December 27, 2012 
< | 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 | >
December 27, 2012 [00:00:24] <lundh> "10.0.1.10 netmask 255.255.255.0 up" is all in that file[00:00:37] <lundh> never mind[00:00:48] <lundh> no, thats right[00:11:05] <lundh> and it still doesnt work as a time machine server...[00:13:20] <lundh> /etc/nodename ![00:16:02] <konobi> ah yes[00:16:43] <lundh> now I just have to figure out why it wont work as a time machine server[00:17:59] <lundh> and get flexget installed and a few other things but I'm feeling better about it for every minute[00:25:57] <lundh> time machine server done[00:35:56] *** tonyarkles has quit IRC[00:43:33] *** gkyildirim has quit IRC[00:59:34] *** jamesd has quit IRC[00:59:48] *** Livid has joined #smartos[01:00:18] <Livid> How can I set hostname for SmartOS booted from USB?[01:03:46] <Alasdairrr> edit /usbkey/config[01:04:00] <Alasdairrr> set hostname=[01:07:36] <konobi> lundh: you may need to change a setting on your mac... "Secrets.app" ftw[01:09:37] <Livid> @Alasdairrr thanks. Is there a page about all possible keys in /usbkey/config?[01:10:13] <Alasdairrr> Livid: i'm not sure actually, hmm[01:10:25] <Alasdairrr> there's a lot of info over at wiki.smartos.org[01:11:00] <Livid> I'll try now, installing a new box now ;)[01:11:56] *** Cpt-Oblivious has quit IRC[01:13:20] <konobi> yup... "extra configuration options" is on the wiki[01:15:21] *** jamesd has joined #smartos[01:15:37] <Livid> Got it. thanks. http://wiki.smartos.org/display/DOC/Setting+a+static+hostname+at+boot+time[01:16:01] <Livid> hostname in /usbkey/config doesn't work[01:16:12] <Livid> I'll try the SMF method[01:18:06] <konobi> http://wiki.smartos.org/display/DOC/extra+configuration+options[01:19:27] <Livid> konobi: thank you.[01:19:28] *** Livid has left #smartos[01:19:52] *** Livid has joined #smartos[01:21:27] *** newlix_ has joined #smartos[01:21:27] *** newlix has quit IRC[01:23:13] <konobi> Livid: hostname in the config file should work[01:23:37] <konobi> you can check generic/lib/svc/method/identity-node[01:23:55] <Livid> I have hostname=vh0 in /usbkey/config and rebooted, and my hostname is still MAC address[01:26:15] <konobi> oh... looks like it'a a headnode=true thing[01:26:27] <konobi> you can also set it on the boot line[01:38:17] *** ryancnelson has joined #smartos[01:43:18] <konobi> howdy ryan[01:44:12] <ryancnelson> hey[01:47:20] *** xmerlin has joined #smartos[01:54:31] <lundh> konobi: no need, I get netatalk to speak fluent OS X 10.8 :)[01:54:40] <lundh> *got[01:55:44] <lundh> some issues with transmission though, doing something wrong when trying to start it as a service[01:56:52] <konobi> lundh: nice[01:57:18] <lundh> how do I unregister a service?[01:57:33] <konobi> svcadm disable[01:57:47] <lundh> remote the manifest from the known one[01:58:54] <ryancnelson> having it there, but disabled hurts nothing.[01:59:03] <ryancnelson> why do you want to "unregister" it?[01:59:25] <lundh> cause I made it myself and it is not wrking[01:59:28] *** Vod has joined #smartos[01:59:37] <ryancnelson> svccfg delete <fmri>[01:59:46] <lundh> fmir?[02:00:08] <lundh> ah[02:00:10] <lundh> got it[02:00:17] <ryancnelson> Fault Management Resource Indicator[02:00:39] <ryancnelson> ... it's basically the same thing as "URL", but talking about an smf thing, instead of a website-location[02:02:50] <lundh> got it :)[02:05:29] <konobi> mmm... looks like you can't create a file backed zpool in a zone[02:05:48] <ryancnelson> there's a list of allowed filesystems in the zone[02:05:55] <ryancnelson> zfs isn't one of 'em, by default[02:06:12] <ryancnelson> ... actually, might be other reasons why, though, w/r/t zfs[02:12:35] <konobi> ryancnelson: yeah, was for a jpc zone[02:12:57] <konobi> wanted to take advantage of zfs compression[02:16:05] <ryancnelson> in jpc, you'll basically never get FS passthru. fs == kernel, and that means "any jerk can crash the box"[02:18:14] <ryancnelson> if you ask nicely, though, an op might turn it on for you :)[02:18:33] * konobi flutters eyelids at ryan[02:19:21] * wesolows walked in at the wrong time, departs[02:19:53] <xmerlin> the password stored in SINGLE_USER_ROOT_PASSWORD.20121213T231116Z.txt doesn't work ...why?[02:20:10] <xmerlin> I've started the image with noimport=true[02:20:44] <ryancnelson> you tried the one you set earlier?[02:20:52] <ryancnelson> (or are you trying to reset that?)[02:21:11] <konobi> shadow on the boot line?[02:21:39] *** Vod has quit IRC[02:22:06] <ryancnelson> ... yeah, what's the shadow passwd in the grub menu item look like? it should (i think) start with "$5$...."[02:22:15] <ryancnelson> it was messed up in the dec 3 build[02:22:40] <ryancnelson> you're using the 20121213T231116Z build, right? 'cause that password changes, every build[02:23:10] *** codingstream has joined #smartos[02:23:39] <xmerlin> ryancnelson, I had some problems "installing" smartos on some disks with old zfs headers ...now it doesn't boot fine ...and I want to wipe the begin and the end of the disks to reinstall it[02:25:20] <ryancnelson> oh... then try just booting with "destroy_zpools=true"[02:25:25] <ryancnelson> in grub[02:25:54] <xmerlin> ok[02:25:58] <ryancnelson> you won't need to dd onto the disks, it should import them and then destroy them[02:26:13] <ryancnelson> ... then let you setup again[02:26:55] <ryancnelson> *-*-*-*-*-* insert data loss IRC warning here *-*-*-*-*-*-[02:27:09] <xmerlin> :)[02:28:03] <konobi> i thought that option was removed[02:28:30] <xmerlin> fyi the password in grub start with $5$.... as pointed[02:28:59] <ryancnelson> so, that should be ok[02:29:59] <xmerlin> yes it's ok ..destroy_zpool is fine ;)[02:53:29] *** Livid has quit IRC[02:53:55] *** Vod has joined #smartos[03:02:39] *** ryancnelson has quit IRC[03:04:24] *** ryancnelson has joined #smartos[03:05:35] *** cjones__ has quit IRC[03:05:48] *** cjones_ has joined #smartos[03:15:41] *** Vod has quit IRC[03:17:43] *** miine_ has joined #smartos[03:19:06] *** miine has quit IRC[03:19:06] *** miine_ is now known as miine[03:54:43] <xmerlin> is it safe to clone the disk of a vm to another one using dd ? .../dev/rdsk/ .... ?[03:54:59] <xmerlin> I plan to stop the vm before[03:56:16] <ryancnelson> you can do that, or, just zfs clone the "origin" disk[03:57:06] <xmerlin> ok ...because cloning with bs=4M gives me "write: File too large"[03:57:44] <ryancnelson> what you just said makes no sense. "zfs clone" doesn't take a bs= argument[03:58:08] <xmerlin> dd[04:00:38] <nahamu> wouldn't it make more sense to attach the disk and then split the pool?[04:00:54] <nahamu> or create a second pool and the snapshot and zfs send all the bits?[04:01:09] <ryancnelson> he's talking about the "disk" of a vm[04:01:12] <ryancnelson> ... a zvol[04:01:19] <ryancnelson> not a real disk[04:01:58] <konobi> xmerlin: use zfs[04:02:00] <nahamu> oh[04:02:19] <nahamu> snapshot and clone, or if you don't want the relationship, send and receive.[04:02:41] <ryancnelson> that's the right answer. but you *can* dd from a zvol to another.[04:02:54] <ryancnelson> i suspect he's got the devices wrong, though.[04:03:22] <nahamu> /dev/zvol/rdsk/zones/<volume-name> or something like that, right?[04:03:32] *** darjeeling has quit IRC[04:03:32] <xmerlin> right[04:03:38] <ryancnelson> ... you typed ".../dev/rdsk/..." ... don't do that. tell us what you *actually* did, xmerlin[04:03:50] <xmerlin> /dev/rdsk...[04:03:55] <ryancnelson> that's wrong[04:04:03] <xmerlin> :([04:04:21] * nahamu doesn't have a GZ console at the moment...[04:04:29] <konobi> nahamu: there's also promote[04:04:35] <ryancnelson> ... /dev/zvol/rdsk/zones/<your volume name>[04:04:40] <nahamu> konobi: promote just reverses the relationship...[04:04:42] <xmerlin> right[04:04:54] <ryancnelson> ... /dev/rdsk is *real disks*[04:04:56] <xmerlin> I've read the right devices from vmadm get .....[04:05:04] <nahamu> if you want them to be completely unrelated you have to do send and receive[04:05:36] <xmerlin> dd if=/dev/zvol/rdsk/zones/99641ccb-5024-4698-a8d2-bd0f2d7e1c4c-disk0 of=/dev/zvol/rdsk/zones/c6e1c1eb-c329-462b-8a0b-8b3a68f58146-disk0[04:05:59] <nahamu> are the zvols identical in size?[04:06:07] <xmerlin> but if I add bs=4M it complains about "write: File too large"[04:06:12] <xmerlin> yes[04:06:14] <nahamu> oh, weird[04:06:26] <xmerlin> it works only without bs=4M[04:06:45] <ryancnelson> so, use a smaller block size[04:07:22] <nahamu> it's indeed a terrible error message if the actual problem is the block size...[04:07:51] <xmerlin> ryancnelson, yes of course ...but I'm tring to understand what's the problem[04:07:59] <nahamu> truss or dtrace![04:08:08] <ryancnelson> oh. ... *is /dev/zvol/rdsk/zones/c6e1c1eb-c329-462b-8a0b-8b3a68f58146-disk0 actually a zvol* ?[04:08:12] <ryancnelson> or did you make that up?[04:08:18] <ryancnelson> it's not a file.[04:08:24] <ryancnelson> you need to create the zvol.[04:08:36] <xmerlin> yes I've create a blank vm using almost the same payload[04:08:42] <xmerlin> yes[04:08:58] <wesolows> OS-1776?[04:09:10] <xmerlin> ?[04:09:10] <wesolows> oh, fuck us for not having a public bug database.[04:09:34] <nahamu> what's the one-line description of OS-1776?[04:09:35] *** jamesd has quit IRC[04:09:49] <wesolows> first write to a new zvol can fail with EFBIG[04:10:10] <nahamu> is upstream illumos unaffected?[04:10:22] <wesolows> I would assume it's affected as well but do not know.[04:10:38] <nahamu> https://www.illumos.org/issues/3425[04:10:59] <wesolows> well there we go; good job rm. :-)[04:11:04] <nahamu> looks like rmustacc was kind enough to file it there too[04:11:16] <ryancnelson> yeah... in the meantime, just write smaller blocks. if you're really dd'ing, 128k is probably the right size, huh[04:11:25] <wesolows> you should be able to tell whether this is the same problem from his analysis.[04:11:49] <ryancnelson> ... but really, zfs send/receive is the right way, and clone is even more clever/efficient[04:11:55] <xmerlin> wesolows, thank you[04:13:00] <wesolows> np... EFBIG + dd + bs= triggered a gigantic fucking L1 cache hit :-)[04:13:01] <xmerlin> ryancnelson, ook[04:13:31] <wesolows> that said, it may be different.[04:14:22] <ryancnelson> if you want to make a vm from a template, that's precisely what zfs clone is for, and it's instantaneous. It's good to know about that bug, though.[04:15:24] <ryancnelson> also, if you zfs clone a 100gb disk, and change just 2gb of it, you've only used 2gb extra disk for the 2nd copy[04:16:15] <wesolows> quiet, you! I own stock in STX.[04:16:24] <wesolows> (just kidding, I don't really)[04:17:37] <ryancnelson> the whole "zfs clone" +[04:18:01] <ryancnelson> sorry, "zfs clone" + "zfs rename" cycle is how i swap disks around when i'm building datasets[04:19:00] <ryancnelson> or, zfs send a zvol snapshot from one box to another, and then provision a like-sized machine on the target, then replace it's disk with zfs recv , zfs rename[04:19:01] <xmerlin> ryancnelson, is there also a gain in the larc/l2arc using clone?[04:19:43] <nahamu> wesolows: they're in the S&P 500... if you own an index fund you might indeed own some... :-P[04:19:59] <ryancnelson> i guess there would be... if you had 10 windows machines all reading that screensaver with the pipes on it... every time the kernel read one of those blocks in, it'd be cached for the others[04:20:03] <nahamu> I certainly do... everyone go buy disks! :-P[04:20:15] <wesolows> yes. fortunately I have disposed of all such funds and no longer own any. precisely because they tend to hold a lot of garbage I don't like or understand. :-)[04:20:30] <ryancnelson> ... assuming you were using clones of your windows image, and nobody "defragged" their disks[04:21:36] <nahamu> defrag before cloning?[04:22:03] <wesolows> guess that would beat defrag after cloning :-)[04:22:19] <wesolows> 2x the I/O, and wasted space.[04:23:38] <ryancnelson> we did make our linux datasets *WAAAY* more compressable by zeroing out swap (why is that even in there?) and empty space[04:24:57] <wesolows> swap are the divining sticks Linux users employ to keep the oom killer at bay.[04:25:43] <wesolows> it works about as well.[04:38:01] *** sachinsharma has joined #smartos[04:51:31] *** ryancnelson has quit IRC[04:57:28] *** newlix has joined #smartos[05:01:53] *** newlix has quit IRC[05:21:55] *** daleg has quit IRC[06:57:44] *** newlix has joined #smartos[07:01:55] *** newlix has quit IRC[07:05:40] *** codingstream has quit IRC[07:06:57] *** wolfeida_ has quit IRC[07:24:20] *** miine has quit IRC[07:32:17] *** forexample_ has quit IRC[08:09:58] *** cjones_ has quit IRC[08:52:39] *** dysinger has quit IRC[09:02:24] *** potatosalad has quit IRC[09:06:10] *** alucardX has joined #smartos[09:06:39] <alucardX> morning[09:11:56] *** kschiess has joined #smartos[09:13:45] <Licenser> morning alucardX[09:23:29] *** texarcana has quit IRC[09:24:42] *** texarcana has joined #smartos[09:58:31] *** cjones_ has joined #smartos[09:59:56] *** cjones_ has quit IRC[09:59:56] *** cjones__ has joined #smartos[10:18:50] *** kschiess has quit IRC[10:21:53] *** kschiess has joined #smartos[10:24:18] <lundh> how many cores does a zone have available per default?[10:30:35] <konobi> all[10:31:11] *** KermitTheFragger has joined #smartos[10:40:18] <lundh> good[10:41:10] <lundh> how do I check resource usage in a zone and in the gz?[10:41:29] <lundh> memory, cpu etc[10:43:44] <konobi> prstat[10:44:33] <lundh> look at that :)[10:46:16] <lundh> so last question for now: my cron jobs (in a zone) uses a different environment then my normal shell environment. how can I make sure that at least charset matches?[10:51:36] *** bayoda has quit IRC[10:53:06] *** bayoda has joined #smartos[10:56:39] *** bayoda has left #smartos[11:01:55] *** bayoda has joined #smartos[11:04:01] *** newlix_ has quit IRC[11:09:07] *** bens1 has joined #smartos[11:53:59] *** Cpt-Oblivious has joined #smartos[12:36:38] *** linuxprofessor has quit IRC[12:58:36] *** newlix has joined #smartos[13:00:45] *** sachinsharma has quit IRC[13:02:56] *** newlix has quit IRC[13:04:18] *** bens1 has quit IRC[13:37:01] *** bens1 has joined #smartos[13:38:29] *** linuxprofessor has joined #smartos[13:41:57] *** ira has joined #smartos[14:03:50] *** avrntsv has joined #smartos[14:16:36] *** avrntsv has quit IRC[14:58:55] *** newlix has joined #smartos[15:02:13] *** des2 has quit IRC[15:03:12] *** newlix has quit IRC[15:09:23] *** des2 has joined #smartos[15:14:07] *** bens1 has quit IRC[15:14:11] *** avrntsv has joined #smartos[15:36:51] *** alucardX has quit IRC[15:37:06] *** neophenix has joined #smartos[15:38:32] *** jamesd has joined #smartos[16:02:27] *** cjones__ has quit IRC[16:02:41] *** cjones_ has joined #smartos[16:26:04] *** daleg has joined #smartos[16:26:49] *** Cpt-Oblivious has quit IRC[16:27:04] *** Cpt-Oblivious has joined #smartos[16:29:06] *** kschiess has quit IRC[16:43:52] *** blu has quit IRC[16:44:14] *** blu has joined #smartos[16:46:42] *** bayoda has quit IRC[16:48:59] *** potatosalad has joined #smartos[16:51:12] *** bayoda has joined #smartos[16:59:12] *** newlix has joined #smartos[17:03:22] *** newlix has quit IRC[17:42:36] <kfr-> I was watching puppetconf2012 videos[17:43:06] <kfr-> they were really interesting since they were all, this is who we are, this is what we do with puppet[17:43:33] <kfr-> It was video after video of that[18:00:01] *** kfr-_ has joined #smartos[18:00:28] *** kfr- has quit IRC[18:05:31] <killfill> im using JDC, and my compute node just went down. i reboot it. some VM's wont start, vnc shows a black screen. i vmadm halt them manually, and now when i login to the portal the machines that are running shows as 'offline'[18:06:01] <killfill> how does the head node knows that VM's are online?..[18:09:04] *** rcamera has joined #smartos[18:09:25] *** bens1 has joined #smartos[18:32:29] *** dysinger has joined #smartos[18:43:09] *** tonyarkles has joined #smartos[19:07:58] <LeftWing> killfill: Do you mean SDC?[19:08:51] <LeftWing> lundh: If your scripts depend on particular environment settings, then you ought to set them in the script.[19:09:15] <LeftWing> e.g. if you need LANG=C, then set that in the script.[19:09:20] <lundh> ok :)[19:10:22] <lundh> just found out why lofs might not be the best of ideas :)[19:10:42] <lundh> the zone have no idea how much space is left on the partition[19:11:29] <ira> lundh: Actually, they do… they just don't know how to ask. ;) If you use statvfs it should work.[19:12:15] <ira> (That's the system call… now how to turn that into something useful, I'll leave as an exercise to the reader alas.)[19:13:12] *** user01 has joined #smartos[19:13:28] <lundh> :)[19:13:47] <LeftWing> So if your lofs is mounted at, say, /testing[19:13:57] <LeftWing> What happens if you "df -h /testing"[19:14:21] <LeftWing> (from inside the zone, that is)[19:14:30] <user01> This may not be the proper place to ask, but are there virtio drivers for OmniOS/illumos as a guest using SmartOS KVM host?[19:14:37] <lundh> since I have file systems below testing I only get 168k used[19:14:53] <LeftWing> user01: VirtIO block (disk) drivers were recently integrated into illumos-gate[19:15:06] <LeftWing> I'm not sure what happened to the network drivers, but I don't believe they've been integrated yet.[19:15:46] <LeftWing> lundh: Ah, you have nested filesystems under the lofs mount?[19:16:01] <lundh> yeah[19:16:22] <LeftWing> I guess that could be a problem. If you df -h /testing/some/nested/filesystem ?[19:16:34] <LeftWing> Also, is the "Avail" figure correct, at least?[19:16:52] <lundh> no[19:17:06] <lundh> avail will only tell me the total size of the root volume[19:17:18] <ira> lundh: Are you using samba?[19:17:21] <lundh> no[19:17:23] <lundh> lofs[19:17:26] <user01> LeftWing: Okay, cool. My bigger problem is to get proprietary software at work to not core dump in smartos...[19:17:41] <ira> Ok… Samba + lofs has this behavior… and I have some patches in my queue to deal with it ;)[19:18:11] <lundh> I use netatalk though and netatalk seems to report the correct amount of free space but wrong total[19:19:19] <ira> lundh: That makes sense… Because each volume reports its own "used" not the total of everything under it. So, you're getting the report of the "base" of your lofs tree.[19:21:03] <ira> LeftWing: I've got a tricker one, that I'm considering patching. What do you do with lofs mounted snapshot trigger directories? (IE: A situation where you lofs mount a volume that has sub volumes…)[19:21:30] <ira> The triggers may not have been hit… and the zone doesn't have permissions to mount them today.[19:21:51] <LeftWing> I am only distantly aware of snapshot trigger directories.[19:22:02] <lundh> you just think of these things to confuse us noobs :)[19:22:08] <ira> Be thankful ;)[19:22:39] <ira> lundh: Alas no… you'll want them too. .zfs/snapshot is a useful directory ;)[19:24:08] <lundh> :)[19:25:33] *** des2 has quit IRC[19:28:20] *** des2 has joined #smartos[19:29:41] *** KermitTheFragger has quit IRC[19:32:08] *** user01 has quit IRC[20:06:03] <konobi> mornin[20:15:14] <nahamu> hi konobi[20:16:50] *** rcamera has quit IRC[20:19:16] *** serzhit has joined #smartos[20:21:31] <serzhit> Is there anybody in here?[20:22:02] <wesolows> yes[20:22:43] <serzhit> Could you help? I cant load SmartOS in 'no install' mode password is not accepted[20:22:51] <serzhit> latest version[20:23:04] <wesolows> hmm, someone said they had this problem yesterday.[20:23:09] <serzhit> password is from file[20:23:48] <nahamu> did you try "root" as the password too?[20:23:52] <serzhit> no[20:24:02] <nahamu> try that.[20:24:26] <nahamu> (does anyone know if the hash in grub is now being changed every release?)[20:24:49] <serzhit> ok. I'll disappear for a moment )[20:25:03] *** serzhit has left #smartos[20:25:53] <wesolows> what hash?[20:26:18] <wesolows> I believe the password hash is different in each build, yes.[20:26:26] *** bens1 has quit IRC[20:26:34] <wesolows> The boot archive hash is different for every build also, because it just is.[20:26:47] <nahamu> not the one in shadow, the one embedded in the grub menu.lst file that gets used in certain cases.[20:27:37] <nahamu> I should ask e^ipi how he does releases these days.[20:29:18] <e^ipi> run a script.[20:30:03] <nahamu> do you use tools/build_iso and tools/build_usb?[20:30:09] <e^ipi> the noinstall password should be 'root' , since one of the init services whose name i forget sets it based on the hash passed via the kernel boot line[20:30:39] <e^ipi> nahamu: no, build_usb doesn't work. i use build_iso & then download a template usb image (with grub installed)[20:31:50] *** wolstena has joined #smartos[20:32:02] <nahamu> e^ipi: sorry that build_usb doesn't work for you... some day I should get that fixed for non-root users...[20:32:39] <e^ipi> builds are done in a zone as root. i don't think you *could* get it working for non-root, since you have to fiddle with lofi[20:33:43] <nahamu> hmmmm... then I'm genuinely surprised that it doesn't work... I'll have to take a look at that when I get a chance.[20:34:40] <nahamu> still, as long as you're not changing the hash in grub from release to release, then my understanding was still correct.[20:34:51] <e^ipi> yes[20:40:26] <nahamu> does the usb template have a different menu.lst that includes a "no install" boot option?[20:43:30] <e^ipi> shouldn't, it's the same one that's been used forever[20:48:02] <nahamu> so that person asking probably read some bit of documentation and figured out how to modify the kernel command line. cool.[20:48:25] <ira> e^ipi: I use build_usb, I think.[20:48:49] <ira> Actually, I know I do… That's how I built my most recent production images. :)[20:48:50] <nahamu> ira: I think I originally wrote it for you, so I'm glad it at least works for you. :)[20:49:16] <nahamu> (well, I also thought it was an interesting challenge...)[20:49:20] <ira> nahamu: I appreciate it, I actually changed it a bit so we can image 4GB sticks. Otherwise works like a champ :)[20:52:51] *** wolfeidau has joined #smartos[20:54:36] <nahamu> does anyone know where the commits that introduced vrrp support to overlay/generic/lib/svc/method/net-physical ended up? I don't see them in either the smartos-live repo or the smartos-overlay one...[20:55:12] <nahamu> it's like when the file was moved from overlay to live it gained the vrrp features, but the commits where the work happened weren't in either...[20:55:54] <nahamu> I'm trying to forward port some changes joshie made, but the vrrp support really refactored the sections of that file...[21:00:01] *** newlix has joined #smartos[21:04:26] *** newlix has quit IRC[21:21:27] *** tonyarkles has quit IRC[21:24:40] *** potatosalad has quit IRC[21:33:12] *** potatosalad has joined #smartos[22:03:38] *** wramthun3 has joined #smartos[22:05:17] *** wramthun2 has quit IRC[22:08:54] *** avrntsv has quit IRC[22:09:40] *** molybdenum has joined #smartos[22:09:54] <nahamu> anyone know how to change the string reported in "uname -v" to use something other than "joyent"?[22:10:12] <nahamu> I want to make it easier to identify when I'm running a non-joyent build...[22:12:14] *** potatosalad has quit IRC[22:12:54] <nahamu> hmmm... perhaps the PUBLISHER in configure.smartos[22:13:13] <nahamu> though I think that has to be done before it generates the illumos.sh file for illumos.[22:16:11] <molybdenum> hi, should I expect this system, http://www.supermicro.com/products/motherboard/xeon3000/946/pdsbm-ln2.cfm , to have trouble booting smartos? 4GB RAM, Xeon 3210, no drives[22:17:20] <nahamu> are you having trouble, or are you considering buying it?[22:17:48] <molybdenum> having trouble[22:18:19] <molybdenum> i get a couple of lines beyond the grub menu and then nothing[22:18:34] <molybdenum> it complains about the ACPI SRAT[22:18:40] <wesolows> that warning is harmless[22:18:46] <wesolows> so you are seeing the SunOS banner?[22:18:57] <molybdenum> sunos 5.11 … joyent..[22:18:58] <molybdenum> yes[22:19:05] <molybdenum> then one more line beyond that and then hangs[22:19:35] <wesolows> are you using serial or legacy framebuffer?[22:19:55] <molybdenum> i've no idea, I selected 'Live 64-bit (text)'[22:20:05] <molybdenum> i'm guessing that implies framebuffer?[22:20:21] <molybdenum> i'm certainly only looking at the VGA display[22:20:23] <wesolows> ok[22:20:34] <konobi> wesolows: btw, that SRAT warning on the dual cpu box the other day was fixed by disabling 'Node Interleaving' in the bios[22:20:41] <molybdenum> should I be peering at the serial port outputs?[22:20:54] <wesolows> since we're booting, the next step is to take another lap and add -kv to the arguments[22:20:59] <molybdenum> hmm, there is an off-chance I've redirected the output, now I think about it[22:21:05] <wesolows> conveniently, this can be done by selecting the kdmb boot menu item[22:21:21] <wesolows> konobi: I saw, thanks. I've passed that on to someone else who was asking about the r720xd also.[22:21:55] <molybdenum> i can get the kmdb prompt, but was unsure what to do with it[22:22:18] <wesolows> oh, so you get that far? ok, type :c and press enter[22:22:19] <molybdenum> i need to double-check the serial console redirection in the BIOS at this point I think[22:22:31] <wesolows> no. if you're not using the serial console, then what you're doing is fine.[22:23:13] <wesolows> are you seeing any messages scroll by? if so, when they stop, what's the last one?[22:23:24] <molybdenum> no messages scrolling by[22:23:37] <molybdenum> just stopping one or two lines after the sunos 5.11 line[22:23:39] <wesolows> hmm, I thought we did -v...[22:23:41] <molybdenum> well, hanging[22:23:42] <wesolows> let me double check that[22:24:04] <molybdenum> i used the USB image from Dec. 13, I believe[22:24:29] <molybdenum> thanks for the pointers, i've got some diagnostic ideas now. need to head out for a bit[22:24:31] <wesolows> ah, we don't. It's -kd. You'll need to edit the GRUB menu to add -v.[22:24:36] <wesolows> ok[22:24:49] <wesolows> also, -m milestone=none, as it may be hanging in userland[22:24:53] <wesolows> good luck![22:25:24] *** bens1 has joined #smartos[22:26:14] * konobi divines an issue with kbd8042 and mouse8042[22:27:12] <wesolows> this is why we need a machine architecture that only has serial.[22:27:14] *** ryancnelson has joined #smartos[22:27:28] <wesolows> VGA = bios, 8042 = crap, and most machines today are even worse because they only have USB[22:29:39] <konobi> yup... it's just come up a few times... though funny thing is that OI works in those cases[22:32:41] <konobi> would be good to get some kvm access to such boxen just to hammer through[22:33:45] <wesolows> well we don't have any delta there relative to illumos[22:34:05] <nahamu> konobi: you should offer people that if they send you a box you'll spend up to n hours trying to fix the issue and you get to keep the machine...[22:34:24] <konobi> ha[22:34:38] <wesolows> the biggest problem is that we don't have systems that fail this way, in part because they tend to be the consumer-grade systems that are far too numerous to test and not our market anyway[22:34:41] <konobi> unfortunately my kmdb isn't super[22:35:20] <wesolows> this one also happens to be extremely old, though there are sometimes problems with brand new systems too[22:35:49] <wesolows> fortunately within a few years the PC will be gone and we won't have to worry about these problems any more[22:36:26] <jamesd> wesolows, they have been saying the same thing about the mainframe for the last 20 years.... there are still mainframes alive and kicking.[22:37:02] <wesolows> by "gone" I don't mean that every single PC will turn to dust overnight[22:37:26] <wesolows> I mean that the number of them being sold will be very small, and the number of distinct makers and models much more manageable than it is today[22:37:45] <wesolows> just like how there used to be dozens of mainframe makers; now there is... one?[22:38:35] <nahamu> wesolows: I'm doing my part... with spice for smartos you don't need a desktop... just a thin client and a smartos server...[22:38:46] <wesolows> yep[22:39:24] <konobi> nahamu: get yerself a raspi yet?[22:39:33] <nahamu> it's either that, or the chromebooks and and ios and android devices will take over the world...[22:39:56] <nahamu> konobi: nope. haven't had the time. but people in the spice IRC channel were talking about it. supposedly the arm stuff is there in the repos.[22:40:05] <wesolows> nahamu: I suspect and fear the latter, if only because the thing I don't want is always what happens. Yes, always.[22:41:02] <konobi> nahamu: yeah, the openelec distro might make a good starting point[22:41:06] <nahamu> well, if you ever want to help me hack on the latest QEMU to help fight the good fight... let me know.[22:42:55] <ryancnelson> i got raspbian booting under qemu's arm emulation on my mac on christmas... fun![22:45:56] <nahamu> between that, and the guy who's trying to get MacOS to run under QEMU/KVM... all we'll need is someone smart enough to hack up spice drivers for MacOS who knows what craziness you could achieve...[22:46:25] <nahamu> wow... that wasn't quite English.[22:47:39] <nahamu> ...and my smartos build failed trying to build binutils after it already finished building illumos... wtf?[22:48:15] *** bens1 has quit IRC[22:51:11] <wesolows> building -extra is a 2-phase process[22:51:23] <nahamu> http://fpaste.org/nKWh/[22:51:36] <wesolows> first we build a bootstrap subset needed to build illumos, then we build illumos, then we build -extra against the proto area.[22:51:51] <nahamu> wesolows: makes sense.[22:52:12] <wesolows> your system's linker is too old[22:52:14] <nahamu> oh, I bet it's something wrong with my build zone... it's probably too old.[22:52:18] <wesolows> upgrade your platform[22:52:30] <nahamu> it's the right platform, but the zone was created on an older platform.[22:52:37] <nahamu> gotta recreate the zone from scratch.[22:52:39] <wesolows> oh, right... fucking fake-subset[22:52:43] <nahamu> yeah...[22:52:53] <nahamu> I should have known being lazy would bite me in the ass.[22:52:55] <wesolows> although we really ought to be using the linker from the proto area there... hmm.[22:53:27] <wesolows> but gcc has the path to the linker hard coded, and it may not be possible to override it. Blecch.[22:53:37] <wesolows> These GNU people really have no idea how to build a system from scratch.[22:53:50] <konobi> yeah, just look at hurd[22:54:41] <wesolows> indeed. in the 25 or so years since the hurd was first mentioned, we've seen dozens of new OSs, but the hurd wasn't one of them.[22:54:59] <nahamu> how do I adjust the size of the filesystem for a zone in the json?[22:55:30] <wesolows> quota[22:55:41] <nahamu> thank you... searching for all the wrong terms in the man page...[22:57:38] <nahamu> heh, it's in the wiki page for building... I probably put it there too...[22:59:54] *** kimc has joined #smartos[22:59:54] *** kimc has joined #smartos[23:06:38] *** wolfeidau has quit IRC[23:13:27] <molybdenum> hmm, -v and -m milestone=none didn't really change the outcome much. while I do see a few dozen lines of kernel devices being discovered, hangs after the ehci3… line[23:13:36] <nahamu> USB3![23:13:41] <nahamu> turn that off in the BIOS[23:13:48] <molybdenum> er, sorry, uhci3[23:13:58] <nahamu> yeah, that's USB3. turn it off.[23:14:14] <molybdenum> ok[23:14:59] <konobi> usb3 == xhci iirc[23:15:06] <nahamu> odd... I don't see USB3 specified on the motherboard specs.[23:15:13] <nahamu> oh, then I'm wrong.[23:15:52] <wesolows> yeah this isn't that. it's just the third uhci instance[23:15:56] <wesolows> (fourth)[23:16:48] <wesolows> that said, it's whatever comes after uhci3 that's hanging[23:17:57] <wesolows> does this system have a way to send an NMI? if so, that's probably the easiest way to figure out where you are.[23:20:37] <wesolows> if not, my next step would be to set a breakpoint in uhci_attach, and then probably in the module loading code when that fires[23:21:08] *** wolstena has quit IRC[23:21:16] <wesolows> this kind of stuff is a pain in the ass to debug unless you happen to get lucky, or can guess the problem[23:23:13] <konobi> molybdenum: looks like that board should be nmi-able[23:24:29] *** wolfeidau has joined #smartos[23:39:10] *** iyp has joined #smartos[23:39:27] <konobi> though now that i think of it, there wouldn't be anywhere for the dump to go ^_^[23:41:52] *** wolstena has joined #smartos[23:43:33] <nahamu> no disks, no dump...[23:44:16] *** Vod has joined #smartos[23:46:41] <wesolows> you're in kmdb[23:46:48] <wesolows> it'll drop you there[23:47:17] <konobi> molybdenum: ping?[23:47:38] *** potatosalad has joined #smartos[23:52:58] *** dysinger has quit IRC[23:53:41] *** dysinger has joined #smartos[23:53:55] *** iyp has quit IRC[23:56:21] *** dysinger has quit IRC