Switch to DuckDuckGo Search
   March 15, 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
[00:00:02] <rmustacc> KVM or OS?
[00:00:30] <trentster1> rmustacc: both
[00:00:33] <rmustacc> There is nothing for KVM.
[00:00:43] <rmustacc> As far as the host cares it locks down all of its memory at the beginning of time.
[00:00:49] <rmustacc> It's always using 100% of its memory.
[00:00:49] *** bcantrill has quit IRC
[00:00:55] <rmustacc> What the guest itself is doing with said memory is a mystery.
[00:00:58] <rmustacc> You have to ask it directly.
[00:01:14] <rmustacc> For OS, look at the implementation of the zomememstat command.
[00:01:14] <trentster1> rmustacc: that makes sense now why i can find any "mem" stuff when grepping kstats for kvm vm zoen id
[00:02:54] <trentster1> rmustacc: thanks will do.
[00:04:30] <rmustacc> Well, there will be stats there.
[00:04:34] <rmustacc> But they will be mostly meaningless.
[00:05:05] *** potatosalad has quit IRC
[00:08:08] *** deirdres_ has joined #smartos
[00:10:35] *** deirdres has quit IRC
[00:10:35] *** deirdres_ is now known as deirdres
[00:25:59] *** denizr has quit IRC
[00:26:26] *** denizr has joined #smartos
[00:27:53] *** nefilim1 has joined #smartos
[00:27:56] *** xmerlin has quit IRC
[00:32:42] *** deirdres has quit IRC
[00:32:58] *** deirdres has joined #smartos
[00:43:57] *** xmerlin has joined #smartos
[00:45:48] *** denizr has quit IRC
[00:49:02] *** theCzar has quit IRC
[00:49:03] *** dap has quit IRC
[00:49:35] *** dap has joined #smartos
[00:54:19] *** des2 has quit IRC
[00:59:19] *** deirdres has quit IRC
[01:06:11] *** bcantrill has joined #smartos
[01:08:11] *** denizr has joined #smartos
[01:08:38] *** tonyarkles has quit IRC
[01:15:38] *** ken_tcc has joined #smartos
[01:16:15] *** ken_tcc has left #smartos
[01:17:16] *** ken_tcc has joined #smartos
[01:27:32] <Licenser> hmm is it popssible that mpstat w/o a interval is entirely useless? I always get the same results o.O
[01:29:29] *** tonyarkles has joined #smartos
[01:33:05] <brendang> it's summary since boot
[01:33:46] <brendang> it's a little useful. but not for showing what is happening now
[01:34:38] <Licenser> ah darn and I hoped I'd get around finding something tricky
[01:34:46] <ken_tcc> I am trying to create a VM for the most resent centos-6 image, version 1.3.0, I get the error:
[01:34:48] <Licenser> thanks mate
[01:34:55] <ken_tcc> Command failed: cannot create 'zones/e76e6440-0053-4ad4-ac7a-548d24eae805-disk0': 'refreservation' is greater than current volume size
[01:35:18] <ken_tcc> {
[01:35:18] <ken_tcc>     "brand": "kvm",
[01:35:19] <ken_tcc>     "autoboot": false,
[01:35:19] <ken_tcc>     "vcpus": 2,
[01:35:19] <ken_tcc>     "ram": 2048,
[01:35:19] <ken_tcc>     "alias": "$HOSTNAME",
[01:35:19] <ken_tcc>     "hostname": "$HOSTNAME",
[01:35:20] <ken_tcc>     "nics": [{
[01:35:20] <ken_tcc>         "mac": "$MAC",
[01:35:21] <ken_tcc>         "nic_tag": "admin",
[01:35:21] <ken_tcc>         "model": "virtio",
[01:35:22] <ken_tcc>         "primary": 1,
[01:35:49] <ken_tcc> I am running SmartOS Live Image v0.147+ build: 20130307T214308Z
[01:36:28] <ryancnelson> you don't have enough disk
[01:37:13] <ken_tcc> I have 612G available in the global zone
[01:39:29] *** darjeeling has quit IRC
[01:44:45] <nahamu> ken_tcc: can you pastebin the full json?
[01:47:13] <nahamu> I'm guessing in your disk section you specified a size bigger than the size of the centos image. which would be a very understandable mistake.
[01:48:32] *** Triluch_ is now known as Triluch
[01:49:31] <ryancnelson> it's "disk0" that is complaining for him, though, which should be of fixed-size, based on the dataset
[01:50:06] <nahamu> ryancnelson: I agree, but I think vmadm might still allow, or even require specifying a size in that section.
[01:50:21] <ken_tcc> nahamu: your guess was correct
[01:50:24] <ryancnelson> i would, in fact, like to see the whole json
[01:50:53] <ken_tcc> I decreased the image_size by a factor of 10 and the command succeeded. thanks
[01:50:56] <ryancnelson> ... specifying a bigger disk won't actually give you a bigger filesystem in the boot disk
[01:51:20] <ryancnelson> you'll just have wasted space after the linux volume on that "device"
[01:51:25] <jeffpc> out of curiosity, does this mean anything to anyone?
[01:51:26] <jeffpc> 1.0.0.127.dnsbugtest.1.0.0.127.in-addr.arpa.
[01:51:27] <nahamu> ryancnelson: I think that the "size" parameter on the boot disk is used to set refreservation
[01:51:38] <jeffpc> (the GZ sends that out during boot)
[01:51:47] <jeffpc> sends that out == DNS query
[01:52:38] <ken_tcc> whole JSON is pasted here: http://pastebin.com/44TbNb5q
[01:54:05] <ryancnelson> jeffpc, that's a healthcheck as multicast dns setup... google for "dnsbugtest"
[01:54:08] <nahamu> right. the man page says that image_size should be set to match the size of the source image
[01:54:45] <nahamu> which makes sense if you know that it's going to be used as refreservation and that you could potentially use a lower number if you want to "oversell" disk.
[01:55:00] <jeffpc> ok
[01:55:06] <nahamu> but highly confusing if you think "the disk already has a size... why do I have to tell you what size it is?"
[01:55:33] <nahamu> at any rate, ken_tcc are you good to go now?
[01:55:39] <jeffpc> I also noticed some "random" IPs being pinged... 69.65.40.29, 128.177.27.52, 169.229.70.183
[01:56:05] *** copec has quit IRC
[01:56:10] <ryancnelson> those aren't random... those are pools.ntp.org
[01:56:20] <ryancnelson> time sync
[01:56:44] <ryancnelson> ### comment:  add this to the FAQ
[01:57:12] <jeffpc> that's was my only guess... the gz just tries to ping them (and since I didn't have routing going it got stuck pinging)
[01:58:22] <ryancnelson> that should timeout, and then the ntp service should go into maintenance
[01:58:38] *** xmerlin has quit IRC
[01:58:49] <jeffpc> I didn't wait long enough for that, but I've seen ntp going into maintenance
[01:58:59] <jeffpc> (on previous boots)
[01:59:37] <ryancnelson> well, without routing, you're not going to get to any ntp servers
[02:00:35] <jeffpc> more context... the router is a zone; I had a dumb config problem and gz was trying to get at DNS before zones were up
[02:00:43] <jeffpc> so I did some packet tracing
[02:00:45] *** CarlosC has quit IRC
[02:00:50] <jeffpc> which pointed at my /etc/hosts being incomplete
[02:03:25] *** bcantrill has quit IRC
[02:04:40] *** copec has joined #smartos
[02:15:15] *** Cpt-Oblivious has quit IRC
[02:16:27] *** dap has quit IRC
[02:22:27] *** ryancnelson has quit IRC
[02:27:04] <Licenser> hmm is it correct that kstat doesn't give integers but floats for things like this: cpu:1:sys:cpu_nsec_idle 1.03946832498533e+15 / aka can I force it to give a number is not rounded
[02:32:09] *** darjeeling has joined #smartos
[02:33:11] *** darjeeling has quit IRC
[02:33:46] *** darjeeling has joined #smartos
[02:48:10] *** alpharender has joined #smartos
[02:52:42] <richlowe> No
[02:53:22] <richlowe> I mean, no, it's not correct
[02:54:27] <richlowe> or perhaps "a bit"
[02:55:59] *** deedubs has left #smartos
[02:56:02] *** deedubs has joined #smartos
[02:56:06] *** deedubs has left #smartos
[03:02:49] *** ipalreadytaken has quit IRC
[03:07:40] <ken_tcc> nahamu: yes, thanks, my immediate problem is solved though I have others I will post about later
[03:21:11] *** ipalreadytaken has joined #smartos
[03:25:19] *** ipalreadytaken has joined #smartos
[03:26:02] *** ipalreadytaken has quit IRC
[03:29:58] *** ipalreadytaken has joined #smartos
[03:32:25] *** axonpoet has joined #smartos
[03:33:17] *** ipalreadytaken has joined #smartos
[03:38:21] *** ipalreadytaken has quit IRC
[03:52:15] <sheppard> shitty internet ftw
[04:01:38] *** ira has quit IRC
[04:02:02] <rbrown2> How big is joyent people wise today?
[04:03:35] *** ipalreadytaken has joined #smartos
[04:03:47] *** axonpoet has quit IRC
[04:05:45] <wesolows> |<-----    ----->|
[04:06:02] <wesolows> 'bout yea big
[04:08:03] *** ipalreadytaken has quit IRC
[04:11:38] <richlowe> today we learn that joyent is font-dependent.
[04:13:20] *** tonyarkles has quit IRC
[04:41:00] *** des2 has joined #smartos
[04:44:30] *** _jgrafton_ has quit IRC
[04:54:56] *** richlowe has quit IRC
[04:57:12] *** bcantrill has joined #smartos
[05:04:03] *** ipalreadytaken has joined #smartos
[05:08:24] *** tonyarkles has joined #smartos
[05:08:42] *** ipalreadytaken has quit IRC
[05:09:09] *** dalethion has quit IRC
[05:22:08] *** ipalreadytaken has joined #smartos
[05:22:12] *** denizr has quit IRC
[05:22:38] *** ipalreadytaken has quit IRC
[05:22:54] *** ipalreadytaken has joined #smartos
[05:25:06] *** denizr has joined #smartos
[05:35:22] *** denizr has quit IRC
[05:47:00] <arekinath> hmm.. why does the joyent brand take a fresh snapshot of the template ds before cloning it every time?
[05:47:41] <arekinath> with it done that way, there's no way to move zones around using incremental sends
[05:48:00] <arekinath> if they all cloned from the same snapshot, you could freely use incremental sends to transfer zones around
[05:59:42] *** axonpoet has joined #smartos
[06:03:47] <ken_tcc> probably it is the snapshots being cloned. I don't think cloning is allowed for non-snapshots items in the pool.
[06:03:47] *** leecallen has quit IRC
[06:04:27] *** leecallen has joined #smartos
[06:13:02] *** tonyarkles has quit IRC
[06:22:13] *** nefilim has quit IRC
[06:42:40] <arekinath> ken_tcc: correct. but if we made one snapshot of a template and then always cloned that for a new zone
[06:42:51] <arekinath> ken_tcc: then incremental transfers between machines would work
[06:43:07] <arekinath> versus currently where they don't unless you transfer all the snapshots of the template in the exact order they were created
[06:43:17] <arekinath> and then you can only receive from one of each template-source pair
[06:44:24] *** nikolam has joined #smartos
[06:44:57] <arekinath> also, this is really cool. I can consistently panic a smartos box by trying to rsync a zone
[06:45:13] <arekinath> that's a nice feature
[06:48:32] <nikolam> arekinath, how do you rsync it, is it doing it inside the zone, rsyncing all zone files or outside or what
[06:48:38] *** wolfeidau has quit IRC
[06:48:40] <ken_tcc> she is definately a high-strung race horse. I wonder if rsync would trigger a panic if dedupe was turned off
[06:48:59] <nikolam> sounds like a bug, what kind of zone you made.
[06:53:08] *** theCzar has joined #smartos
[07:02:15] <arekinath> nikolam: outside the zone
[07:02:23] <arekinath> and dedupe was off at both ends
[07:03:11] <arekinath> http://pastie.org/pastes/6507665/text?key=fog5tgvyzwyrhvunicvq
[07:03:32] <arekinath> same trace all of the 3 times I've done it so far, from inside zfs
[07:05:26] *** wolfeidau has joined #smartos
[07:08:12] <nikolam> arekinath, you were rsyncing system parts of a zone, from outside the zone, while zone was running?
[07:10:53] <nikolam> aren't you suppose to make snapshot/clone the zone or make new zone and install changes in it and then start the new zone?
[07:11:21] <nikolam> (while retaining old one)
[07:16:00] <nikolam> maybe if you want to see what VM will do if you change it's system files while running, you should do that in KVM machine, that is inside the zone, and when KVM machine pull the legs up, it will kill only itself.
[07:19:01] *** theCzar has quit IRC
[07:23:23] *** ChrisPartridge has quit IRC
[07:24:35] *** marsell has quit IRC
[07:28:01] <arekinath> nikolam: the rsync was restricted to looking at things from zfs diff -- it wasn't trying to sync /proc or anything
[07:28:16] <arekinath> nikolam: but even still, it shouldn't panic the kernel
[07:28:20] *** bcantrill has quit IRC
[07:28:33] <arekinath> just because I did something silly like that
[07:37:03] *** alpharender has quit IRC
[07:58:37] *** sachinsharma has joined #smartos
[07:58:57] <nikolam> I am sure devs would welcome well funded project ideas that could improve features in illumos.
[07:59:56] <MerlinDMC> morning / hello
[08:12:23] *** theCzar has joined #smartos
[08:20:40] *** theCzar has quit IRC
[08:20:49] *** texarcana has quit IRC
[08:22:03] *** texarcana has joined #smartos
[08:22:12] *** alucardX has joined #smartos
[08:22:54] <alucardX> morning guys
[08:28:22] <MerlinDMC> morning alucardX
[08:38:40] *** bens1 has joined #smartos
[08:58:22] *** bens1 has quit IRC
[09:00:15] *** chh has joined #smartos
[09:05:08] *** mamash has joined #smartos
[09:19:51] *** ipalreadytaken has quit IRC
[09:24:47] *** haydock has joined #smartos
[09:26:32] *** alcir has joined #smartos
[09:27:26] <scarcry> jperkin: ping
[09:32:30] *** chh has quit IRC
[09:45:53] *** darjeeling has quit IRC
[09:50:31] *** ipalreadytaken has joined #smartos
[09:51:24] *** marsell has joined #smartos
[10:01:49] *** defa has joined #smartos
[10:01:56] *** ipalreadytaken has quit IRC
[10:02:54] *** chh has joined #smartos
[10:11:15] *** bens1 has joined #smartos
[10:15:15] <jperkin> scarcry: pong
[10:30:46] *** sachinsharma has quit IRC
[10:58:35] *** ipalreadytaken has joined #smartos
[11:12:47] *** ipalreadytaken has quit IRC
[11:26:54] <scarcry> jperkin: I created another zone from multiarch-12.4.1 for puppet. I noticed that /opt/local/share/examples/ruby193-puppet/svc/puppetmasterd.xml from ruby193-puppet-0.25.5nb1 pkgsrc uses /opt/local/lib/svc/method/svc-puppetmasterd for its start method but said method file actually contains /opt/csw as the prefix
[11:27:27] <scarcry> instead of /opt/local I mean
[11:28:03] <scarcry> should we consider that a mistake and if so, want me to file an issue for it on GH ?
[11:28:42] <jperkin> scarcry: yes please, thanks.
[11:28:52] <scarcry> jperkin: cool, will do
[11:30:35] *** alcir has quit IRC
[11:31:36] *** alcir has joined #smartos
[11:34:11] *** robinbowes has joined #smartos
[11:35:19] *** chh has quit IRC
[11:37:01] <scarcry> jperkin: https://github.com/joyent/pkgsrc/issues/43
[11:38:18] <mamash> that package is a mess
[11:38:30] <scarcry> mamash: agreed
[11:38:52] *** chh has joined #smartos
[11:39:38] <mamash> ah
[11:39:42] <mamash> i remember now
[11:39:42] <trentster1> jperkin: hi there…. are there know problem with the 1.9.0 datasets?
[11:39:52] <trentster1> *known
[11:39:53] <scarcry> my intention is to run puppetmaster from a zone and the clients in other zones
[11:40:08] <scarcry> all based on multiarch 12.4.1
[11:40:25] <mamash> scarcry: actually i had massacred that package to update to puppet-2.7.18 in the fall, but it didn't make it into the 2012Q4 branch
[11:40:37] <scarcry> trentster1: it's dev stuff so breakage is to be expected
[11:40:44] <scarcry> not sure what you're seeing?
[11:40:50] <mamash> trentster1: just found one so far
[11:41:33] <scarcry> mamash: did your puppet package end up elsewhere?
[11:41:53] <mamash> mozilla-rootcerts is using the platform openssl and it populates /opt/local/etc/openssl/certs with incorrect hash names, so validation of root certs doesn't work properly in openssl enabled apps
[11:41:59] <mamash> i'm fixing that
[11:42:19] <mamash> scarcry: it was a local hack on our 2012Q2 branch, i didn't push that upstream
[11:42:24] <mamash> need to look at it again
[11:42:30] <scarcry> mamash: is that related to git clone actions using https scheme too ?
[11:42:37] <scarcry> openssl I mean
[11:42:43] <mamash> most likely
[11:42:47] <scarcry> got some ssl errors from that
[11:42:51] <scarcry> switched to using ssh keys
[11:43:24] <mamash> it's relatively easy to fix by replacing the openssl path in the mozilla-rootcerts script, then run 'mozilla-rootcerts rehash' in /opt/local/etc/openssl/certs again
[11:43:27] <mamash> just a tad annoying
[11:43:58] <mamash> 1.9.1 will show up soon
[11:47:14] <scarcry> mamash: good to know although I've never used the 1.9.0 image..I like what multiarch brings and it needs some testing love so jperkin can bring it under imgadm control
[11:47:37] <trentster1> mamash: scarcry where do i see which images are dev etc?
[11:48:01] <scarcry> trentster1: my bad, I got it mixed up ..1.9.0 is not dev AFAIK
[11:48:22] <scarcry> I was confused with the 0.99.0 trunk image that jperkin has listed on his blog
[11:48:57] <trentster1> I found 2 errors thus far, 1st one is when using the image it totally ignores any setting in .curlrc file where on other images it works fine, e.g if you tell it to ignore https check for git/curl.
[11:48:58] <scarcry> I'm not sure what is new in 1.9.0
[11:49:12] <mamash> it's not 'dev' in any way
[11:49:14] <mamash> just fresh
[11:49:24] <scarcry> mamash: right, that's what I figured
[11:49:27] <trentster1> 2nd one is if you install postgresql 191 from pkgin and start it service goes immediately into maintenance
[11:49:37] <mamash> http://wiki.joyent.com/wiki/display/jpc2/SmartMachine+Base+Release+Notes
[11:49:43] <mamash> just not updated
[11:49:45] *** sachinsharma has joined #smartos
[11:49:48] <scarcry> I've only used a base64 1.8 image once and then switched to multiarch 12.4.1 and haven't gone back since
[11:49:49] <mamash> dunno what's taking the guys so long
[11:50:09] <trentster1> mamash: yeah thats first place I looked when 1.9.0 came out and no info was there.
[11:50:39] <mamash> i passed the info to the doc guys, not sure what's going on
[11:51:43] <mamash> the only major change (beyond the switch to 2012Q4) is that /opt/local/gnu is no more, back to everything in PATH, and we killed some legacy entries in /etc/passwd and used a fixed map of UIDs/GIDs for stuff like www, mysql etc
[11:51:58] <mamash> so that those users do not pollute the normal uid/gid space of real users
[11:52:51] <trentster1> mamash: is my postgresql bug confirmed?
[11:56:03] *** dizko has quit IRC
[11:56:04] *** dizko_ has joined #smartos
[11:59:30] *** sachinsharma has quit IRC
[12:08:55] *** mamash has left #smartos
[12:09:28] *** ipalreadytaken has joined #smartos
[12:10:37] *** ira has joined #smartos
[12:11:57] <opeth__> hey; is there a howto on rolling packages only I seem to need to be fed locally to pkgin?
[12:12:02] <opeth__> as in, properly
[12:14:26] *** mamash has joined #smartos
[12:30:11] *** bluezenix has joined #smartos
[12:30:15] *** mikl has joined #smartos
[12:30:41] *** ipalreadytaken has quit IRC
[12:31:24] *** chh has quit IRC
[12:32:50] <jperkin> opeth__: by rolling do you mean creating your own packages from binaries you have already built, or rebuilding packages already available in pkgsrc?
[12:34:26] <opeth__> the former
[12:35:15] <jperkin> opeth__: http://www.perkin.org.uk/posts/creating-local-smartos-packages.html
[12:35:28] <opeth__> thank you very much
[12:36:47] <opeth__> yes, seems a perfect fit
[12:36:48] <opeth__> cheers
[12:37:03] <opeth__> apologies for my not finding this beforehand on my own
[12:39:48] *** darjeeling has joined #smartos
[12:58:29] *** JBrickley has joined #smartos
[12:58:50] *** JBrickley has left #smartos
[13:01:52] *** chh has joined #smartos
[13:04:38] *** szaydel has joined #smartos
[13:06:25] *** xmerlin has joined #smartos
[13:09:11] *** chh has quit IRC
[13:09:42] *** chh has joined #smartos
[13:26:56] *** alucardX has quit IRC
[13:27:18] *** ipalreadytaken has joined #smartos
[13:29:03] <jesse_> jperkin, what happened to the jdk/jre 6 packages? the summary says there are two, but at least the newest binary packages do not exist
[13:31:11] <jperkin> jesse_: somewhat uncertain redistribution agreements, the native pkgsrc packages do not permit the resulting binaries to be made available but our internal joyent one does.
[13:31:35] <jperkin> one of the reasons I really want to get openjdk in for 2013Q1 ;)
[13:33:47] <mamash> we were using a different source and license for the jre/jdk tarballs (one which allowed redistribution, not oracle's site), but it's no longer available, so openjdk7 is the only way to go
[13:36:21] <jesse_> jperkin, well... some jdk6 is required to compile openjdk7=)
[13:36:34] <jesse_> I don't see much other use for the 6
[13:36:47] <jperkin> yes, and you can still build it manually
[13:37:10] <mamash> that jdk6 can be the non redistributable oracle java too - it will only be a build dependency
[13:38:54] <jesse_> true, but it still needs to be redistributable so that anyone can compile openjdk with it
[13:39:20] <jperkin> no, you download the binaries from oracle
[13:39:21] <jesse_> ofcourse, this problem will go away after we have openjdk7 binary package and the newest java is 8
[13:39:51] <jesse_> what's the difference between pkgsrc and joyent packages?
[13:40:02] <jperkin> the pkgsrc sun-jdk6 package then installs it under $PREFIX and the openjdk package then uses that for bootstrap
[13:40:13] <jesse_> and can I add joyent repo to pkgsrc somehow to be able to dl it?
[13:40:15] <jperkin> not a lot other than version
[13:40:25] <jperkin> https://github.com/joyent/pkgsrc-joyent
[13:40:53] <jesse_> thinking how to do that in practice with trunk
[13:40:55] <jperkin> unpack to pkgsrc/joyent
[13:41:33] <jperkin> if you want to use the joyent jdk you'll need to adjust the dependencies in mk/java-vm.mk
[13:42:00] <jesse_> maybe I should check if openjdk7 build will accept itself as a seed package
[13:42:17] <jperkin> that's what I'm currently looking at
[13:42:34] <jperkin> but first I'm trying to avoid the libgcc dependency, but -static-libgcc doesn't appear to be holding for bin/java
[13:43:03] <jesse_> hmm?
[13:43:07] <jesse_> you want it to be static?
[13:43:58] <jperkin> if possible, else it gets messy
[13:44:18] <jesse_> oh, it's possible
[13:44:29] <jesse_> some parts of it are probably already static
[13:44:56] <jesse_> it will just require some changes to pretty fragile build system
[13:45:16] <jesse_> I didn't see it as a priority when I made it compile
[13:45:59] <jesse_> and, obviously, I didn't ask ;)
[13:53:10] *** mamash has left #smartos
[14:01:22] *** Melinko2003 has joined #smartos
[14:03:58] *** mamash has joined #smartos
[14:05:19] *** Melinko has quit IRC
[14:06:12] *** ipalreadytaken has quit IRC
[14:09:00] <jesse_> it should be ok to mirror the openjdk hg repos as-is to github, right?
[14:10:03] <jesse_> it being gpl and all
[14:10:03] *** darjeeling has quit IRC
[14:54:48] <MerlinDMC> someone here who can tell if there is some more info about progress in the internal ticket IMGAPI-88 than in smartos-live#142 ... just want to know if it's worked on currently or if I can look at that myself again
[14:57:31] *** scarcry has quit IRC
[15:01:55] *** ipalreadytaken has joined #smartos
[15:02:29] *** Trixboxer has joined #smartos
[15:08:26] *** nikolam has quit IRC
[15:13:50] *** scarcry has joined #smartos
[15:13:59] *** scarcry_ has joined #smartos
[15:16:31] *** denizr has joined #smartos
[15:22:18] *** Melinko2003 has quit IRC
[15:31:47] *** neophenix has joined #smartos
[15:39:41] *** darjeeling has joined #smartos
[15:42:44] *** khushildep has joined #smartos
[15:45:48] *** ipalreadytaken has quit IRC
[15:47:04] *** nefilim has joined #smartos
[15:47:20] *** szaydel has quit IRC
[15:50:15] *** AlainODea has joined #smartos
[15:51:15] *** szaydel has joined #smartos
[15:51:53] *** scarcry_ has quit IRC
[15:52:11] <AlainODea> Is there a way to run imgadm through a web filter/proxy?  I am using Squid for the web proxy.
[15:52:40] *** tonyarkles has joined #smartos
[15:53:14] *** scarcry has quit IRC
[15:57:37] *** scarcry has joined #smartos
[15:57:43] <MerlinDMC> AlainODea, I was asking about the status of ticket IMGAPI-88 ... if I get the time and trentm is not working on that currently i'll look into the new imgapi and try to include some sort of support
[15:58:25] <rmustacc> MerlinDMC: I'll try and ping him today to ask about that.
[15:59:07] <MerlinDMC> rmustacc, would be cool thx
[16:00:21] <MerlinDMC> some other thing ... can a zone be configured as NAT gateway for other zones with the standard joyent brand? ... do I have to allow ip spoofing or anything else?
[16:00:45] <_dumfries> Would one expect a block device for when you attach a ISO file to a KVM instance? I am not booting of the ISO, rather I need to access the ISO from within the instance. I've ensure the ISO is in the proper location (/zones/<UUID>/foo.iso) and started the VM with 'cdrom=/foo.iso' and the instances comes up but I don't see the ISO.
[16:01:19] <jesse_> I thought I'd do a quick patch up to jdk7u17-b02 (which is the one oracle ships itself)
[16:01:35] <jesse_> the changes here are baffling
[16:01:36] <MerlinDMC> _dumfries, /zones/<uuid>/root/foo.iso ... don't forget the /root/ part
[16:01:55] <rmustacc> _dumfries: Also, waht format did you specify there?
[16:02:25] <_dumfries> MerlinDMC: rmustacc sorry, I did include root as well.
[16:02:28] <rmustacc> MerlinDMC: May have to allow ip spoofing and get the router flag on the nic there, I don't really remember. I think nahamu would know.
[16:02:38] <_dumfries> rmustacc: cdrom=/foo.iso,ide
[16:02:48] <rmustacc> What guest?
[16:02:56] <_dumfries> Ubuntu
[16:03:32] <rmustacc> Well, it will show up as some kind of block deviec in the guest.
[16:03:37] <rmustacc> *device
[16:03:42] <jesse_> instead of old code that read the stack pointer from a register
[16:03:47] <jesse_> they now do
[16:03:49] <jesse_> volatile int dummy;
[16:03:50] <_dumfries> rmustacc: that's what I was expecting.
[16:03:55] <jesse_> address sp = (address)&dummy + 8;     // %%%% need to confirm if this is right
[16:04:06] <rmustacc> ls
[16:04:31] <rmustacc> _dumfries: But some kernels will also point it in the /dev/srXXX on GNU/Linux.
[16:05:09] <jesse_> is volatile automatic var even guaranteed to be on the stack?
[16:05:18] <jesse_> (on gcc, at least)
[16:05:19] <_dumfries> rmustacc: yeah, I've been looking for srXXX, cdrom, etc. hmmm
[16:05:48] <rmustacc> But if it's actually showing up as a *block* device and its ide, I'd expect /dev/hdXXX.
[16:06:04] <rmustacc> But they change that stuff all the time.
[16:06:20] <rmustacc> And udev is anything but consistent.
[16:06:21] <_dumfries> Love a moving target ;)
[16:06:55] <MerlinDMC> on newer kernels those ide devices can also get listed as /dev/sdXX
[16:07:20] <rmustacc> Ah, right. I guess those drivers are all no longer experimental.
[16:07:43] *** szaydel has quit IRC
[16:08:26] *** CarlosC has joined #smartos
[16:08:49] <nahamu> I think I ran into a bug with the latest patch upgrading to bunyan 0.19.0... http://paste.ec/?6f9c62bac23194e1#vR7DoAqQgUbUvs/enE9nrSqqg5iUE0YoyqN6mRi/tcM=
[16:10:33] <rmustacc> nahamu: File a bug please.
[16:10:45] *** szaydel has joined #smartos
[16:11:20] <AlainODea> MerlinDMC: thanks :)  Hopefully there's a solid solution.  Node.js does not appear to have integrated support for HTTP proxies.  It's easy to do it manually, but the libraries like restify don't implement it
[16:12:15] <rmustacc> Yes, node's http client can often be underwhelming.
[16:16:57] *** szaydel has quit IRC
[16:32:00] *** mamash has left #smartos
[16:32:44] <nahamu> rmustacc: https://github.com/joyent/smartos-live/issues/184 should probably be for Josh W.
[16:33:37] <nahamu> I'll probably reroll my build from the commit before that one...
[16:34:02] *** ipalreadytaken has joined #smartos
[16:34:48] *** ryancnelson has joined #smartos
[16:36:06] <Trixboxer> What are the known issues for SmartOS ? unable to get link from google.  ( May be its bug free :) )
[16:36:26] <nahamu> Trixboxer: https://github.com/joyent/smartos-live/issues
[16:37:10] <Trixboxer> nahamu: are all issues including KVM are tracked there ?
[16:37:49] <nahamu> https://github.com/joyent/illumos-kvm/issues
[16:38:16] <nahamu> but if you're filing new issues, you can put stuff into https://github.com/joyent/smartos-live/issues and the developers will still do the right thing.
[16:40:16] <Trixboxer> ok thanks
[16:40:45] <Trixboxer> atm my disklayout command weird info which does not matches to pool
[16:41:35] *** chh has quit IRC
[16:41:39] *** dap has joined #smartos
[16:49:05] *** dalethion has joined #smartos
[16:49:27] <rmustacc> Trixboxer: disklayout is not designed to give you the layout of the current pool.
[16:50:26] <dalethion> Hello
[16:52:48] <rmustacc> Hello
[16:53:10] <Trixboxer> rmustacc :it shoes vdevs, vdev type, spares etc
[16:53:53] <jesse_> Trixboxer, it's the layout sdc would create with your disks
[16:54:00] <rmustacc> Trixboxer: Yes. The purpose of the tool is to show how we would ideally lay out your pool.
[16:54:51] <ryancnelson> if you want to see how it's actually laid out, that's zpool status
[16:54:53] <Trixboxer> another smart feature :)
[16:55:05] <dalethion> rmustacc: You have helped me yesterday...however it seems that there is still some issue with the memory. I have created one kvm with 128 GB RAM, started it, created a second, started it...The second one does not finish booting and I get an error message in vm.log about qemu_mlock not being able to lock al;l the memory....
[16:55:15] <ryancnelson> disklayout is a helper script chunk of our setup process.
[16:55:16] <rmustacc> How much swap do you have?
[16:55:21] <ryancnelson> which is why there's no manpage
[16:55:35] <Trixboxer> ryancnelson: yes
[16:55:42] <rmustacc> It should arguably have one.
[16:55:54] <rmustacc> And hopefully at some point we'll integrate it into the smartos setup process as well.
[16:56:06] <ryancnelson> … i suppose so… only 'cause it ships with the system
[16:56:12] <rmustacc> daelthion: The swap question was for you, just to be clear since there are are a few different conversations.
[16:56:15] <dalethion> rmustacc: After I have started one kvm (128 GB RAM) this is how much I have: swap 245 GB
[16:56:49] <dalethion> rmustacc: With no KVM started it is 490 GB
[16:57:01] <rmustacc> I just mean, how large is the physical swap volume?
[16:57:15] <rmustacc> In total, regardless of what's running.
[16:57:21] <dalethion> Oh sorry, 7 GB.
[16:57:39] <dalethion> Are we talking about the physical swap device?
[16:57:46] <dalethion> That is 7 GB.
[16:58:14] <rmustacc> Really, the GZ only has a 7 GB swap zvol?
[16:58:59] <dalethion> I think I am a bit slow here, I will give you a paste of df -h
[16:59:11] <rmustacc> Actually, no.
[16:59:21] <rmustacc> Give me zfs list
[17:00:15] <dalethion> rmustacc: http://pastebin.com/DkkKeGXS
[17:00:38] *** alcir has quit IRC
[17:01:08] <rmustacc> dalethion: Did you manually create the zpool or did you go through setup?
[17:01:23] <dalethion> Setup.
[17:01:47] <rmustacc> Wait, you only have 33 GB of HDD Space?
[17:02:09] <dalethion> I have a total of 72 GB. There are 2 hardware mirrored drives.
[17:02:12] <dalethion> Is that an issue?
[17:02:22] <rmustacc> Yes.
[17:02:30] <rmustacc> You have no swap partition.
[17:02:42] <dalethion> We are planning on adding FC storage to this server.
[17:03:03] <rmustacc> We don't support putting zones on the FC.
[17:03:27] <ryancnelson> ### comment: add this to the FAQ
[17:03:31] <rmustacc> Or rather, we expect all the primary storage to be on local disk.
[17:03:39] <rmustacc> That includes swap.
[17:03:49] <dalethion> I know, but we would reinstall SmartOS on the FC storage and have that be the global zone. Is that possible?
[17:03:57] <wesolows> FWIW, SDC will not allow you to setup unless total local disk is 2x or more the size of DRAM
[17:04:02] <nahamu> ryancnelson: are those comments for you to be able to find them yourself later in the logs?
[17:04:09] <ryancnelson> yes
[17:04:12] <nahamu> cool
[17:04:25] <nahamu> I'll try not to abuse that knowledge ;)
[17:04:36] <ryancnelson> it's a good-enough defacto thing i'm gonna grep for, later.  please join in, if you like!
[17:05:28] <rmustacc> rayncnelson: Thanks.
[17:05:37] <rmustacc> dalethion: Right now we don't support fc.
[17:06:00] <wesolows> and for those who want to split hairs on what "support" means, it doesn't work.  at all.
[17:06:14] <dalethion> Hm, ok. So ideally I would need how much local disk for this to work?
[17:06:30] <wesolows> 2 TB plus whatever you want to store in VMs/zones
[17:06:49] <wesolows> a pair of $140 2 TB SAS disks will do you.
[17:06:51] *** richlowe_ is now known as richlowe
[17:07:01] <wesolows> or even 1, really.
[17:07:49] <dalethion> Ok, thanks :) I guess I will have to do that if I want this to work.
[17:08:12] <ryancnelson> is this a blade server or something?  seems like a LOT of ram and a TEEEEENY disk
[17:08:25] <wesolows> you spent $40k on DRAM; hopefully you can scrape up a couple hundred to get this across the line. :-)
[17:08:30] <ryancnelson> (no offense :) )
[17:09:28] <dalethion> It is a HL DL580. We don't really care about disk, we keep everything on network storage and usually install a minimal Solaris.
[17:09:37] <nahamu> https://i.chzbgr.com/maxW500/4738386176/h8FA07781/
[17:09:40] <dalethion> *HP
[17:10:15] <dalethion> (none taken)
[17:11:12] <rmustacc> Understandable. SmartOS is basically all about local storage.
[17:13:43] <wesolows> depending on exactly what configuration is required to make a particular FC setup operable (i.e., to make disk nodes appear in devfs), it may be as simple as adding fcp to the platform.  But I suspect that in virtually all SAN use cases that will not be sufficient and you will run up against the brick wall of the nonpersistent root filesystem.
[17:15:10] <richlowe> I imagine that boot services and lofs can get you out of that, if you care enough.
[17:16:11] <wesolows> yes
[17:16:16] <Trixboxer> what could be a good way to create "Active/Passive" bond in smartos ?
[17:16:21] <wesolows> anything's possible
[17:16:28] <Trixboxer> I created an aggr mentioned as http://wiki.smartos.org/display/DOC/Managing+NICs
[17:16:43] <wesolows> what you're describing is ipmp, which is not supported at all in SmartOS
[17:16:51] <Trixboxer> but the link is LACP with mode=off
[17:20:15] <dalethion> wesolows : We are installing the storage today, I will give it shot and see. If not, bigger local storage it is.
[17:28:59] *** bcantrill has joined #smartos
[17:35:00] *** khushildep has quit IRC
[17:44:32] *** Melinko has joined #smartos
[17:48:20] *** xmerlin has quit IRC
[17:48:52] *** tonyarkles has quit IRC
[17:49:58] <_dumfries> rmustacc: still unable to expose a ISO from the to an Ubuntu instance, where it works flawlessly in Windows. Very odd .
[17:50:05] *** axonpoet has quit IRC
[17:53:43] <jesse_> jperkin, €%&/&€%#"€#€, it didn't do static-libgcc because it only enables it if gcc version is 3. As gcc 2 doesn't support it.=)
[17:54:00] <jperkin> jesse_: I changed that to 4, it didn't help
[17:58:54] *** szaydel has joined #smartos
[17:59:44] <jesse_> jperkin, is it linked dynamically with the libraries or the binaries?
[17:59:48] <jesse_> or everything?
[18:02:51] *** xmerlin has joined #smartos
[18:03:33] <xmerlin> Alasdairrr, ping
[18:04:03] <jperkin> jesse_: I didn't check the libs, but at least bin/java was still dynamically linking so it wasn't necessary to check further..
[18:04:18] *** leecallen has quit IRC
[18:04:47] <nahamu> _dumfries: did you copy the ISO into the zone root? how did you do it for windows?
[18:04:53] *** leecallen has joined #smartos
[18:05:09] <JT-EC> xmerlin: Sup?
[18:05:17] <_dumfries> nahamu: for windows I copied into /zones/<UUID>/root/<iso>.iso
[18:05:26] <_dumfries> nahamu: and did the same for Ubuntu
[18:05:44] <nahamu> _dumfries: can you double check that it did get copied into the ubuntu zone?
[18:05:54] <_dumfries> nahamu: Yep
[18:05:58] <nahamu> (double check the uuid of the ubuntu zone)
[18:06:02] <xmerlin> JT-EC, I've installed the openwebstack repo ...and I've setted crle -l /lib:/usr/lib:/ec/lib
[18:06:02] <xmerlin> crle -64 -l /lib/64:/usr/lib/64:/ec/lib/amd64
[18:06:15] <xmerlin> but /ec/bin/pkg gives me http://pastebin.com/8BadPFxH
[18:06:46] <rmustacc> You need to put the ec stuff before ours.
[18:06:53] <rmustacc> Or you need to have a proper run path.
[18:06:56] <_dumfries> nahamu: Yep, lives in /zones/<UUID>/root/<iso>.iso
[18:06:59] <JT-EC> You've set crle in the wrong order for the newer SmartOSs
[18:07:09] <JT-EC> as rmustacc says.
[18:07:17] *** wramthun has quit IRC
[18:07:21] <_dumfries> nahamu: and I am currently passing in "cdrom=/<iso>.iso,ide" when booting the instance
[18:07:24] <xmerlin> I've followed the wiki page
[18:07:33] <JT-EC> That's fixed in recent pkg packages I think but apparently not in the bootstrap
[18:07:41] <xmerlin> ok
[18:07:44] <JT-EC> xmerlin: Yeah, it's not your fault.
[18:08:04] <xmerlin> ok now it works
[18:08:13] <nahamu> _dumfries: so what's the failure? is it failing to boot from the iso? not seeing it from an already installed system?
[18:08:31] <_dumfries> nahamu: unable to see the ISO from within the running instance.
[18:08:42] <xmerlin> JT-EC, is your repo integrated with svc?
[18:08:57] <nahamu> _dumfries: did you stop the VM before trying to attach the CD?
[18:09:08] <_dumfries> nahamu: yes, several times now.
[18:09:10] *** wramthun has joined #smartos
[18:09:26] <JT-EC> xmerlin: The packages in it should be using svc where applicable, yep
[18:09:35] *** sachinsharma has joined #smartos
[18:09:35] <jesse_> jperkin, there's also the problem of using the name 'openjdk', it's apparently trademarked
[18:09:45] <nahamu> _dumfries: look in /zones/<uuid>/root/tmp/vm.log to make sure that qemu is getting the extra arguments.
[18:09:55] <jesse_> jperkin, http://en.wikipedia.org/wiki/OpenJDK "It was also the first to use OpenJDK for the package name (via the OpenJDK trademark agreement) instead of IcedTea."
[18:10:10] <jesse_> so, SmartJDK?=)
[18:10:12] <nahamu> and make sure you didn't accidentally put in a typo somewhere.
[18:10:42] <xmerlin> JT-EC, I've installed nginx ...and as I can see there is a var/svc/manifest/site/http-nginx.xml but I cannot see it using the command svcs
[18:10:53] <jperkin> jesse_: that makes no sense, fedora etc distribute it as 'openjdk'
[18:11:04] <_dumfries> nahamu: not seeing much in the way of it indicating whether or not the ISO was found, are you perhaps talking about the startvm script?>
[18:11:36] <xmerlin> is there a way to add another "search path" to svcs & friends?
[18:11:50] <jesse_> jperkin, because of the trademark agreement
[18:11:53] <jesse_> apparently
[18:12:02] <jesse_> didn't really look more into it
[18:12:16] <JT-EC> xmerlin: svcs -a | grep nginx
[18:12:18] <rmustacc> xmerlin: A script needs to either import it or tell smf to rescan and do manifest import again.
[18:12:23] <jperkin> sigh, licensing
[18:12:34] <jperkin> one day everyone will use the ISC license and the world will be a better place.
[18:12:45] <JT-EC> rmustacc: Yeah, manifest-import should have been restarted when he installed it.
[18:12:47] <_dumfries> nahamu: https://gist.github.com/anonymous/5171464 looks like it is
[18:13:47] <nahamu> so what makes you thing that ubuntu can't see it?
[18:14:31] <jesse_> jperkin, http://openjdk.java.net/legal/openjdk-trademark-notice.html
[18:14:32] <xmerlin> rmustacc, now it's more clear ..thank you very much
[18:14:33] <_dumfries> nahamu: the block device is not created, and I have gone through all the usual naming suspects, it seems that there is no disk
[18:15:27] <nahamu> _dumfries: yeah, check the startvm script and make sure the iso is mentioned in it.
[18:15:27] *** bcantrill has quit IRC
[18:15:31] <jesse_> jperkin, looks like one needs approved repo on openjdk.java.net
[18:17:25] <jperkin> jesse_: I don't see anything wrong with that notice, however I am very much not one for legalese
[18:17:47] <jperkin> but it seems like a standard 'you can't call this "openjdk" if you've gone and modfied it beyond recognition'
[18:17:52] <jperkin> which we haven't
[18:18:20] <jesse_> they want to have the repo at their site
[18:18:27] <jesse_> which I kind of understand
[18:18:53] <jperkin> it's questionable whether this constitutes a 'new operating system'
[18:19:02] <_dumfries> nahamu: yep, it's in there
[18:19:17] <nahamu> is this the Joyent ubuntu image, or your own?
[18:20:01] <jesse_> jperkin, in other news, apparently the build it ok with openjdk7 as the bootstrap jdk
[18:20:08] <jesse_> buils is
[18:20:17] *** khushildep has joined #smartos
[18:20:20] <jesse_> b u i l d  i s
[18:20:37] <jperkin> yeh, that's how the *BSD ports do it
[18:20:43] <_dumfries> nahamu: joyent's
[18:20:54] <jesse_> jperkin, one can always fork and call it something else
[18:20:56] <jesse_> ofcourse
[18:21:02] <xmerlin> JT-EC what's the best way to start to create a new package that depends on this repo?
[18:21:20] <jesse_> and refer to it as "our packaging of openjdk n"
[18:22:27] *** enmand_ has quit IRC
[18:23:44] *** denizr has quit IRC
[18:23:45] <jesse_> jperkin, actually, if I didn't read that wrong, you can't call it openjdk if it's not port to new os or hw architecture=)
[18:24:23] <jesse_> it would be nice if someone qualified checked that, though
[18:25:32] *** des2 has quit IRC
[18:25:44] <nahamu> _dumfries: I'm stumped. I haven't played with that image, so I don't know if my usual tricks would work on it (cat /proc/partitions, etc.)
[18:25:52] *** nikolam has joined #smartos
[18:26:02] <_dumfries> nahamu: likewise, no idea.
[18:26:33] <JT-EC> xmerlin: We use the userland framework that should be linked to in the wiki page - unfortunately I don't think the latest repo has made it public yet.
[18:26:41] *** denizr has joined #smartos
[18:27:42] *** Cpt-Oblivious has joined #smartos
[18:29:14] <xmerlin> JT-EC, actually I've downloaded this https://hg.openindiana.org/users/aszeszo/s10-userland and I've followed the last path of the guide http://wiki.smartos.org/display/DOC/3rd+Party+Software+Repos
[18:31:01] *** wramthun has quit IRC
[18:31:26] *** sachinsharma has quit IRC
[18:31:51] *** nikolam has quit IRC
[18:33:12] *** wramthun has joined #smartos
[18:33:39] *** des2 has joined #smartos
[18:37:52] *** enmand has joined #smartos
[18:38:03] *** bluezenix has quit IRC
[18:38:38] *** bcantrill has joined #smartos
[18:41:01] *** nikolam has joined #smartos
[18:44:03] <jesse_> jperkin, are you sure -static-libgcc works at all with the pkgsrc gcc?
[18:44:32] *** kfr- has joined #smartos
[18:47:06] *** denizr has quit IRC
[18:48:12] <wesolows> who was having trouble with diskinfo crashing?  was that pringlescan?
[18:48:21] <wesolows> I have a test binary for you if you're here.
[18:54:32] *** hij1nx has quit IRC
[18:55:52] *** hij1nx has joined #smartos
[18:58:53] *** axonpoet has joined #smartos
[19:00:32] *** axonpoetsphere has joined #smartos
[19:03:28] *** axonpoet has quit IRC
[19:03:29] *** axonpoetsphere is now known as axonpoet
[19:09:43] *** szaydel has quit IRC
[19:12:49] *** riceo has joined #smartos
[19:17:08] *** denizr has joined #smartos
[19:19:06] *** ken_tcc has quit IRC
[19:19:34] *** ipalreadytaken has quit IRC
[19:27:26] *** khushildep has quit IRC
[19:29:59] *** bluezenix has joined #smartos
[19:37:47] *** bens1 has quit IRC
[19:49:21] *** ira has quit IRC
[19:52:38] *** riceo has quit IRC
[19:58:16] *** nhubbard has joined #smartos
[20:00:04] <nhubbard> I'm setting up a smartos server and to be honest have no experience with solaris.  I'm having a hard time figuring out how I can set the hostname of the global zone, I created the XML file as described here: http://wiki.smartos.org/display/DOC/Setting+a+static+hostname+at+boot+time but even after a reboot the hostname was still the mac address, am I missing something?
[20:02:08] *** porkbelt_ has joined #smartos
[20:02:19] *** porkbelt has quit IRC
[20:02:20] *** porkbelt_ is now known as porkbelt
[20:08:15] <Trixboxer> I can mount NFS export but its readonly, any idea ?
[20:12:03] *** ipalreadytaken has joined #smartos
[20:17:26] *** bluezenix has quit IRC
[20:19:05] <nikolam> Trixboxer, I compiled and use ntfs-3g on Openindiana, it is working here, I suppose it should also work on omnios?
[20:19:28] <Trixboxer> nikolam: Im on smartos
[20:22:22] *** kfr- has quit IRC
[20:23:54] *** nikolam has quit IRC
[20:26:06] *** xmerlin has quit IRC
[20:26:55] <richlowe> wesolows: it was pringlescan
[20:27:44] <richlowe> wesolows: does your "freeing stack garbage" cover the executable name in the core being screwed?
[20:27:51] <richlowe> (see, for eg, $m)
[20:30:11] *** des2 has quit IRC
[20:30:34] *** ipalreadytaken has quit IRC
[20:35:59] <jperkin> nhubbard: if it didn't work then perhaps you didn't copy/paste it correctly as it should work fine.
[20:36:19] *** des2 has joined #smartos
[20:36:44] *** neophenix has quit IRC
[20:38:41] <jperkin> nhubbard: alternatively you could use the rc.local method described here https://gist.github.com/jclulow/2606370, that one definitely works for me
[20:44:16] <wesolows> richlowe: Probably not; the case I was thinking of is the one where we never even get the controller attributes and then go down and nvlist_free them later.
[20:45:10] *** jim80net has joined #smartos
[20:45:27] <nhubbard> thanks jperkin that seems to have done the trick
[20:47:07] <jesse_> jperkin, static tries to link with -lgcc -lgcc_eh, dynamic with -lgcc_s
[20:47:32] <jesse_> in both cases, ld generates dynamic link to gcc_s
[20:47:45] <jesse_> doesn't make sense, what am I missing?=)
[20:48:31] <jperkin> dunno, will look at it on monday, time for some weekend ;)
[20:49:30] <jesse_> this _is_ my weekend=)
[20:49:45] <jesse_> dodging other stuff I need to do=)
[20:50:28] *** AlainODea has quit IRC
[20:54:45] <richlowe> wesolows: if you look at the mappings in the guy's core, wherever we get the executable name from has been crudded.
[20:57:40] <opeth__> are there any plans to enable serving nfs off os zones?
[20:58:12] <opeth__> that'd be something I'd miss migrating off solaris
[20:58:44] <jperkin> opeth__: I'm not aware of anyone currently working on it
[20:59:07] <jperkin> sounds like a great project to get involved with ;)
[20:59:31] <opeth__> I wish I could; I'm a mere sysadmin :/
[21:00:00] <richlowe> nexenta had someone working on it, and they were doing it in the obvious fashion
[21:00:05] <richlowe> I'm convinced the obvious fashion is flawed.
[21:00:15] <richlowe> (though it'd be nice if I could remember _why_ I think that...)
[21:02:35] <opeth__> seems I'll have to make do with cifs until this gets unleashed perhaps later on
[21:02:40] <opeth__> thanks ;]
[21:04:28] <opeth__> I'm planning the move from s11 to smartos on my playground, this seems to impose some new challenges
[21:06:03] *** dysinger has joined #smartos
[21:07:51] *** vsomes has joined #smartos
[21:16:57] <opeth__> yeah I've found a discussion online with Garrett referring to the Nexenta work having been done towards NFS in zones
[21:17:11] <opeth__> hope dies last
[21:17:24] <opeth__> same goes for integrated zfs crypto
[21:17:38] <opeth__> don't yet know what to do instead, apart from perhaps lofi
[21:21:26] *** Cpt-Oblivious has quit IRC
[21:22:26] *** miine has joined #smartos
[21:26:09] <Trixboxer> is there a place to submit smartos coredumps ?
[21:26:29] <rmustacc> What dumped core?
[21:26:38] <Trixboxer> nfs
[21:27:04] <rmustacc> A userland process or the operating system?
[21:27:08] <richlowe> the difference with zfs crypto is that nobody seems to care enough to work on it.
[21:27:08] <Trixboxer> it was mounted on client and did unshare
[21:27:19] <richlowe> whereas nfs-in-zones they've worked on but probably done wrong.
[21:27:26] <richlowe> ... that didn't come out as optimistic as I intended.
[21:31:42] <opeth__> thanks for the info, I'll bear with that for now
[21:32:24] <Trixboxer> rmustacc:  do think its of any use ?
[21:32:37] <Trixboxer> I tred to reproduce but dint work
[21:41:07] *** ipalreadytaken has joined #smartos
[21:45:32] *** ipalreadytaken has quit IRC
[21:52:09] *** bluezenix has joined #smartos
[22:17:24] *** Trixboxer has quit IRC
[22:17:31] *** denizr has quit IRC
[22:22:12] *** deirdres has joined #smartos
[22:25:53] *** denizr has joined #smartos
[22:31:02] *** trentm has joined #smartos
[22:32:40] *** theCzar has joined #smartos
[22:48:16] *** wolfeidau has quit IRC
[22:49:34] *** bens1 has joined #smartos
[22:56:25] *** denizr has quit IRC
[23:00:34] *** theCzar has quit IRC
[23:00:57] *** _0x783czar has joined #smartos
[23:03:48] *** des2 has quit IRC
[23:05:03] *** axonpoet has quit IRC
[23:05:37] <jesse_> jperkin, does libz.so link with libgcc.so, intentionally?
[23:06:20] <jesse_> jperkin, does libz.so link with libgcc_s.so, intentionally?
[23:08:28] *** bens1 has quit IRC
[23:50:25] *** _0x783czar has quit IRC
[23:53:06] *** wramthun has quit IRC
[23:53:20] *** jim80net has quit IRC
[23:53:41] *** axonpoet has joined #smartos
[23:53:46] <richlowe> Well, intentionally in what sense?
[23:54:29] <richlowe> basically everything will get linked to libgcc_s because the libgcc unwinder will get used, and, regardless, we don't have ld --as-needed
[23:54:37] <richlowe> it almost certainly doesn't really need it, though.
top

   March 15, 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 | >