Switch to DuckDuckGo Search
   December 23, 2011  
< | 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 | >

Toggle Join/Part | bottom
[00:00:28] <aszeszo> can't imagine upgrading such box if individual zoned had their own /usr
[00:00:38] <bcantrill> Yeah, no.
[00:00:59] <aszeszo> *zones
[00:02:00] <aszeszo> that's very impressive
[00:03:12] <rmustacc> So, maybe this is a stupid question. But if we had originally only done a ctfconvert/merge on the merged stabs/dwarf in a binary or so, what would we be losing?
[00:08:27] *** ryancnelson has joined #illumos
[00:10:28] <richlowe> I couldn't actually find anything, presuming -xs was used in the stabs case
[00:10:46] *** ryancnelson has left #illumos
[00:11:07] <richlowe> I had thought we may lose the ability to assert that each input object built from C source had data, but we still have the STT_FILE symbols, and the dwarf (at least) CU information has the source info, also.
[00:12:33] *** postwait has quit IRC
[00:13:23] *** CarlosC has quit IRC
[00:15:06] <rmustacc> Do we currently use -xs?
[00:15:36] <rmustacc> I guess, the answer is probably no, since nothing really is ever built with debug info that isn't ctf.
[00:15:47] *** mtzaurus has quit IRC
[00:15:50] *** voidcoder has quit IRC
[00:16:16] *** yakov_ has quit IRC
[00:19:59] <Shamgar> Hmmm. Lazy question here - I've got a fresh install of smartos, and the resolv.conf is right, and I can do a successful lookups with the host command (or dig) but any apps that actually try to do lookups fail. Ideas?
[00:20:06] <Shamgar> (i.e. ping fails, wget fails, etc)
[00:20:24] <elijah-mbp> is /etc/nsswitch.conf right
[00:20:25] <elijah-mbp> ?
[00:20:42] <elijah-mbp> e.g. hosts: files dns
[00:20:49] <Shamgar> right after I hit enter I thought of that, since that would make that kind of result.
[00:20:50] <elijah-mbp> or what have ye.
[00:20:52] <Shamgar> and no, it's not.
[00:20:56] <Shamgar> it's all files.
[00:21:04] <Shamgar> that's weird...is that just the default?
[00:21:20]
[00:21:28] *** voidcoder has joined #illumos
[00:21:46] * Shamgar nods.
[00:21:56] <Shamgar> okay - so long as I didn't do something stupid myself. (this time)
[00:22:00] <elijah-mbp> in our public cloud deployment, dns by default is off. we actually flip it on with chef in post-config.
[00:22:00] <rmustacc> It's a regression from a previous version.
[00:22:25] *** bcantrill has quit IRC
[00:22:49] *** auronandace has quit IRC
[00:24:08] <richlowe> rmustacc: we don't, no. We'd need to add -_cc=-xs to CTF_FLAGS
[00:30:30] <rmustacc> richlowe: Okay, thanks.
[00:32:14] <rmustacc> It seems that if we do that, then really what we need to do is just a single pass convert/unique on the generated object.
[00:32:36] <rmustacc> But I haven't really dived into too many of the details yet, so I'm sure there's going to be more to it than that.
[00:35:07] *** auronandace has joined #illumos
[00:38:00] *** zdunn has quit IRC
[00:44:29] *** _ams_ has quit IRC
[00:44:34] *** bcantrill has joined #illumos
[00:48:23] *** andy_js has quit IRC
[00:53:56] *** ryancnelson has joined #illumos
[00:53:58] *** bcantrill has quit IRC
[00:54:08] *** ryancnelson has left #illumos
[00:56:14] <LeftWing> Based purely on the quantity of mail, today feels a busy day.
[01:00:46] <richlowe> all I've done is ask for re-review of rmustacc's kstat fix, and tell people to stop being losers.
[01:02:36] <LeftWing> Perhaps it's a poor metric then. :)
[01:02:40] <e^ipi> LeftWing: busy implies work. email is the opposite of that
[01:02:57] <LeftWing> Indeed.
[01:03:29] <LeftWing> richlowe: I'm running some gcc4 bits now. hooray.
[01:04:00] <richlowe> the more the merrier.
[01:04:40] <LeftWing> AMD KVM module still works, too.
[01:05:12] <bdha> !
[01:05:29] <LeftWing> ?
[01:05:33] <bdha> .
[01:05:38] <LeftWing>
[01:05:42] <e^ipi> and being told i suck by rich, apparently
[01:05:42] <richlowe> rmustacc: when we talked about things to do to CTF I missed part of an aside out. Adam and I had been talking about change to check the ctf parent label, and he suggested that perhaps CTF uniquification against CTF genunix and ip may no longer win sufficiently to be worthwhile. Especially in the dwarf case (where we have less type data, as unused types are dropped).
[01:05:54] <richlowe> rmustacc: so amongst my end goals is to try that.
[01:06:07] <richlowe> rmustacc: I got screwed by studio's dwarf being busted, but I still plan to measure with dwarf that isn't.
[01:06:11] <richlowe> e^ipi you answered my question, so no.
[01:06:19] <richlowe> e^ipi: as I said, arguing and _not_ answering is sucking.
[01:08:39] <richlowe> LeftWing: KVM doesn't build with gcc4 though, unless Joyent fixed that.
[01:08:46] <richlowe> for problems all on that side, iirc.
[01:08:55] <richlowe> aszeszo pasted all the errors, and they were all legitimate as best as I recall.
[01:09:01] <LeftWing> Yeah, I haven't tried that yet.
[01:09:05] <richlowe> artifacts of the port, inlines prototyped but never declared, stuff liket hat.
[01:09:26] <e^ipi> richlowe: i'm getting to it
[01:09:29] <e^ipi> eventually
[01:09:39] <richlowe> e^ipi: Sure, just seemed polite to warn him.
[01:11:16] <rmustacc> I can add it into the never-ending list of todos.
[01:11:33] <richlowe> that'd make it on 3 such lists (me, John, you).
[01:15:30] <rmustacc> A part of me just wants to beat on gcc4 and slowly forget Studio.
[01:15:54] <rmustacc> Or not so slowly even.
[01:16:10] <e^ipi> lord knows a big part of me does
[01:16:21] <e^ipi> which is why every smartos user is now testing the bits
[01:16:34] <e^ipi> and one of the build boxes
[01:17:59] *** ganbold has joined #illumos
[01:19:39] <richlowe> John just likes to keep me terrified, really.
[01:19:47] <richlowe> he has his excuses, sure, but ultimately, it's all about the terror.
[01:22:55] <LeftWing> He doesn't seem, in person, to be particularly terrifying. But perhaps that's the most terrifying part.
[01:23:02] <e^ipi> it's certainly a fringe benefit
[01:24:08] *** mfernest has left #illumos
[01:27:33] *** tsoome has quit IRC
[01:28:06] *** tsoome has joined #illumos
[01:30:01] *** CarlosC has joined #illumos
[01:31:09] *** tsoome1 has joined #illumos
[01:31:10] *** tsoome has quit IRC
[01:31:10] *** tsoome1 is now known as tsoome
[01:33:10] <bdha> Sneaky freakin' Canadians...
[01:34:57] <marsell> They had to be sneaky to set the Whitehouse on fire.
[01:35:00] <_Tenchi_> nevertrustem
[01:42:16] *** postwait has joined #illumos
[01:59:21] *** DrLou has joined #illumos
[02:05:49] *** raichoo has quit IRC
[02:21:06] *** nachox has joined #illumos
[02:21:06] *** nachox has joined #illumos
[02:21:58] *** Darkproger has quit IRC
[02:51:54] <danmcd> WIsh me luck, I'm rebooting kebe.com now. :)
[02:52:01] <nahamu> good luck!
[02:54:54] *** danmcd` has joined #illumos
[02:55:18] <danmcd`> dammmit.
[02:55:26] <danmcd`> IRC won't let me be me again.
[02:55:28] *** danmcd` is now known as kebe
[02:55:31] <kebe> THere we go.
[02:56:29] *** danmcd has quit IRC
[02:57:41] <jbk> if you register your nick, you can ghost the other clients
[02:57:51] <jbk> (basically disconnects the other sessions)
[02:58:30] <jbk> also works if someone so rudely tries to connect using your nick :)
[02:58:53] <kebe> neuromancer(~)[0]% uname -a
[02:58:54] <kebe>
[02:59:02] <kebe> SunOS neuromancer 5.11 1918 i86pc i386 i86pc
[02:59:09] <kebe> neuromancer(~)[0]%
[02:59:21] <kebe> Can anyone visit www.kebe.com to see if bits move there as well?
[02:59:36] * kebe wants someone to verify nat shuffling still works.
[02:59:42] *** kebe is now known as danmcd
[03:01:30] <nachox> works for me
[03:01:56] *** ajkn has quit IRC
[03:02:08] <nachox> though you might want to change that powered by opensolaris logo :)
[03:02:19] *** ajkno has joined #illumos
[03:04:03] <jbk> danmcd: seems to be working..
[03:04:39] <nahamu> danmcd: I see the site from here.
[03:05:20] *** CarlosC has quit IRC
[03:06:22] <danmcd> Thank you folks.
[03:07:05] <danmcd> I have powered-by-Illumos drafts, but the "dark background" version needs to match the "light background" version and I need community sign-off before I publish them.
[03:07:52] <nachox> nice
[03:07:56] <danmcd> I may spend a little of my vacation time next week dinking 'em.
[03:09:47] <nachox> kebe.com looks pre 1995 btw
[03:10:39] <nachox> which, considering how web pages have grown these days, i'd be happy to see others mirror
[03:12:03] <jbk> nachox: what, you don't think <1k of text should be conveyed by less than 10mb of html, css, and javascript? :)
[03:12:46] <jeffpc> so, I just zfs send | recv'd my rpool (original pool was rpool2, destination is rpool); then I installed grub on the new one, halt, remove disk with rpool2
[03:13:11] <jeffpc> booting off the new pool works, I had to fix up dumpadm's device
[03:13:15] <nachox> jbk, i work as a security consultant, the more code there is in their web pages, the better the chance i'll find a way in, so i wont complain :)
[03:13:22] <jeffpc> anything else I should take fix?
[03:16:20] <Alasdairrr> morning all
[03:16:26] * Alasdairrr can't sleep
[03:16:57] <herzen> hi guy
[03:17:18] <Alasdairrr> hi herzen
[03:17:39] <herzen> how's the weather in London? damp?
[03:18:02] *** ajkn has joined #illumos
[03:18:20] <herzen> I was just listening to Lush.
[03:18:47] <Alasdairrr> It's dry!
[03:18:52] <Alasdairrr> i think
[03:19:37] <herzen> the main reason I respect you Brits is that no one equals you when it comes to pop.
[03:19:50] *** ajkno has quit IRC
[03:20:56] *** proberts has quit IRC
[03:23:24] <danmcd> nachox: "Edited with either emacs, vi, or cat!"
[03:24:31] <jbk> my friend used to had a 'made with vi' logo or such on his website
[03:25:29] *** miine has quit IRC
[03:26:20] *** enmand has joined #illumos
[03:29:45] <danmcd> I've seen logos for each of emacs, vi, and cat + Control-D.
[03:30:07] *** proberts has joined #illumos
[03:30:10] <danmcd> Ahhh, I can actually join the Mrs. this evening. later!
[03:30:11] *** danmcd has quit IRC
[03:30:29] <Triskelios> I'd be more impressed if the logos were hand-edited XPMs
[03:31:52] *** miine has joined #illumos
[03:33:54] <Alasdairrr> "Made in od"
[03:35:04] *** nemysis has quit IRC
[03:35:21] *** trn has quit IRC
[03:37:41] *** trn has joined #illumos
[03:43:29] <jelmd> "ssl_support.cc", line 185: Error: Different types for "?:" (extern "C" void(*)(GENERAL_NAME_st*) and void(*)(GENERAL_NAME_st*)).
[03:43:42] <jelmd> any hint, how to convince CC to go ahead ?
[03:43:56] <richlowe> either make both, or neither, extern "C"
[03:44:26] <jelmd> hmm - it is openssl stuff
[03:45:02] <jelmd> somewher burried in /usr/include/openssl :(
[03:48:34] *** eschrock has joined #illumos
[03:50:17] <jelmd> perhaps I should just go with gcc and close my eyes :(
[03:50:22] *** eschrock has quit IRC
[03:53:33] *** POloser has joined #illumos
[04:07:15] <jelmd> richlow: can you give me a hint? It results into:
[04:07:29] <jelmd> sk_pop_free(CHECKED_STACK_OF(GENERAL_NAME, altnames), ((void (*)(void *)) (1 ? GENERAL_NAME_free : (void (*)(GENERAL_NAME *))0)));
[04:09:23] <jbk> what are you compiling?
[04:09:46] <richlowe> I don't actually remember if you can specify extern "C" in a cast.
[04:11:18] <jelmd> jbk: squid
[04:12:06] *** alanc has quit IRC
[04:12:13] *** mjanthony has left #illumos
[04:12:23] <jbk> that code just seems a bit horrid :)
[04:13:10] <jelmd> yes, unbelievable, that the changed to c++ - when it was C, everything was fine ...
[04:17:35] *** |wilsonkk| has quit IRC
[04:19:09] *** nachox has quit IRC
[04:20:33] *** freakazoid0223 has joined #illumos
[04:21:42] *** alanc has joined #illumos
[04:28:31] *** kart_ has joined #illumos
[04:29:55] *** enmand has quit IRC
[04:30:34] *** |wilsonkk| has joined #illumos
[04:33:30] *** zdunn has joined #illumos
[04:35:26] *** trn has quit IRC
[04:38:37] *** trn has joined #illumos
[04:42:03] *** trn has quit IRC
[04:42:47] *** trn has joined #illumos
[04:43:02] *** freakazoid0223 has quit IRC
[04:43:26] *** mavhc has quit IRC
[04:44:58] *** marsell has quit IRC
[04:46:10] *** freakazoid0223 has joined #illumos
[04:47:48] *** mavhc has joined #illumos
[05:20:55] *** rmustacc has left #illumos
[05:29:14] *** Botanic has quit IRC
[05:29:28] *** zdunn has quit IRC
[05:38:07] *** libkeiser has joined #illumos
[05:41:17] *** libkeise1 has quit IRC
[05:43:09] *** libkeise1 has joined #illumos
[05:43:50] *** libkeiser has quit IRC
[05:45:25] *** zdunn has joined #illumos
[05:58:16] <timclassic> I'm getting nasty sound skips when resizing evince windows and some other tasks, and I was thinking about investigating with dtrace, but I'm not sure where to begin.
[06:06:18] *** radu has quit IRC
[06:09:17] *** keremet has joined #illumos
[06:13:59] *** hando has joined #illumos
[06:15:45] *** zdunn has quit IRC
[06:34:10] *** libkeiser has joined #illumos
[06:35:00] *** descipher_ has joined #illumos
[06:36:30] *** libkeise1 has quit IRC
[06:36:36] *** descipher has quit IRC
[06:36:42] *** descipher_ is now known as descipher
[06:42:28] *** DrLou has quit IRC
[06:46:11] <richlowe> so very many bugs.
[06:48:07] *** bcantrill has joined #illumos
[06:50:39] <bdha> richlowe: Indeed.
[06:57:52] *** bcantrill has quit IRC
[07:05:38] *** igork has quit IRC
[07:08:27] *** gdamore has quit IRC
[07:12:44] *** gdamore has joined #illumos
[07:12:45] *** ChanServ sets mode: +o gdamore
[07:18:20] *** Botanic has joined #illumos
[07:18:20] *** Botanic has quit IRC
[07:18:20] *** Botanic has joined #illumos
[07:37:14] *** igork has joined #illumos
[07:42:09] *** bcantrill has joined #illumos
[07:45:30] *** Botanic has quit IRC
[07:54:36] *** proberts1 has joined #illumos
[07:57:10] *** proberts has quit IRC
[08:09:42] *** _ams_ has joined #illumos
[08:12:29] *** markc13 has quit IRC
[08:12:45] *** Worsoe has joined #illumos
[08:14:23] *** markc13 has joined #illumos
[08:14:27] *** voidcoder has quit IRC
[08:19:13] <richlowe> Current nemeses: People who succeed in not passing -v to cc
[08:19:21] <richlowe> "How can this code be wrong, there'd be warnings?!"
[08:21:39] *** mavhc has quit IRC
[08:22:23] *** mavhc has joined #illumos
[08:26:23] <richlowe> https://www.illumos.org/issues/1926
[09:35:43] *** |AbsyntH| has joined #illumos
[09:37:04] *** bcantrill has quit IRC
[09:49:04] *** nikolam has quit IRC
[09:50:29] *** keremet has quit IRC
[09:52:28] *** keremet has joined #illumos
[09:53:07] *** nikolam has joined #illumos
[09:58:45] *** mtzaurus has joined #illumos
[10:03:31] *** alhazred has joined #illumos
[10:10:30] *** nikolam has quit IRC
[10:10:42] *** tsoome has quit IRC
[10:15:38] *** lblume has quit IRC
[10:36:21] *** gdamore has quit IRC
[10:44:29] *** tsoome has joined #illumos
[10:57:30] *** yakov__ is now known as yakov
[11:20:42] *** hajma has joined #illumos
[11:21:00] *** libkeise1 has joined #illumos
[11:23:41] *** libkeiser has quit IRC
[11:25:12] *** _ams_ has quit IRC
[11:27:46] *** libkeiser has joined #illumos
[11:28:07] *** _ams_ has joined #illumos
[11:31:46] *** libkeise1 has quit IRC
[11:37:19] *** feizhou_ has joined #illumos
[11:41:11] *** feizhou__ has quit IRC
[12:01:56] *** andy_js has joined #illumos
[12:03:21] *** tnn_ has joined #illumos
[12:14:40] *** nemysis has joined #illumos
[12:20:30] *** |AbsyntH| has quit IRC
[12:22:30] *** hajma has quit IRC
[12:23:53] *** c2n2 has joined #illumos
[12:24:20] *** DreamCatcher has quit IRC
[12:25:49] *** DreamCatcher has joined #illumos
[12:26:29] *** mtzaurus has quit IRC
[12:58:38] *** keremet has left #illumos
[13:06:38] *** tsoome has quit IRC
[13:08:43] *** hajma has joined #illumos
[13:12:25] *** mtzaurus has joined #illumos
[13:29:17] *** POloser has left #illumos
[13:39:27] *** |AbsyntH| has joined #illumos
[13:53:59] *** tnn_ has quit IRC
[14:02:47] *** randw has joined #illumos
[14:12:19] *** enmand has joined #illumos
[14:25:49] *** voidcoder has joined #illumos
[14:30:29] *** postwait has quit IRC
[14:30:38] *** randw has quit IRC
[14:32:25] *** randw has joined #illumos
[14:43:08] *** spanglywires has joined #illumos
[14:43:30] *** tsoome has joined #illumos
[14:51:04] *** alhazred has quit IRC
[14:51:23] *** alhazred has joined #illumos
[15:04:07] *** andy_js has quit IRC
[15:07:26] *** Worsoe has quit IRC
[15:10:38] *** mtzaurus has quit IRC
[15:19:34] *** gdamore has joined #illumos
[15:19:34] *** ChanServ sets mode: +o gdamore
[15:23:20] *** voidcoder_ has joined #illumos
[15:23:21] *** randw_ has joined #illumos
[15:25:41] *** nachox has joined #illumos
[15:25:45] *** randw has quit IRC
[15:26:52] *** voidcoder has quit IRC
[15:28:16] *** voidcoder_ has quit IRC
[15:28:35] *** voidcoder has joined #illumos
[15:29:20] *** voidcoder has quit IRC
[15:29:31] *** voidcoder has joined #illumos
[15:34:18] *** mavhc has quit IRC
[15:35:43] *** mavhc has joined #illumos
[15:41:16] *** DrLou has joined #illumos
[15:42:59] *** yakov has quit IRC
[15:43:10] *** CarlosC has joined #illumos
[15:44:08] *** voidcoder has quit IRC
[15:44:27] *** voidcoder has joined #illumos
[15:51:33] *** man_u has quit IRC
[15:54:34] <miine> Hi. I'm trying to get apcupsd working correctly on oi 151a . problem: the "kill the ups power" script provided by apcups isn't fitting anymore into /sbin/rc0 . It should be applied after some umounts , but this seems to be handled by smf now.
[15:54:42] <miine> so: is there an official way smf does support ups power-down scripts?
[15:59:47] *** zdunn has joined #illumos
[15:59:57] *** zdunn has joined #illumos
[16:02:05] *** eschrock has joined #illumos
[16:06:28] *** voidcoder has quit IRC
[16:06:43] *** postwait has joined #illumos
[16:09:35] *** jurikm has joined #illumos
[16:10:21] *** PiotrSikora has quit IRC
[16:11:15] *** PiotrSikora has joined #illumos
[16:11:43] *** yakov has joined #illumos
[16:12:17] *** movement has quit IRC
[16:13:03] *** herzen has quit IRC
[16:13:56] *** voidcoder has joined #illumos
[16:16:37] *** randw_ has quit IRC
[16:27:54] <e^ipi> miine: i don't know what the script is meant to do. run continuously?
[16:31:05] *** |AbsyntH| has quit IRC
[16:32:29] *** nachox has quit IRC
[16:33:11] *** andy_js has joined #illumos
[16:34:12] <miine> e^ipi: the old solaris /sbin/rc0 script did umount the volumes. after that the server could be shut down. this part is now in svc.startd (function do_uadmin). so it would be nice to have a hook into that or function uadmin to execute a binary which pulls the ups power...
[16:34:33] <e^ipi> "pulls the ups power" ?
[16:34:50] <e^ipi> i don't understand what you mean by that
[16:35:23] *** andy_js has quit IRC
[16:36:27] <miine> e^ipi: the ups runs on battery. the ups daemon invoces a shutdown. at the end of the shutdown procedure the ups power must be pulled before the mainboard sends it signal to the power supply to power off. otherwise it won't turn on when the ups is on power line again...
[16:37:22] <e^ipi> ahh, i see what you mean
[16:37:28] <miine> and that was done in the apcupsd scripts in /sbin/rc0 after the umounts. these are now in svc.startd and everything like the rc* scripts are called before...
[16:38:01] *** danmcd has joined #illumos
[16:38:21] <miine> so there seems no clean way to use an ups on smf based systems...
[16:38:41] <miine> at least without patching svc.startd ...
[16:38:47] <tsoome> why not? you still can build up dependency tree
[16:39:18] <tsoome> so that the upsoff will be sent near last thing
[16:39:25] <miine> tsoome: look at http://src.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/cmd/svc/startd/graph.c#3615
[16:39:34] *** johest has quit IRC
[16:40:05] <miine> tsoome: if I send the ups kill power signal in an rc script, there are thing to be done after that..
[16:40:30] *** evanl has joined #illumos
[16:40:47] <miine> tsoome: so the command should be executed in uadmin(A_SHUTDOWN, halting, mdep) ...
[16:41:18] <tsoome> not really, you are overthinking:P
[16:41:34] <miine> tsoome: what should I do? googled and found nothing...
[16:41:47] <tsoome> as there is an dependency tree to *start* the services, the same tree will apply for stopping them
[16:43:13] <tsoome> also, even if you are using rc script, you can check how far the shutdown is from smf, and send upsoff then. itsnot too clean tho
[16:43:48] <tsoome> if i recall the poweroff for ups was not really immediate as well, so you still have some time to finish cleanup
[16:43:49] <miine> tsoome: the code isn't started at a legacy service. ok, I could change that. but the problem is: the services are shutdown BEFORE svc.startd does some other cleanup things. so I either have to replicate that behaviour in case of an emergency shutdown or there will be something missing.
[16:44:11] <miine> tsoome: sorry: at the moment the code is start as a legacy service (rc script)
[16:44:24] <tsoome> well, then the code is *broken*
[16:44:30] <tsoome> and needs to be fixed.
[16:44:46] <miine> tsoome: which code? for me it's svc.startd which is broken...
[16:44:58] <tsoome> why?
[16:45:26] <tsoome> because now suddenly the rc scripts are not the last thing to be run?
[16:45:34] <miine> it does things which won't be done if an service would kill the power...
[16:46:02] <miine> tsoome: look at the code. it's exactly the case. the umount code was in /sbin/rc0 and is now in svc.startd ...
[16:46:29] <tsoome> rc scripts were ok before SMF, its nearly 7 years since SMF was introduced, get over it:P
[16:47:11] <miine> tsoome: I have no problem with that (just using the rc scripts because I don't know which exact dependencies I have to specify).
[16:47:45] <miine> but the problem remains: svc.startd will do things after the last service is stopped. and if I kill the power before that code won't run...
[16:48:02] <tsoome> but, you are correct about startd, in sense that upsoff support should be there
[16:48:59] <miine> it's the same in macos x : launchd was (or still is) braindead in the same way... I wrote an ups daemon and had to replicate the whole shutdown code of launchd...
[16:51:34] <miine> in my view there should be a special ups service - so that startd could communicate with it. if the os starts up while on battery all other services don't need to be started so it could go directly into shutdown mode...
[16:52:34] <miine> also putting that into startd and a special service could ensure that on shutdown everything is loaded for the ups daemon to shutdown even after everything is umounted...
[16:52:38] <tsoome> uhm, i still fail to see why it cant be done as "normal" service tho
[16:53:04] <miine> tsoome: did you look at the code?
[16:53:20] <tsoome> if the ups service is the first thing to be started, then it will be last thing to be stopped
[16:54:24] <miine> but svc.startd does something with the system AFTER the last service is stopped. on a power failure the ups service would kill the power when it is stopped, so svc.startd won't run completly..
[16:54:46] <tsoome> the ups power kill wasnt immediate
[16:55:04] <miine> tsoome: the ups power kill is immediate :-(
[16:55:24] <tsoome> are you sure? i recall they had some time setting there
[16:55:57] <miine> tsoome: as there a "dumb" ups which won't power off if the power comes back during a shutdown...
[16:56:29] <miine> tsoome: some people did a ups killpower; sleep 60; reboot sequence in /sbin/rc0 because of that ...
[16:57:14] <miine> as ups is such a basic feature in power managment it's a shame it's not in svc.startd ...
[16:57:19] <tsoome> in that case you need to integrate ups poweroff code into do_uadm()
[16:57:31] <tsoome> do_uadmin()
[16:57:51] <miine> tsoome: yep. just thought about that...
[16:58:09] <tsoome> well… we dont have any ups...
[16:58:18] <e^ipi> patch it in a way that doesn't break everyone else, integrate it
[16:58:18] <tsoome> we have power generators
[16:59:30] <tsoome> of course the problem with USP is, every ups is kind of its own, so some config facility is needed:P
[16:59:48] <miine> but at the moment I need a "quick hack". otherwise that would be a project of its own as to do it right needs some thinking...
[17:00:15] *** DreamCatcher has quit IRC
[17:00:30] <miine> not really. svc.startd should just communicate with a special ups service. so every ups daemon could use that "api"...
[17:01:02] <miine> or with some predefined binaries...
[17:01:30] <storkone> miine: Please look at this page from the NUT docs http://www.networkupstools.org/docs/man/apcsmart.html. You can program the ups to kill power after XXX seconds. This setting is in the ups regardless of the software which controls the ups.
[17:01:49] <miine> ok - at the moment I feel lucky that 151a doesn't have the usb bug of 148b which prevents from communicating over usb...
[17:02:52] <tsoome> storkone: exactly thats what I did remember
[17:02:53] <miine> storkone: the problem is that won't work with every ups. as there are some which won't go thru an complete power cycle in case the main power comes back...
[17:03:48] <storkone> This final grace period allows you to clean up and even powerdown your machine. That's explained as well in the page. That's the difference between a hard and softwa hibernate.
[17:04:22] <storkone> I didn't invent these descriptions...
[17:06:13] <miine> storkone: so that would work for the ups I have here. but it isn't assured it will work with every ups. I think there should be a mechanism in svc.startd to ensure just that...
[17:07:05] <miine> I have read many messages regarding smf based *solaris and ups shutdown. no real solution found so far...
[17:07:05] *** arethusa has quit IRC
[17:07:15] <tsoome> the thing is, there is no clean way to do it.
[17:07:33] *** nachox has joined #illumos
[17:07:33] *** nachox has joined #illumos
[17:07:38] <storkone> It's the softhibernate you want to perform. That way the ups returns when power comes back AND is charged up to a level set by you (also in firmware)
[17:07:41] *** arethusa has joined #illumos
[17:13:11] <storkone> I've got no issues at all with nut. Even a very busy network can be shutdown in order. But I must admit that this depends on a grace period in the ups, which (some) APC ups models provide. After I issue the soft hibernate command some final cleanup is done by openindiana if that takes to long the ups will kill the power. But the machine is gone anyway because the network services are already shutdown etc.
[17:14:08] *** jurikm has quit IRC
[17:16:27] <miine> I think the correct way would be to execute a binary on shutdown and then signal when it is ok to kill the power after everything else is done. as the binary has access to the whole filesystem when it is started it could check if it the ups is on battery or not and also load everything to communicate with the ups into ram...
[17:16:28] *** kart_ has quit IRC
[17:17:38] <storkone> Sorry I'm off.
[17:17:40] *** carlomagno has quit IRC
[17:19:48] *** carlomagno has joined #illumos
[17:27:17] *** herzen has joined #illumos
[17:28:33] *** davemq has joined #illumos
[17:34:49] *** nachox has quit IRC
[17:39:24] *** jurikm has joined #illumos
[17:40:57] *** andy_js has joined #illumos
[17:43:30] *** voidcoder has quit IRC
[17:46:23] *** voidcoder has joined #illumos
[17:46:50] *** voidcoder has quit IRC
[17:47:24] *** voidcoder has joined #illumos
[17:52:11] *** nemysis has quit IRC
[17:52:45] *** nemysis has joined #illumos
[17:55:45] *** voidcoder has quit IRC
[17:57:20] *** spanglywires has quit IRC
[18:05:49] *** raichoo has joined #illumos
[18:09:04] *** man_u has joined #illumos
[18:10:23] <jeffpc> what's an easy way to figure out the offset of a partition on a disk? (I want to make sure it's aligned properly)
[18:16:00] *** hajma has quit IRC
[18:17:21] <nahamu> e^ipi: the smartos wiki page about building smartos makes no mention of downloading studio, but the readme in smartos-live does. Is studio still needed to build smartos-live?
[18:17:34] <e^ipi> yes. for lint
[18:17:37] <e^ipi> :-/
[18:17:42] <nahamu> gah! lint!!
[18:17:52] <nahamu> thanks
[18:17:56] <e^ipi> np
[18:18:16] <e^ipi> i haven't looked if there's a way to disable that yet though
[18:25:47] *** alhazred has quit IRC
[18:32:04] *** bcantrill has joined #illumos
[18:35:28] *** CarlosC has quit IRC
[18:40:09] *** bcantrill has quit IRC
[18:40:31] *** bcantrill has joined #illumos
[18:43:20] <bdha> moo.
[18:46:30] *** mtzaurus has joined #illumos
[18:46:53] *** davemq has quit IRC
[18:47:03] *** kart_ has joined #illumos
[18:47:18] *** lblume has joined #illumos
[18:50:03] *** bcantrill has quit IRC
[19:14:39] <nahamu> moo?
[19:14:58] <nahamu> did you ever get a dtrace enabled mysql running?
[19:18:30] *** jurikm has left #illumos
[19:28:35] *** voidcoder has joined #illumos
[19:33:51] *** kart_ has quit IRC
[19:40:26] *** randw has joined #illumos
[19:54:36] *** _wilsonkk_ has joined #illumos
[19:57:30] *** |wilsonkk| has quit IRC
[19:58:51] *** tsoome has quit IRC
[19:59:07] *** tsoome has joined #illumos
[20:07:46] *** drwrr has joined #illumos
[20:25:48] *** enmand has quit IRC
[20:26:11] *** enmand has joined #illumos
[20:26:52] *** Botanic has joined #illumos
[20:26:54] *** Botanic has quit IRC
[20:26:54] *** Botanic has joined #illumos
[20:29:09] *** rbarraud has quit IRC
[20:33:14] *** xinkeT has quit IRC
[20:37:42] *** danmcd has quit IRC
[20:40:50] *** drwrr has quit IRC
[20:48:03] *** cyanogene has joined #illumos
[20:51:41] *** c2n2 has quit IRC
[20:58:03] *** drwrr has joined #illumos
[21:07:24] *** freakazoid0223 has quit IRC
[21:25:35] *** pizdec has quit IRC
[21:29:49] *** cyanogene has quit IRC
[21:40:22] *** Botanic_ has joined #illumos
[21:40:29] *** Botanic has quit IRC
[21:51:11] *** nemysis has quit IRC
[21:52:57] *** nemysis has joined #illumos
[21:56:25] *** freakazoid0223 has joined #illumos
[21:57:15] *** enmand has quit IRC
[22:05:41] *** smrt has joined #illumos
[22:09:46] *** evanl has left #illumos
[22:10:17] <dre2kse> Merry Christmas to all
[22:13:30] *** randw has quit IRC
[22:16:45] *** spanglywires has joined #illumos
[22:28:50] *** agrajag has quit IRC
[22:34:38] *** lblume has quit IRC
[22:36:23] *** agrajag has joined #illumos
[22:36:23] *** spanglywires has quit IRC
[22:38:03] *** d_k has left #illumos
[22:39:26] *** d_k has joined #illumos
[22:45:48] <herzen> still a couple of days to go. IRC is real time. :)
[22:47:52] <herzen> and then there our Orthodox friends, who are still on the old calendar...
[22:51:36] *** Darkproger has joined #illumos
[22:51:54] *** postwait has quit IRC
[22:53:23] *** arethusa has quit IRC
[22:53:56] *** Botanic has joined #illumos
[22:53:56] *** Botanic has quit IRC
[22:53:56] *** Botanic has joined #illumos
[22:54:14] *** arethusa has joined #illumos
[22:54:18] *** Botanic_ has quit IRC
[22:54:56] <e^ipi> herzen: and it won't be christmas for them for almost 3 weeks
[22:56:20] <lystra> Triskelios: you there?
[23:02:20] *** lblume has joined #illumos
[23:10:29] *** matt_c has quit IRC
[23:10:38] *** zdunn has quit IRC
[23:12:09] *** yakov_ has joined #illumos
[23:12:27] *** matt_c has joined #illumos
[23:12:57] *** smrt has quit IRC
[23:13:12] *** smrt has joined #illumos
[23:16:59] *** yakov has quit IRC
[23:17:05] *** yakov_ is now known as yakov
[23:17:32] *** yakov_ has joined #illumos
[23:20:55] *** matt_c has quit IRC
[23:21:17] *** matt_c has joined #illumos
[23:24:10] *** carlomagno has quit IRC
[23:24:26] *** carlomagno has joined #illumos
[23:28:37] *** matt_c has quit IRC
[23:30:24] *** matt_c has joined #illumos
[23:36:59] *** andy_js has quit IRC
[23:38:52] <richlowe> LeftWing: opengrok doesn't find typedefs in rpcgen input (foo.x)
[23:38:58] *** gdamore has left #illumos
[23:39:00] <richlowe> not sure if that's easy to rectify or not.
[23:41:29] *** arethusa has quit IRC
[23:41:45] *** arethusa has joined #illumos
[23:42:50] *** arethusa has quit IRC
[23:43:44] *** arethusa has joined #illumos
[23:47:01] *** lblume has quit IRC
[23:51:18] *** eschrock has quit IRC
top

   December 23, 2011  
< | 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 | >