joined #smartos
[05:37:09] <arai> rmustacc: I do, does this allow you to specify an arbitrary dataset or is it just a boolean?
[05:37:26] *** dysinger has quit IRC
[05:38:02] <arai> testing time...
[05:39:26] <rmustacc> I think at this time it may only be a boolean, but that may change in the future.
[05:39:50] <arai> also, it looks like it's on create only
[05:40:50] <jesse_> *unsupported* you can edit them with zonecfg
[05:41:24] <arai> jesse_: I tried to while I was gone, no dice....
[05:41:25] <rmustacc> Correct, the manual page describes all of that.
[05:41:26] <jesse_> I've used zonecfg to delegate whole pools to a zone
[05:41:42] <arai> Then again I think I've been at this for about 10 days
[05:41:56] <arai> zonecfg ... ooh
[05:42:36] <jesse_> zonecfg -z <uuid> ; add dataset ; set name=<zfsname> ; end ; verify ; commit ; exit
[05:42:43] <jesse_> (where ; is enter)
[05:43:38] <jesse_> you can add lofs mounts the same way *hint*
[05:44:04] *** denizr has quit IRC
[05:44:16] <arai> I was using filesystems in the create json for that
[05:45:06] <arai> Well snap, I wish you guys were awake a week ago
[05:45:49] <jesse_> I'm not sure if vmadm can even delegate arbitrary zfs
[05:46:09] <arai> Not zfs, but it can do lofs mounts
[05:46:20] <arai> but only on create, not on updates
[05:46:37] <jesse_> there is delegation of /data, I think
[05:46:42] <arai> yes
[05:46:53] <arai> Thats what rmustacc just pointed out
[05:48:09] <arai> I've been creating datasets right under the zones pool and using lofs to mount them into place.  It felt ... kludgey at best
[05:48:58] <jesse_> I've only used delegation with pools I had on the boxes from days before smartos
[05:49:10] <arai> Then I find out about zone delegation after reading the solaris docs and I facepalm
[05:49:44] <jesse_> for other uses it has been enough to adjust the root fs size
[05:53:26] <jesse_> ### comment: add this to the FAQ
[05:53:59] <rmustacc> Hopefully ryancnelson or his bot is in the channel.
[05:56:37] <jesse_> I think he said he's going to go through the logs
[05:57:03] <jesse_> I assumed the ones behind http
[05:57:29] <rmustacc> Yeah, I guess that's probably the case.
[05:59:13] <ChrisPartridge> LeftWing: same deal on the 1.9.0 base, i don't think the postgresql 9.2 package is SMF enabled, 9.1 works fine
[06:00:08] <rmustacc> Consider filing a bug in joyent/pkgsrc
[06:00:32] *** axonpoet has quit IRC
[06:10:46] *** leecallen35 has joined #smartos
[06:10:46] *** leecallen has quit IRC
[06:22:29] *** nefilim1 has quit IRC
[06:50:47] *** sachinsharma has quit IRC
[07:00:24] *** sachinsharma has joined #smartos
[07:07:19] *** wolfeidau has quit IRC
[07:11:52] *** ipalreadytaken has quit IRC
[07:16:44] *** des2 has quit IRC
[07:22:51] *** des2 has joined #smartos
[07:30:22] <MerlinDMC> morning / hello
[07:42:05] <ChrisPartridge> afternoon
[07:57:25] *** mamash has joined #smartos
[07:59:28] *** mamash has left #smartos
[08:14:05] *** emocakes has quit IRC
[08:15:18] *** khushildep has joined #smartos
[08:17:59] *** alucardX has joined #smartos
[08:18:10] <alucardX> morning
[08:22:19] *** texarcana has joined #smartos
[08:26:59] *** khushildep has quit IRC
[08:40:32] *** kamilr has joined #smartos
[08:45:02] *** mamash has joined #smartos
[08:45:49] *** mikl has quit IRC
[08:53:06] *** mikl has joined #smartos
[09:03:15] *** alucardX has quit IRC
[09:06:03] *** alucardX has joined #smartos
[09:06:10] *** ChrisPartridge has quit IRC
[09:29:29] *** bens1 has joined #smartos
[09:33:53] <jbergstroem> I have a itx-rig with three segate drives and a intel ssd for cache. for some reason the disks flags as offline more often than not in zpool. If i restart, they come online again and gets resilvered. I've seen this happen on and off for a year with smartos. I've run every disk check tool there is. Suggestions on what to do?
[09:35:15] <jbergstroem> s/segate/seagate
[09:35:55] <sheppard> jbergstroem: some powersaving setting putting the drives to sleep?
[09:36:21] <jbergstroem> sheppard: thats a good suggestion. i'll check bios
[09:37:39] *** alcir has joined #smartos
[09:39:56] *** alcir has quit IRC
[09:41:56] *** wolfeidau has joined #smartos
[09:56:04] <jbergstroem> sheppard: i disabled whatever power-related functionality i could find on my z68 board. anything else I can do os-wise? power.conf doesn't seem to contain much
[09:56:27] <jbergstroem> smartos just flags the device as unavailalbe, so it probably out of the os's control
[09:59:28] *** KermitTheFragger has joined #smartos
[10:09:47] *** avrntsv has joined #smartos
[10:29:36] *** dysinger has joined #smartos
[10:37:24] *** dysinger has quit IRC
[10:49:36] *** ktkNA is now known as ktk
[10:54:49] *** kamilr has quit IRC
[11:08:59] *** sachinsharma has quit IRC
[11:35:40] *** avrntsv1 has joined #smartos
[11:36:47] *** avrntsv has quit IRC
[11:38:47] *** alcir has joined #smartos
[12:05:16] *** calmh has joined #smartos
[12:11:39] *** khushildep has joined #smartos
[12:11:51] <khushildep> Hi all
[12:11:59] <khushildep> Where does the libgcc's sit in SmartOS?
[12:11:59] <khushildep> ld.so.1: mongod: fatal: libstdc++.so.6: open failed: No such file or directory
[12:12:27] <jperkin> it's part of gcc47-runtime in newer images
[12:12:58] <jperkin> er gcc47-libs sorry
[12:15:30] <MerlinDMC> jperkin, any possibility that you may have an already build ruby 2.0 somewhere? ;)
[12:15:40] *** xinkeT has joined #smartos
[12:16:24] <jperkin> MerlinDMC: no, it isn't in pkgsrc yet
[12:17:20] <MerlinDMC> did look into pkgsrc ... thought of maybe some jpc customer that maybe wanted to have that also
[12:17:48] <MerlinDMC> I got some configure errors with weird could not calculate size of long long or something like that
[12:19:19] <jperkin> taca will likely have it ready for after 2013Q1 freeze
[12:20:48] <MerlinDMC> oh that would be cool
[12:21:27] <MerlinDMC> thx ... will watch out for that if I can't get it compiled somehow ;)
[12:38:39] *** szaydel has joined #smartos
[12:40:12] *** darjeeling has joined #smartos
[12:44:00] *** darjeeling has quit IRC
[12:44:28] *** darjeeling has joined #smartos
[12:44:49] *** ira has joined #smartos
[12:45:37] <khushildep> thx jperkin
[12:46:00] <khushildep> Now for another one - does anyone know how to get a script to be executed POST install? pkgin seems to want to run the script PRE install?
[12:46:29] <jperkin> the INSTALL scripts take a $STAGE argument, you can set it to PRE-INSTALL or POST-INSTALL
[12:50:19] <khushildep> but I want the script itself executed post install - so I expect extraction of the tgz to real prefix and then the script run?
[12:50:34] <khushildep> i.e where would I set $STAGE?
[12:50:47] <khushildep> in the script itself?
[12:51:00] <khushildep> si I have -i postinstall.sh - in there?
[12:51:37] <jperkin> "git grep POST-INSTALL -- '*/INSTALL'" in a checked-out pkgsrc tree, there are plenty of examples
[12:51:58] <jperkin> if building through pkgsrc then STAGE and PKGNAME are already set
[12:52:30] <jperkin> if creating from scratch with pkg_create then you need to set them manually (PKGNAME is $1, STAGE is $2)
[12:55:27] <khushildep> is in the script I can test $2 = "POST-INSTALL" and that should work?
[12:55:56] <khushildep> I'm using pkg_create here
[13:03:59] <jperkin> yes
[13:05:39] *** emocakes has joined #smartos
[13:15:33] *** sachinsharma has joined #smartos
[13:33:04] <leecallen35> guys I am trying to work through my 'domain join persistence issue', picking through the code...
[13:33:32] <leecallen35> my challenges: I was never much of a programmer, and I am completely unfamiliar with the solaris and SMB code bases...
[13:33:56] <leecallen35> it looks like it is storing the domain join context in the SMB SMF service, somehow
[13:34:13] <leecallen35> calling functions such as scf_property_get_value()
[13:34:39] <leecallen35> could someone help me understand that?  does that translate to any sort of persistent storage?
[13:38:12] *** enmand has joined #smartos
[13:43:18] *** xmerlin has joined #smartos
[13:44:31] <mamash> leecallen35: SMF properties are stored with the respective SMF service(s)
[13:44:53] <mamash> are you trying to set a property to a non-default value for your purposes?
[13:45:46] <leecallen35> No... I think the 'smbadm join' is storing the domain join state in the SMF properties. I am trying to find a way to persist the domain join across GZ reboots...
[13:46:22] <leecallen35> I am trying to figure out how it is storing that information in the service (if in fact it is), and whether that would make it persistent on another opensolaris implementation
[13:46:40] <mamash> i suggest you create your own transient SMF service under /opt/custom that sets the property and enable the samba service then
[13:47:19] <ira> I think he's using kernel cifs here.
[13:47:31] <leecallen35> It's no samba, it's CIFS.  And I am not yet sure it is saving this stuff in the SMF service.  First I want to confirm that.
[13:47:33] <mamash> (that would be the case for SmartOS, where GZ is not persistent. on other SunOS implementations SMF properties should be persistent unless overwritten)
[13:47:46] <leecallen35> ah good that answers part of my question
[13:48:55] <leecallen35> would this function call scf_property_get_value() display exactly the same things I see using 'svcprop -v svcname' ?
[13:50:05] <mamash> i guess it should (not sure really)
[14:04:14] <xmerlin> hi to all
[14:04:49] <xmerlin> I've a strange problem with pkgin ...pkgin upgrade generates a core dump
[14:04:56] <xmerlin> pkgin install xxxx works
[14:05:09] <mamash> what image/packageset?
[14:05:24] <xmerlin> SmartMachine base64 1.8.4
[14:05:37] <xmerlin> without any addition ...only fifo
[14:06:12] <mamash> that would nail it down to fifo i guess. what changes does adding fifo entails?
[14:06:39] <xmerlin> these ones http://project-fifo.net/display/PF/Installation+Guide
[14:06:49] <xmerlin> only ...it's a clean installation
[14:07:00] *** emocakes has quit IRC
[14:07:40] <mamash> that adds a 2nd repository to pkgin? that could be it
[14:07:46] <mamash> pkgin never worked well with multiple repositories
[14:07:48] <jperkin> yeh, multiple repos isn't supported
[14:08:00] <mamash> sometimes it works, but it's pretty unpredictable
[14:08:42] <xmerlin> :(
[14:08:53] <xmerlin> the same setup on another machine works as expected
[14:08:59] <xmerlin> and it's a very bad news
[14:09:26] <mamash> are you looking for something specific to 'upgrade', or just checking?
[14:09:59] <xmerlin> the second one
[14:15:43] <xmerlin> sorry for the bad question ...but why did you choose a package manager that doesn't support multiple repositories?
[14:16:02] *** jefferai has left #smartos
[14:16:43] *** avrntsv1 has quit IRC
[14:16:43] <jperkin> because the alternative was nothing
[14:17:30] <jperkin> well, not nothing, but certainly nothing close to being as convenient as pkgin
[14:17:38] <xmerlin> apt?
[14:17:47] <xmerlin> for example
[14:17:58] <jperkin> there is currently no working .deb backend for pkgsrc
[14:18:05] <jperkin> though the infrastructure is there to support it
[14:18:16] <jperkin> it would be quite a fun project to work on if you fancy it
[14:19:57] <mamash> also back when we picked pkgin, its developer seemed to have much more time to work on it, compared to last year or so
[14:20:58] <xmerlin> I understand
[14:21:32] <mamash> i should say that pkgin does 'support' multiple repos, but the level of support is the lowest possible. it merges all the sources in its db and doesn't concern itself about absolutely any problems rising from the new situation
[14:21:35] <jperkin> I would certainly welcome a working apt/dpkg support for pkgsrc, it would certainly make things simpler for people used to linux, and may give us better support for things like package upgrades with config file management
[14:22:17] <jperkin> there was a gsoc to add rpm backend support to pkgsrc, I don't think the project was completed but it would certainly be useful for any work
[14:22:56] <jperkin> I wouldn't recommend going for the rpm/yum route directly though as it would add a dependency on python
[14:23:18] <jperkin> at least pkgin is nice and simple C/sqlite
[14:23:40] <mamash> "nice" - not if you actually read the code ;)
[14:24:30] <jperkin> well.. :)
[14:28:24] <xmerlin> I'm agree with you apt is better than rpm/yum
[14:28:34] <xmerlin> I don't like the python dep
[14:29:03] *** pringlescan has quit IRC
[14:29:46] <xmerlin> from a distro development perspective something like the gentoo portage is even better ...but it's less focused on binary packages
[14:30:05] <jperkin> well, that's what pkgsrc is at its core
[14:30:21] <jperkin> we just happen to provide binary packages by default as that's what users want
[14:30:40] <jperkin> but there's nothing stopping you tracking pkgsrc trunk and building everything yourself, there are plenty of tools to help you with that
[14:31:29] <jperkin> portage was inspired by freebsd ports (hence the name), and pkgsrc is a direct descendant of ports (was forked around 1997)
[14:33:23] <xmerlin> yes but in portage you can recompile your system with only 1 command ...live ...and it's very simple to recompile broken dependencies
[14:34:02] <jperkin> yes, pretty similar with pkgsrc, and you can even recreate the true gentoo experience by putting CFLAGS+=-funroll-loops -O6 etc in mk.conf ;)
[14:35:04] <xmerlin> lol
[14:35:19] <jperkin> pkgtools/pkg_chk will let you 'pkg_chk -u' to rebuild everything you already have installed from the latest source
[14:36:37] <xmerlin> mmm --> http://pastebin.com/mjVWerNX
[14:37:00] <xmerlin> pk :(
[14:39:04] *** neophenix has joined #smartos
[14:40:35] *** mamash has left #smartos
[14:51:26] *** denizr has joined #smartos
[15:16:33] *** russiane39 has quit IRC
[15:24:40] *** sachinsharma has quit IRC
[15:30:19] *** kim0 has joined #smartos
[15:30:53] <kim0> Folks, can I run smartos under some virtualization tool (ESX better, or vbox) and still play with kvm ?
[15:31:50] <jperkin> esx perhaps, not vbox - it works with vmware fusion certainly
[15:32:01] <jperkin> version >=5
[15:32:17] <kim0> my ESX is 5.1
[15:32:25] <kim0> cool ..
[15:32:35] <kim0> can I run a full cloud ... with smart DC
[15:32:41] <jperkin> I've never used esx, so I can only suggest you just try it and see what happens
[15:32:41] <kim0> simulated on esx
[15:32:49] <kim0> ookkaayy
[15:33:00] <jperkin> we develop SDC on fusion
[15:33:05] <kim0> I guess I'll try to install solaris additions as well
[15:33:12] <kim0> that should be close enough :)
[15:33:28] <wesolows> solaris definitely is not useful for simulating a cloud
[15:33:34] <kim0> SDC at minimum needs 2 nodes right ? head and cpu correct?
[15:33:42] <wesolows> head and compute, yes
[15:33:50] <kim0> awesome
[15:33:57] <wesolows> and that can definitely be simulated in fusion
[15:41:10] <xmerlin> jperkin, "pkg_chk -u" was a disaster ...removed some key packages ...and stopped because circular dependecies
[15:42:02] <jperkin> this is why we provide binary packages ;)
[15:42:15] <xmerlin> argh ...I hate you :-P
[15:42:39] *** theup has joined #smartos
[15:42:40] <jperkin> circular deps suggest an incorrect setup though
[15:43:24] <xmerlin> the automatic uninstall is bad
[15:43:58] <jperkin> for live systems, yes, you'd normally do the builds inside a chroot
[15:44:56] <xmerlin> I've a "builder" zone ...but now is quite useless
[15:45:07] <jperkin> http://mail-index.netbsd.org/tech-pkg/2005/01/05/0006.html is an old post by me on how I used to do it, though I don't think those URLs still exist
[15:45:34] <jperkin> or there is the longer http://www.perkin.org.uk/posts/pkgsrc-on-solaris.html, but again a lot of that is obsolete these days, was written a long time ago
[15:46:23] <xmerlin> thanks
[15:47:02] <jperkin> when I get a spare afternoon I'll write the next in my 'pkgsrc on SmartOS' series which will go into pbulk and how to do bulk source builds
[15:48:23] *** tonyarkles has joined #smartos
[15:49:45] <theup> Hi everyone... I am currently trying to set up a KVM machine alongside a couple of OS zones... following the user guide, I downloaded an ubuntu image and successfully created the machine (showing as started). However, I can't seem to obtain any info using 'vmadm info <uid>' - the output is just blank. I'm trying to get the VNC settings. Does anyone know what I could likely be doing wrong?
[15:52:23] *** axonpoet has joined #smartos
[15:56:05] *** bens1 has quit IRC
[16:00:35] *** nefilim has joined #smartos
[16:09:04] *** khushildep has quit IRC
[16:14:27] *** avrntsv has joined #smartos
[16:14:43] *** jim80net has joined #smartos
[16:16:30] *** mamash has joined #smartos
[16:20:25] *** mamash has left #smartos
[16:21:17] <sebasp> question of the zfs gurus out there... why does zpool get version shows a version of - on smartos ?
[16:22:00] <EMH_Mark3> because of feature flags I assume
[16:22:45] <sebasp> ok...I'm trying to migrate ou of Solaris to smartos and now I need to zfs send and zfs receive about 2Tb of data :)
[16:23:05] <sebasp> on Solaris I'm at version 31
[16:23:31] <EMH_Mark3> think highest version supported by smartos/opensolaris/etc is 28...
[16:23:36] <sebasp> ok...
[16:23:54] <EMH_Mark3> but that might not affect zfs send/receive
[16:24:50] <sebasp> I get this on the source host: cannot receive: stream has unsupported feature, feature flags = 24
[16:24:57] <sebasp> actually...
[16:25:10] <sebasp> it's on the receiving end ...
[16:25:11] <sebasp> sorry
[16:25:27] <EMH_Mark3> :|
[16:25:38] <EMH_Mark3> guess you'll be rsync'ing then...
[16:26:03] <sebasp> got a bunch of snapshots that I need to keep...
[16:26:14] <sebasp> :)
[16:26:55] <sebasp> guess scripting will be my next activity :)
[16:28:34] <MerlinDMC> sebasp, rsync first snapshot to the target ... snapshot there again, rsync second snapshot over the first one and repeat :P
[16:29:56] <MerlinDMC> if you have spare disks you could also try to create a version 28 pool without feature flagsand send there then boot that up in smartos for upgrade and send another time
[16:30:16] <sebasp> MerlinDMC : figured it will be something around those lines...I knew I should not have upgraded that filesystem when I installed Solaris 11 Express way back!
[16:30:17] <sebasp> :)
[16:31:01] <MerlinDMC> yeah ... the missing downgrade is bad in such situation
[16:31:15] <MerlinDMC> but hopefully doesn't happen often
[16:32:06] *** tonyarkles has quit IRC
[16:32:38] <sebasp> actuallly as soon as this is done, I'm re-installing the host with smartos and won't have this issue for a while...
[16:33:48] *** dysinger has joined #smartos
[16:39:23] <rmustacc> The questoin is what is the zfs version.
[16:39:25] *** Forced has quit IRC
[16:40:25] *** Forced has joined #smartos
[16:43:17] *** kfr- has joined #smartos
[16:44:17] <kfr-> is linux kernel 3.0.69 better than linux kernel 2.6.18 in a KVM guest?
[16:44:30] <rmustacc> Generally speaking, yes.
[16:44:51] *** CarlosC has joined #smartos
[16:45:47] <kfr-> I'm currently stuck on RHEL5 running PAE kernel
[16:46:04] <kfr-> going to OL 6.3 with x86_64 kernel
[16:46:32] <kfr-> Sun Ray kernel modules compile fine on 3.6.x
[16:53:00] *** szaydel has quit IRC
[16:53:11] *** nefilim has quit IRC
[16:53:45] *** dap has joined #smartos
[16:53:47] *** kfr_ has joined #smartos
[16:53:59] *** kfr- has quit IRC
[16:54:02] *** kfr_ is now known as kfr-
[16:56:15] *** ktk is now known as ktkNA
[16:57:43] *** alucardX has quit IRC
[16:57:43] *** tonyarkles has joined #smartos
[16:58:14] *** nefilim has joined #smartos
[17:01:05] *** szaydel has joined #smartos
[17:01:13] *** bens1 has joined #smartos
[17:01:21] *** khushildep has joined #smartos
[17:03:22] *** chorrell has joined #smartos
[17:04:38] *** kim0 has quit IRC
[17:08:33] *** jim80net has quit IRC
[17:08:55] <sebasp> rmustacc what do you mean by the question is what is the zfs version?
[17:10:45] *** dysinger has quit IRC
[17:11:54] *** dysinger has joined #smartos
[17:15:30] *** jim80net has joined #smartos
[17:20:52] *** alcir has quit IRC
[17:23:43] *** cxreg has joined #smartos
[17:24:45] <Licenser> xmerlin there is a issue with pkgin that it segfaults when a second repo is present
[17:27:12] *** nefilim has quit IRC
[17:28:12] <rmustacc> sebasp: There is more than one version in zfs.
[17:28:22] <rmustacc> See zfs(1M) and the version property.
[17:28:43] <rmustacc> That historically has been what controls send/receive.
[17:30:36] *** mikl has quit IRC
[17:33:10] *** nefilim has joined #smartos
[17:35:22] *** ryancnelson has joined #smartos
[17:37:22] <sebasp> rmustacc actually both my solaris 11 box and smartos box are at version 5 of zfs get version
[17:38:15] <rmustacc> Ah yes, this one.
[17:38:23] <rmustacc> This came out in a recent illumos developer thread.
[17:38:35] <rmustacc> Let me see if I can get Matt Ahren's comment on it.
[17:40:31] <rmustacc> See: http://www.listbox.com/member/archive/182179/2013/03/sort/time_rev/page/3/entry/21:150/20130315182631:619F9A92-8DBF-11E2-ACEA-A643D5BAC4CE/
[17:41:05] <rmustacc> And probably the related messages in that thread.
[17:41:21] *** nefilim has quit IRC
[17:41:41] <|woody|> oracle did change the zfs format/send/recivce stuff without changing the version. Happend somewhere in solaris 10 u10 patch and in some solaris11 sru
[17:42:57] *** nefilim has joined #smartos
[17:47:02] *** dysinger has quit IRC
[17:47:13] <|woody|> ZFS stream data generated by zfs send on Solaris 10 + 147440-14/147441-14 (or later) or Solaris 11+ SRU3 (or later) , can't be received using zfs recv although zpool version and zfs version are same. This problem happens only with ZFS version 5. If ZFS version is 4 or older, this does not happen.
[17:47:41] <|woody|> so if you have a pool with that config you can send to to any "open" zfs
[17:49:21] <xmerlin> Licenser, ok :(
[17:49:37] *** dysinger has joined #smartos
[17:49:38] <Licenser> xmerlin sorry for the bad news mate
[17:49:39] *** Cpt-Oblivious has joined #smartos
[17:50:51] *** nefilim has quit IRC
[17:52:41] *** nefilim has joined #smartos
[17:57:15] *** ira has quit IRC
[17:58:14] *** theup has left #smartos
[17:58:32] *** theup has joined #smartos
[18:10:57] *** leecallen has joined #smartos
[18:10:57] *** leecallen35 has quit IRC
[18:12:26] *** theup has left #smartos
[18:33:31] *** nefilim has quit IRC
[18:33:38] <opeth__> I'm currently migrating my data pool off S11 to SmartOS by having created a pool with -o version=28 and a dataset with -o version=5 as my S11.1 runs 34 and 6 respectively
[18:34:08] <opeth__> and it seems to be able to recv my v6 zstreams though it does complain about the 'share' attributes' inability to be received
[18:34:21] <opeth__> but that's all, data is intact, pool can be exported, imported, mounted and read well
[18:35:04] *** nefilim has joined #smartos
[18:36:44] *** tonyarkles has quit IRC
[18:37:29] *** KermitTheFragger has quit IRC
[18:38:34] *** nefilim has quit IRC
[18:38:36] <sebasp> opeth__ so if I upgrade my zpool I could be fine?
[18:38:44] <sebasp> rmustacc thanks much...
[18:38:59] <sebasp> |woody| kind of sad from them to do this...
[18:39:36] <opeth__> sebasp: I can say only that what I said above seems to do it for me now
[18:39:49] <opeth__> I have tested by creating a zpool over 2 iscsi luns
[18:39:59] *** nefilim has joined #smartos
[18:40:09] <opeth__> and from 11.1 zpool create -o version=28 foo mirror bar baz
[18:40:29] <opeth__> then zfs create -o version=5 foo/zoo
[18:40:37] <sebasp> opeth__ so both instance are runnins solaris
[18:41:02] <sebasp> not one with illumos and the other with solaris ..
[18:41:06] <opeth__> and zfs send -R source/data@snap | zfs recv -Fduv -o copies=1 foo/zoo
[18:41:13] <opeth__> still on the same 11.1 solaris node
[18:41:17] <sebasp> ok
[18:41:17] <opeth__> and then export this pool
[18:41:19] <opeth__> release the luns
[18:41:30] <opeth__> and import the same pool from the smartos node running the recent illumos kernel and userland
[18:41:36] <opeth__> and it will import and mount.
[18:41:39] <sebasp> ok...
[18:41:49] <opeth__> my smartos image uses zfs version 5 and zpool version 28
[18:42:00] <opeth__> however if I create the pool on the illumos instance
[18:42:09] <opeth__> I can import it on the solaris instance just as fine
[18:42:16] <sebasp> will need to find an alternative, my environemt is geared to do what you state...
[18:42:55] <opeth__> I was surprised to see how well it went but it seems to work fine
[18:43:54] <opeth__> I couldn't do anything else, all my data was on a version 34 zpool on version 6 datasets
[18:44:05] <opeth__> that naturally my illumos target couldn't use
[18:44:16] <opeth__> but if you create legacy pools and datasets, it will.
[18:46:58] <sebasp> but one thing I could try is open a case with Oracle (still pay support :))
[18:47:15] <opeth__> what would your usecase be?
[18:47:55] <sebasp> actually...I though that zfs send to my solaris 10 box would have failed...but no and use that as usecase
[18:49:47] <opeth__> so you have no disks that you could use as a temporary intermediary legacy pool like I do?
[18:50:51] *** kfr- has quit IRC
[18:51:54] <sebasp> nope...
[18:52:16] <opeth__> that's bad news
[18:52:21] <sebasp> got to physical machine only in a lights out location ... both with 4 x 1TB
[18:52:42] <sebasp> already formated...one is running smartos the other is running solaris 11
[18:52:52] <sebasp> so rsync is my solution it seems
[18:53:36] <opeth__> perhaps creating an iscsi-shared zvol on the likely empty illumos box
[18:53:53] <opeth__> and creating a legacy version pool on those from the solaris box
[18:54:01] <sebasp> yep...
[18:54:02] <opeth__> and replicate to that
[18:54:10] <sebasp> hmm...
[18:54:11] <opeth__> kind of what I did
[18:54:12] <sebasp> maybe...
[18:54:34] <opeth__> try it with a few gigabytes
[18:54:42] <sebasp> yep :)
[18:54:44] <opeth__> and see if you can access everything from smartos then
[18:54:51] <opeth__> the method I pasted above works for me
[18:55:01] <opeth__> replicate to the legacy pool, export it
[18:55:23] <opeth__> yada yada
[18:55:33] <sebasp> then remount on smartos then zfs send | receive
[18:55:51] <opeth__> huh?
[18:56:01] <opeth__> I'd send | recv on the solaris image
[18:56:17] <opeth__> then export it and try to import it with smartos
[18:56:17] <sebasp> but I don't want to keep it on iscsi ;)
[18:56:29] <opeth__> you needn't on the long run I'd think
[18:56:35] <opeth__> in the long run, even
[18:56:40] <opeth__> I won't either
[18:56:59] <sebasp> will seee...thanks for the tip...
[18:57:12] <opeth__> you're welcome
[18:57:21] <opeth__> my advice is to plan this on paper
[18:57:24] <opeth__> step-by-step
[18:57:30] <opeth__> and read it through a couple of times
[18:57:33] <opeth__> if it seems sane, try it
[18:57:39] <sebasp> hahaha
[18:57:58] <sebasp> good...will read the irc logs to get the flow ;)
[18:58:44] <opeth__> kk I was being all-serious
[18:58:48] <opeth__> I'm a visual thinker myself
[18:58:58] <opeth__> and I prefer planning before acting
[19:03:55] *** khushildep has quit IRC
[19:08:35] *** szaydel has quit IRC
[19:10:53] *** szaydel has joined #smartos
[19:14:37] *** Melinko has joined #smartos
[19:21:07] *** denizr has quit IRC
[19:51:47] *** denizr has joined #smartos
[19:53:04] *** noahmehl has joined #smartos
[19:53:22] <nahamu> office hours must have come and gone already...
[19:53:44] <wesolows> seems that way
[19:55:17] <nahamu> perhaps everyone was getting real work done.
[19:55:31] <opeth__> real work, eek
[19:55:44] <wesolows> I know I was!  Thursdays often provide rare opportunities to get shit done.
[19:56:41] <linuxprof> do you know of a way to see the cpu usage per zone?
[19:57:45] <opeth__> zonestat exists on solaris for sure
[19:57:52] <opeth__> I'm not yet accustomed to anything else
[19:57:57] <richlowe> rmustacc: may take you up on building pkgsrc to test stuff :\
[19:58:03] <opeth__> prstat -Z may also be enough
[19:58:42] <linuxprof> zonestat doesnt exist in smartos afaik
[19:59:05] <linuxprof> guess prstat -Z could work =)
[20:01:33] <linuxprof> now i just have to figure out how to show all zones and no processes, hehe
[20:02:02] <opeth__> let me look my oneliner up for you
[20:03:38] <linuxprof> much appreciated
[20:06:24] <opeth__> I know this will sound queer, but I don't have it at hand. I won't until tomorrow morning :/
[20:06:37] <opeth__> will be happy to share unless that's too late
[20:08:12] <linuxprof> nah, its fine =)
[20:11:21] *** enmand has quit IRC
[20:12:00] <opeth__> if you query me and give me an e-mail address I could send it to, I will around 9 CET most likely
[20:12:17] <opeth__> but I won't be able to access IRC by then so do so now if you want to
[20:18:39] *** noahmehl has quit IRC
[20:24:33] *** _trev has joined #smartos
[20:28:51] <_trev> Hi all, smartos novice here. I'm having an issue with a module loading in my app
[20:29:08] <_trev> 'Error: ld.so.1: node: fatal: /root/path/to/file.node: unknown file type'
[20:29:11] <sebasp> exit
[20:29:16] <sebasp> oops
[20:29:20] <sebasp> \quit
[20:29:23] *** sebasp has quit IRC
[20:29:25] <_trev> You meant /oops
[20:29:29] <_trev> :P
[20:35:53] *** enmand has joined #smartos
[20:52:34] *** denizr has quit IRC
[20:52:54] *** enmand has quit IRC
[20:54:06] *** denizr has joined #smartos
[20:55:37] *** enmand has joined #smartos
[20:58:56] <nahamu> e^ipi: do you have a script or some other tool you use when you cut releases that trigger checkouts of the release branches or do you do it by hand?
[21:01:24] *** bens1 has quit IRC
[21:01:32] <e^ipi> script
[21:01:51] <rmustacc> _trev: What does the output of file /root/path/to/file.node say?
[21:01:53] <rmustacc> richlowe: Okay.
[21:01:57] <e^ipi> it's specific to the SDC build, so you can't have it
[21:02:02] <e^ipi> sorry
[21:02:07] <nahamu> no worries.
[21:02:56] <e^ipi> at one point it was by hand, obviously we have workload reducing technologies so that ended quickly
[21:03:14] <nahamu> I'm all about workload reducing technologies.
[21:04:49] <e^ipi> i read some blog post about some guy w/ minimal programming experience that ended up in a janky data entry job who scripted away his work and spent the day messing around on reddit & stuff
[21:05:35] <_trev> rmustacc: the file looks like it's binary
[21:05:44] <_trev> unreadable chars in it
[21:05:48] <e^ipi> seems silly to me, that's the sort of thing you bring up the chain past your direct manager so that you get a promotion and a raise, not just waste time on reddit
[21:06:35] <nahamu> indeed.
[21:06:37] <wesolows> well, that and if you still have to go to the office you may as well spend your time there doing something
[21:07:50] <nahamu> I guess I'll just do a gmake update. Should be close enough.
[21:08:29] <nahamu> What I really ought to do is get my patches pulled in so that I can stop spinning my own builds.
[21:08:36] *** tonyarkles has joined #smartos
[21:08:58] <e^ipi> patches?
[21:09:13] <e^ipi> have i ignored more of your pull requests? If so i apologize
[21:09:26] <nahamu> they're old and probably need test cases.
[21:10:04] *** denizr has quit IRC
[21:10:13] <nahamu> and/or documentation in addition.
[21:11:30] <nahamu> it's these two commits, basically https://github.com/nshalman/smartos-live/commit/6578bce332175f5adc9109db9b6dc89cebc1cae1
[21:11:37] <nahamu> and https://github.com/nshalman/smartos-live/commit/a4376266f22387cb0ccac142f586275ae7829078
[21:11:58] <nahamu> the ipv6 stuff is clearly ambitious and needs some careful examination.
[21:12:28] <nahamu> the listen_ip one probably just needs documentation
[21:13:18] <e^ipi> those are ones I'd run by robg & josh
[21:13:23] <nahamu> yeah
[21:13:46] <nahamu> two people who I'm sure have tons of spare time... :-P
[21:14:01] <e^ipi> heh
[21:14:50] *** bens1 has joined #smartos
[21:15:28] *** denizr has joined #smartos
[21:19:05] *** bens1 has quit IRC
[21:27:58] <nahamu> oh! are all the release branches already marked?
[21:28:19] <nahamu> (for today)
[21:28:29] *** bens1 has joined #smartos
[21:28:42] <e^ipi> should be
[21:32:58] *** _trev has quit IRC
[21:33:59] <nahamu> cool
[21:51:39] *** enmand has quit IRC
[21:56:26] *** enmand has joined #smartos
[21:58:02] *** bens1 has quit IRC
[22:06:09] *** iyp has joined #smartos
[22:09:47] *** enmand has quit IRC
[22:16:00] *** denizr has quit IRC
[22:17:19] *** khushildep has joined #smartos
[22:35:20] *** ira has joined #smartos
[22:41:37] *** tonyarkles has quit IRC
[22:42:28] *** Melinko has quit IRC
[22:42:54] *** Melinko has joined #smartos
[22:53:14] *** szaydel has quit IRC
[22:56:19] *** wolfeidau has quit IRC
[23:01:38] *** khushildep_ has joined #smartos
[23:02:06] *** khushildep has quit IRC
[23:02:07] *** khushildep_ is now known as khushildep
[23:05:59] *** jim80net has quit IRC
[23:07:37] *** ChrisPartridge has joined #smartos
[23:07:42] *** enmand has joined #smartos
[23:08:30] *** Forced has quit IRC
[23:08:38] *** neophenix has quit IRC
[23:08:57] *** Forced has joined #smartos
[23:10:01] *** wolfeidau has joined #smartos
[23:11:37] *** bens1 has joined #smartos
[23:16:09] *** bens1 has quit IRC
[23:28:31] *** iyp has quit IRC
[23:48:32] *** gkyildirim has joined #smartos
[23:51:19] *** mrj has joined #smartos
top

   March 21, 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 | >