July 7, 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 | >
July 7, 2011 [00:00:11] *** tsoome1 has joined #openindiana[00:00:17] *** axisys has quit IRC[00:03:01] *** tsoome has quit IRC[00:03:01] *** tsoome1 is now known as tsoome[00:03:02] <sergefonville1> what is wrong with:[00:03:05] <sergefonville1> owner@:read_data/write_data/read_attributes:file_inherit:allow[00:04:29] *** Maylow has joined #openindiana[00:05:16] <Maylow> good evening[00:06:09] *** hsp has quit IRC[00:08:28] *** skeeziks has quit IRC[00:13:08] *** Maylow has quit IRC[00:14:19] *** CoilDomain has quit IRC[00:20:07] *** madwizar1 has quit IRC[00:20:12] *** madwizard has joined #openindiana[00:22:20] *** zynox has quit IRC[00:23:49] *** DanaG has joined #openindiana[00:24:02] <DanaG> okay, this openindiana installer has been sitting at 99% done for the past 3 hours.[00:24:37] <DanaG> How do I get the **** thing to finish?[00:24:55] <longcat> does iostat -xtc 1 show any i/o?[00:24:58] <longcat> on the disk[00:25:09] <DanaG> I don't know -- it's the text installer, and I can't vt-switch.[00:26:25] <randomuser> "BUG: Bad page state in process swapper"[00:26:47] <randomuser> whoops, wrong channel[00:29:11] <DanaG> So now I've taken the other flash drive, that was taking 45 minutes to fail to boot, and it works fine with ubuntu on it.[00:29:34] <DanaG> So all I can do now, with the installer, is quit.[00:29:49] <sergefonville1> it is possible to have newly created files be the same as the owner/group from the dir in wich they are created right?[00:30:35] <DanaG> Okay, not even F9 (to quit) is working.[00:32:04] *** gea has quit IRC[00:32:30] *** jkimball4 has quit IRC[00:33:55] *** sergefonville1 has left #openindiana[00:33:57] *** jkimball4 has joined #openindiana[00:34:30] *** freetown2 has quit IRC[00:34:34] *** sergefonville has joined #openindiana[00:34:56] *** JStoker has quit IRC[00:35:33] *** mnaser has joined #openindiana[00:36:53] <DanaG> How long is it supposed to take to install?[00:36:57] <DanaG> Certainly not hours.[00:37:26] <sergefonville> what does the + mean at the end of te pemissions when I ls?[00:39:02] <alanc> acls[00:39:17] <sergefonville> a + means there are ACLs?[00:40:49] <alanc> yes[00:41:03] <alanc> this character is a plus sign (+) character if a[00:41:03] <alanc> non-trivial ACL is associated with the file or a space char-[00:41:03] <alanc> acter if not.[00:41:07] <alanc> from the ls man page[00:41:28] <richlowe> "non-trivial ACL" being a technical term.[00:41:34] <richlowe> also, a frank admission[00:42:11] <sergefonville> what man page?[00:42:18] <sergefonville> since I tried, chmod, acl and ls[00:42:24] <richlowe> man /usr/bin/ls[00:42:28] <richlowe> you'd probably have got the GNU page[00:43:24] <alanc> oh yes, I have /usr/man in my $MANPATH and not /usr/gnu/man[00:45:41] <sergefonville> I read that page, but cannot find the 'his character is a plus sign...' part[00:46:24] <sergefonville> found :P[00:46:30] *** JStoker has joined #openindiana[00:46:40] <sergefonville> back to my oritinal problem[00:47:03] <sergefonville> making sure all newly created files and dirs have the same owner/group and relevant acls[00:58:10] *** axisys has joined #openindiana[00:59:00] <sergefonville> is it correct that I can only inherit permissions and not ownership?[00:59:31] <sergefonville> or does that mean I have to define acls based on the user instead of the owner@, group@, etc?[01:01:47] <alanc> inherting group is easy[01:02:19] <alanc> I don't know if you can inherit ownership, since that would normally be a security hole admins would want to kill over[01:06:29] <sergefonville> it is a samba option[01:06:39] <sergefonville> force group and force user[01:07:57] *** miine has joined #openindiana[01:10:41] <miine> Hmm. just thinking about ZFS on RAID 10. so the "data layout" should be A&B + A&B. So one A&B (two disks) could fail without problems. But what happens if two "A" or two "B" fail at the same time (result would be either B + B or A + A).[01:11:08] <miine> Does ZFS magically shift some bits in one mirror so that the data is still safe and could be resilvered?[01:12:01] <miine> or I'm totally wrong on ZFS and RAID 10 ?[01:12:21] <sergefonville> it kinda worked[01:12:49] <sergefonville> I know have all newly created files have the right permissions for the relevant user and group[01:13:07] <sergefonville> but, it is owned by root.root with 000[01:13:40] <sergefonville> so I think that with acls ownershipt is not really relevant anymore[01:13:57] <RoyK> miine: zfs is about as magic as fat32[01:14:04] <RoyK> it just works a bit better[01:14:44] <OniAtWork2> miine, on a ZFS "raid 10" if you lose both disks from a vdev, you loose the array (for all practical purposes)[01:14:59] <sickness> raid 10 is risky :)[01:15:10] <RoyK> but fast indeed[01:15:33] <sickness> yeah[01:15:42] <RoyK> I have a box with eight mirrors and some SSDs for L2ARC and SLOG[01:15:50] <OniAtWork2> 3-way mirrors keep you redundent even if you lose a second one before the first can be repaired.[01:15:50] <RoyK> it's very fast indeed[01:16:12] <OniAtWork2> or find bad sectors while rebuilding.[01:16:19] <RoyK> and more iops[01:16:32] <RoyK> since zfs reads from all sides of a mirror[01:16:35] <sergefonville> raidz2 with 4 disks is the same as raid10, without the risk of one set crashing[01:17:02] <RoyK> but worse in terms of iops[01:17:19] <OniAtWork2> It's not really, it has different performance chactoristics.[01:17:22] <RoyK> it all depends what you want or need[01:17:50] *** tsoome1 has joined #openindiana[01:18:00] *** tsoome has quit IRC[01:18:04] *** tsoome1 is now known as tsoome[01:18:06] <RoyK> I read a post on the opensolaris list from a guy that had setup a 35-drive raidz1[01:18:14] <RoyK> it crashed :P[01:18:28] <OniAtWork2> I can't say I'm supprised[01:18:36] <sickness> raid isn't a substitute for a backup anyway...[01:18:38] <miine> ok. maybe I'm wrong on that, but I just thought that by interleaving the bits different in one mirror it should be possible to be able to loose any combination of two disks in a 4 disk raid 10 ... (have only 4 disk tray available). but maybe that interleaving idea doesn't work anyway...[01:19:09] <sickness> do a 4 disk raidz2 and you can surely loose any 2 disks :P[01:19:09] *** mnaser has quit IRC[01:19:29] *** tsoome1 has joined #openindiana[01:19:33] <OniAtWork2> miine: it's kinda possible, but you are talking about "raid-6" or "raid-z2" at that point.[01:19:54] <RoyK> miine: if you want good iops, use striped mirrors, if you want space and safety, use raidz2[01:20:04] <OniAtWork2> how you are useing the disk is different, and performs vary differently[01:21:18] <miine> ok. was a bad idea. won't work even in theory. will stay with RAID-Z1 with either 3 or 4 disks...[01:22:38] *** tsoome has quit IRC[01:22:39] *** tsoome1 is now known as tsoome[01:23:02] <OniAtWork2> if you came to me and asked for a reccomendation for a machine (offsite, of course) to dump backups to, I wouldn't even have to think before saying "raid-z(2)." For other usages the loss in performance can be bad, so "It depends"[01:23:53] * OniAtWork2 has never been accoused of being to quick to commit[01:24:07] <miine> OniAtWork2: yes. the more I think about it, RAID-Z2 makes more sense...[01:24:11] <RoyK> sickness: raid 10 isn't really risky - resilver times are low and it's fast as fuck[01:24:33] <sergefonville> only actual data is resilvered[01:25:01] <OniAtWork2> performance during a resilver on raid-z2 is just abysmal. on the raid-10 it's not so bad.[01:25:21] <RoyK> sergefonville: and metadata[01:25:40] <sergefonville> ur right ofcourse[01:25:50] <OniAtWork2> sergefonville, that's true, I should say for a larger data-sets[01:25:54] <sergefonville> but on a 500G disk with 35 used...[01:26:01] <RoyK> OniAtWork2: performance during resilver on raid-z(something) is quite good if you have SSDs for slog/l2arc[01:26:07] <miine> Just thinking if it makes sense to use 2 x 2 2 TB drives from different vendors in a 4 disk RAID-Z 2...[01:27:19] <OniAtWork2> RoyK, depends on your access patterns. I have battery backed DRAM for Slog, and (more conventional) SSDs for l2arc, and it was a problem[01:28:26] <RoyK> OniAtWork2: I only have vertex 3 for slog and some c300s for l2arc - it's very fast indeed[01:28:48] <RoyK> random access included[01:30:29] * OniAtWork2 shrugs. we had issues, running l2arc and slogs with raid-10 worked much better.[01:30:32] *** DeanoC_ has joined #openindiana[01:30:33] *** ChanServ sets mode: +o DeanoC_[01:32:22] *** DeanoC has quit IRC[01:32:34] *** axisys has quit IRC[01:38:13] <miine> another question: is somebody using puppet and zones here and has already a solution (willing to share) for changing existing zones (e.g. add/remove file systems etc.)?[01:40:17] <DontKnwMuch> how can I check to see if OI detected my 4 cores or not?[01:40:52] <Triskelios> psrinfo[01:41:03] <Triskelios> (-v)[01:41:31] <DontKnwMuch> ah.. ok :)[01:41:51] *** DanaG has quit IRC[01:42:01] *** LizardKing has quit IRC[01:42:52] *** ottom has joined #openindiana[01:45:59] *** tsoome1 has joined #openindiana[01:48:00] <sergefonville> is there a remote gui tool to manage ACLs?[01:48:22] <sergefonville> or an X app to do the same thing locally[01:48:28] <sergefonville> I could forward that :P[01:49:41] *** tsoome has quit IRC[01:49:41] *** tsoome1 is now known as tsoome[01:50:53] <sergefonville> I just noticed its 2AM[01:51:03] <sergefonville> I had to go to sleep 3 hours ago :P[01:51:28] <sergefonville> thanks a lot for the great help again[01:51:35] <tsoome> windows can do it i believe[01:51:38] <tsoome> :P[01:52:20] <sergefonville> the checkmark for propate is greyed out[01:52:29] <sergefonville> and the users are unknown[01:52:38] <sergefonville> so, unfortunately no :([01:52:48] <sergefonville> anyway, I'm off[01:52:59] <sergefonville> bye ;-)[01:53:01] *** LizardKing has joined #openindiana[01:53:12] *** sergefonville has left #openindiana[02:03:19] *** randomuser has quit IRC[02:06:06] *** dws60451 has quit IRC[02:10:10] <miine> are there any plans to use mount_smbfs on gnome desktop instead of the f*****-up builtin gnome stuff?[02:16:54] *** axisys has joined #openindiana[02:27:43] *** InTheWings has quit IRC[02:33:27] *** DanaG has joined #openindiana[02:34:04] *** mnaser has joined #openindiana[03:05:04] *** DanaG has quit IRC[03:06:30] *** mnaser has quit IRC[03:11:13] *** master_of_master has quit IRC[03:12:34] *** master_of_master has joined #openindiana[03:18:55] *** birvin has quit IRC[03:29:13] *** land0 has joined #openindiana[03:42:21] *** mnaser has joined #openindiana[04:03:15] *** redgone has joined #openindiana[04:03:44] *** Reepicheep has joined #openindiana[04:11:05] *** axisys has quit IRC[04:12:17] *** DanaG has joined #openindiana[04:20:47] *** POloser has joined #openindiana[04:29:02] *** Hedonista has quit IRC[04:30:24] *** Hedonista_ has joined #openindiana[04:30:58] *** Hedonista_ is now known as Hedonista[04:35:19] *** edogawaconan has quit IRC[04:36:48] *** edogawaconan has joined #openindiana[04:43:06] *** SH0x has quit IRC[04:43:52] *** SH0x has joined #openindiana[04:51:09] *** Reepicheep has quit IRC[04:51:35] *** cruisereg has quit IRC[04:54:30] *** Naresh``` has joined #openindiana[04:59:58] *** kart_ has joined #openindiana[05:00:07] *** DrLou has quit IRC[05:00:14] *** mnaser has quit IRC[05:03:37] *** Reepicheep has joined #openindiana[05:09:11] *** akamit has joined #openindiana[05:11:17] *** land0` has joined #openindiana[05:14:24] *** land0 has quit IRC[05:14:46] *** land0` is now known as land0[05:21:49] *** jpg has joined #openindiana[05:21:49] *** jpg has joined #openindiana[05:22:55] *** DanaG has quit IRC[05:23:16] *** DanaG has joined #openindiana[05:23:21] *** akamit has quit IRC[05:27:23] *** dijenerate has joined #openindiana[05:37:37] *** Naresh``` has quit IRC[05:49:58] *** DanaG has quit IRC[05:55:00] *** miine has quit IRC[05:56:26] *** Nitial has quit IRC[05:57:34] *** Nitial has joined #openindiana[05:58:37] *** redgone has quit IRC[05:58:40] *** TPickle has quit IRC[06:02:44] *** akamit has joined #openindiana[06:05:55] *** DanaG has joined #openindiana[06:06:09] <DanaG> Hmm, installed successfully again, but again it's booting godawful slowly.[06:06:18] <DanaG> With service start timeouts.[06:07:04] <richlowe> As much detail as possible in a bug report would be really useful.[06:07:08] <DanaG> Notably, this is the second same-model flash drive, and Ubuntu works pretty well on the other one.[06:07:22] <DanaG> It took like 3 hours to install...[06:07:44] <richlowe> I've heard complaints of this nature a whole bunch, from "took forever" to "so much timed out it didn't work" to "appeared to hang at boot"[06:07:49] <DanaG> And now it's sitting there, saying nothing, when trying to boot.[06:07:52] <richlowe> but I've never seen it (I don't think anyone who'd normally try to fix it has)[06:08:08] <DanaG> Wait, I just got "configuring devices"...[06:08:19] <richlowe> so as much description of hardware, media, etc, as possible dumped into a bug would be awesome.[06:08:36] <DanaG> Ah. On bug tracker?[06:08:59] <richlowe> in general, if your hardware isn't _really_ old, and you're not doing anything insane regarding install media (over the network, in some way, I guess), you shouldn't be getting service timeouts at all[06:09:01] <richlowe> never mind the rest.[06:09:46] <richlowe> bugs.openindiana.org[06:12:12] <DanaG> It's already installed, on an HP Microserver.[06:12:43] <DanaG> I've heard somebody suggest disabling swap... How do I do that from recovery mode?[06:13:11] <richlowe> comment the line in /etc/vfstab and reboot[06:14:24] <DanaG> Is there any equivalent to the Linux parameter, "Single"?[06:17:11] <richlowe> if you're wanting to boot single user, add '-s' to the kernel$ line in in grub while booting (hit 'e' to edit, then 'e' to edit the kernel$ line)[06:20:14] <DanaG> Hmm, this time it booted, but it's rejecting all passwords.[06:20:24] <DanaG> Er, wait...[06:21:14] <DanaG> Yup, my username plus my password, rejected. Root plus my password, rejected.[06:21:26] <DanaG> I'll try -s[06:21:28] <richlowe> what's your network environment look like? name service and everything nice and tidy?[06:22:16] <DanaG> Just a home router running dnsmasq.[06:22:25] <DanaG> I'll try without Ethernet.[06:22:26] <richlowe> nothing funny like network auth?[06:22:38] <DanaG> Nope.[06:23:45] <DanaG> Just a fresh install, then that kexec-ish reboot (it certainly didn't do a normal reboot), then the service timeout. Power off, then cold bot and it's rejecting passwords.[06:24:09] <DanaG> Bot. Hah.[06:27:04] <richlowe> configured to use your dnsmasq instance for DNS, or straight upstream?[06:27:23] <richlowe> there's a bug I just saw referenced that will cause services to deadlock and timeout if nameservice requests are slow[06:27:36] <DanaG> Dnsmasq is my dns, yes. To block Charter dns-breakage.[06:27:39] <richlowe> the real nasty one is for user/group info, but it should suffer with hosts in certain cases.[06:29:59] *** mnaser has joined #openindiana[06:30:55] <DanaG> Still rejecting username.[06:34:49] *** akamit has quit IRC[06:35:19] *** dijenerate has quit IRC[06:37:22] <DanaG> Same with ssh.[06:37:28] <DanaG> Of course.[06:38:36] <richlowe> stupid question; You'd leave the install media around boot of it by mistake, right? :)[06:39:37] <DanaG> Nope.[06:40:17] <DanaG> I'm betting I can chroot, but I don't know what to bind mount.[06:45:27] *** descipher_ has joined #openindiana[06:45:36] <DanaG> Maybe nothing, if I'm just doing passwd.[06:46:08] <DanaG> And disabling swap.[06:46:34] *** dijenerate has joined #openindiana[06:47:19] *** descipher has quit IRC[06:47:23] *** descipher_ is now known as descipher[06:48:41] *** land0 has quit IRC[06:50:00] *** land0 has joined #openindiana[06:51:49] <DanaG> I don't even know how to mount the root.[06:56:21] *** land0` has joined #openindiana[07:01:19] *** land0 has quit IRC[07:03:51] <DanaG> Okay, I got some pool named "rpool" to mount, but all it has inside is boot and etc.[07:09:11] *** ChanServ sets mode: +o madwizard[07:12:34] <madwizard> Coffee[07:15:00] *** Naresh``` has joined #openindiana[07:16:55] *** Naresh``` is now known as Naresh[07:17:01] *** mnaser has quit IRC[07:17:26] *** Naresh has quit IRC[07:17:26] *** Naresh has joined #openindiana[07:18:50] *** land0` has quit IRC[07:19:35] *** staticnote has joined #openindiana[07:22:44] *** dijenerate has quit IRC[07:23:55] *** mnaser has joined #openindiana[07:35:03] *** dijenerate has joined #openindiana[07:35:38] <DanaG> Okay, I got root mounted.[07:35:53] <DanaG> Now how do I passwd within that root?[07:36:17] <DanaG> Chrooting in doesn't work. Permission denied on passwd.[07:39:30] *** mnaser has quit IRC[07:42:02] *** staticnote has quit IRC[07:42:39] *** Botanic has quit IRC[07:45:36] <DanaG> Any tips on that (passwd within offline, mounted root)?[07:48:14] *** SH0x has quit IRC[07:48:48] *** SH0x has joined #openindiana[07:49:15] *** akamit has joined #openindiana[07:50:39] <DanaG> Okay, it gets weirder: the user "dana" DOES NOT EXIST on the installed system![07:53:33] <DanaG> I edited it to take out swap, and now I get invalid argument after grub.[07:54:36] <DanaG> I'm honestly about to just give up on this.[07:55:58] *** staticnote has joined #openindiana[08:00:59] *** gea_ has joined #openindiana[08:16:28] *** gea_ has quit IRC[08:16:52] *** Botanic has joined #openindiana[08:28:18] *** staticnote has quit IRC[08:32:41] *** merzo has joined #openindiana[08:33:33] *** DanaG has quit IRC[08:37:03] *** staticnote has joined #openindiana[08:39:34] <madwizard> Coffee[08:40:01] *** tsoome has quit IRC[08:52:05] *** kytibe has quit IRC[08:53:50] *** kytibe has joined #openindiana[08:55:32] *** kinesis has joined #openindiana[08:57:21] *** xmikus01 has joined #openindiana[08:57:39] *** McBofh has quit IRC[09:00:38] <DontKnwMuch> Is anyone here using OI virtualized in ESXI with vmdirectpath for controllers and napp-it? How is you CPU usage?[09:06:14] <DontKnwMuch> How can I have boot backuped *zfs send?) to be able to recover fast, or is it better to just reinstall in such a case?[09:08:04] *** Nitial has quit IRC[09:11:40] *** |AbsyntH| has joined #openindiana[09:11:40] *** Nitial has joined #openindiana[09:13:04] *** McBofh has joined #openindiana[09:18:14] <DontKnwMuch> Can I use SSD for boot and slog or is it not a good idea[09:18:32] *** Botanic has quit IRC[09:21:46] *** bens1 has joined #openindiana[09:24:12] *** hajma has joined #openindiana[09:31:03] *** gber has joined #openindiana[09:34:41] *** McBofh has quit IRC[09:35:08] <DontKnwMuch> can I have the boot system backuped in such a way, that it can be recovered in one (khm...) simple step, like copying snapshot back or something[09:37:25] <DontKnwMuch> I was going with esxi, but it can only use 1 vcpu efficiently, there I had an easy sistem snapshot and can copy it there and back fast[09:40:45] *** Micr0mega has joined #openindiana[09:44:41] <blues> So, I've been doing some soul searching today... I'm curious, is anyone here running OI + napp-it / some other appliance level application in a production environment?[09:47:23] <sickness> I'm curious too about OI+napp-it =)[09:48:01] <DontKnwMuch> blues: I do[09:48:08] <DontKnwMuch> but I have problems[09:48:21] <DontKnwMuch> as I tried to run in in esxi[09:48:46] <blues> Why did you go with it over say, a dell equallogic entry-level san?[09:48:49] <DontKnwMuch> will have to run the backup system again, and reinstall it bare meta[09:48:56] <DontKnwMuch> metal[09:49:22] <DontKnwMuch> only because my boss is a cheap bastard ;)[09:49:53] <blues> How many hosts do you run?[09:50:01] <blues> ESXi hosts, i mean[09:50:43] <DontKnwMuch> I have 4 esxi machines[09:52:01] <blues> Obviously, if the money was no issue you'd go with a true hardware SAN... but how far off are you from being happy with OI ?[09:52:26] <blues> Do you feel like its viable in a small-biz environment?[09:54:08] <DontKnwMuch> yes it is, and it works great, I just made a mistake to install it on esxi (the only guest on that machine) as I thought I will be safe from hardware chipset/lan drivers changes, but have just found out that using more than 1 vcpu does not work so good..[09:56:24] <blues> I'm looking at a virtualization implementation for the small business i work for, and we just don't have the capital to do a hardware san. My options are to do an OI box, or forgo shared storage, do a single host w/ local storage, and then when we add a second host in the next year or two, hope i can afford the san.[09:57:58] <blues> the consensus in the vmware chat is that A) i should probably find another job and B) anything built myself is half-assed and not production worthy and C) I'd be better off just not doing the project until i can get funds to "do it right"[10:01:19] *** staticnote has quit IRC[10:02:31] <DontKnwMuch> hm. Not true. Go with a supermicro barebone, OI and you are done. Has it all (well almost). Something like that for hardware: http://www.supermicro.nl/products/chassis/3U/936/SC936A-R1200.cfm[10:03:16] <blues> Oh i've got everything spec'd.. at this point i'm ready to pull the trigger and get going on building and implementing things... but I've got alot of people telling me that its not a good idea.[10:03:17] <DontKnwMuch> how many users do you have[10:04:33] <DontKnwMuch> I was running freenas on a self built machine for three years, 50 users, iscsi for esxi the works, and it was ok. OI is way better if you ask me[10:04:44] <blues> around 25 users over 3 locations... Total storage needed for vms is < 4 TB, More storage needed for things like document imaging and stuff, but it can be "slow" storage. My first year IOPS need is somewher around 500[10:06:23] <blues> my long term plan is year 1: one host, homegrown san, second host as test environment year 2: second host in production, introduce HA and other goodies year 3: start looking at maxing out existing hardware and looking at client virtualization (virtual desktops)[10:08:38] <blues> obviously my iops needs go up considerably in year 3, but i'm hoping by that time things have progressed to the point where it is more feasible to get the storage infrastructure in place that i need.[10:09:03] <DontKnwMuch> nice plan.[10:09:27] <blues> you are literally the only person that seems to think so.[10:09:31] <blues> so *hug*[10:09:58] <DontKnwMuch> :) perhaps my nick is also an indicator[10:10:09] <blues> well that is also a consideration[10:10:12] <blues> note mine[10:10:57] <blues> thats my hesitation... i'm not the smartest guy in the room... and when guys who know more than me tell me i'm doing it wrong.. i want to stop and re-evaluate... and i have..but the bottom line is, if i'm going to do it, it has to be this way[10:12:34] <DontKnwMuch> I think that if you pick good hardware you will not be any worse than buying a true hardware san...[10:14:09] <blues> Well, like i look at something that is not affordble...but is the closest thign to affordable, a dell equallogic ps4000x... and that thing has 16 disks, small size config is around 6 TB, can handle 2 arrays, has dual gigabit nics for data, 1 for management[10:14:52] <blues> I can build that, using OI + napp it... using enterprise level disks for < 40% of the cost of the quote i've gotten[10:15:26] <DeanoC_> You can build very powerful SAN/NAS using OI and spend whatever your budget is[10:15:50] <blues> obviously i'm not getting hte support... but at that price point i can afford to keep some spare disks on hand, and hell, even have a second box on stand by[10:17:22] <DontKnwMuch> yes, that is true and quite doable[10:17:42] *** lkocman has joined #openindiana[10:17:44] <DeanoC_> I do it as all the time, for VM's i'd go for something like (off the top of my) 8 2TB drives, mirrors (so 8TB useable) with 3 SSDs (whatever makes fits the budget) and a main board and processor that works[10:17:49] *** InTheWings has joined #openindiana[10:18:34] <blues> I think people hear white box and assume i'm gonna use western digital green drives or something[10:18:35] <DontKnwMuch> so blues, you see you are not alone :)[10:18:58] <DeanoC_> even budget drives are okay, just up the redundency and have a few on hand[10:19:46] <DeanoC_> its often people think 8 expensive driver or 8 cheap whereas it should be 8 expensive or 12 cheap and still save money...[10:20:25] <lblume> Heh, the biggest cost when using cheap drives ends up to be the enclosure to put them[10:20:43] <DeanoC_> its sad some many people don't get RAID, as that was the point, not that cheap can replace expensive 1 for 1 but lots of cheap can replace expensive[10:20:52] <DeanoC_> sure that has to factoring in of course :)[10:20:56] *** Infin1ty|work has joined #openindiana[10:21:00] <DeanoC_> factored in*[10:21:01] <blues> I'd like to mix drive types ideally. for my iSCSI target for ESXi, take 8 10 or 15K sas drives, then take 8 or so cheaper higher capacity drives and pool those and use for NFS shares for bulk data storage[10:22:26] <DontKnwMuch> Strangely, I have seen that mostly people are not using iscsi but NFS for ESXi[10:22:37] <DeanoC_> front end SSD and back end SATA and let ZFS do it thing.[10:23:08] <blues> If i go NFS, i'm limited to 1gbps connection between host and pool[10:23:30] <blues> with iSCSI i can do multipath and get 1.5 ~ 1.7 real world across 2 links[10:25:53] *** kart_ has quit IRC[10:28:31] *** tsoome has joined #openindiana[10:36:34] <blues> DontKnwMuch, what are you doing for backup ?[10:37:06] <DontKnwMuch> I have another very cheap OI and do a send -receive thing[10:38:04] <blues> I need to get more familiar with OI / napp-it and ZFS in general.. any reading you recommend? I've got my home lab setup, but haven't really had a chance to get going yet[10:38:07] <DontKnwMuch> and a tape library for file backups[10:38:32] <blues> i tried renaming a damn nic last night going by the solaris how-to and couldn't...decided to drink a beer and go to bed instead[10:38:38] <Infin1ty|work> hmm, i'm considering moving to openindiana, going to build a new storage, so far been using opensolaris, i'm a bit afraid running it on production, is there any reason not to run it on production?[10:41:12] <tsoome> blues: with dladm?[10:41:23] <blues> yeah[10:41:53] *** nbprashanth has joined #openindiana[10:42:09] <tsoome> you probably have to have it all inactive before you can rename it[10:42:20] <tsoome> so, make sure your nwam and friends are disabled[10:43:24] <tsoome> Infin1ty|work: same reasons as with opensolaris really.[10:44:00] <blues> is there a book i can read on solaris that would be helpful for stuff like this?. i havent used it since college.[10:44:37] <tsoome> http://www.oracle.com/technetwork/documentation/solaris-11-192991.html[10:44:46] <tsoome> quite some books[10:45:14] <tsoome> dont be afraid of name solaris 11, basically all will apply to OI as well.[10:45:17] <Infin1ty|work> tsoome, but isn't opensolaris more stable? one of the reasons i'm afraid is "dev release"[10:45:33] <tsoome> opensolaris *is* dev release.[10:45:45] <tsoome> even solaris 11 express is dev release[10:45:57] <tsoome> or preview, if you like to call it that way[10:46:28] <tsoome> it all depends where you draw the line;)[10:46:31] <Infin1ty|work> tsoome, hard decision :)[10:46:55] <tsoome> b148 has been quite stable afaik, loads of people are using it in production[10:47:43] <Infin1ty|work> tsoome, ok :) i will try that, thanks.[10:47:46] <blues> just dont use realtek adapters or gigabyte motherboards[10:47:46] <Infin1ty|work> really thanks :))[10:47:54] <Infin1ty|work> not going to :)[10:49:48] <tsoome> bloody hell, i need to figure out how to build and vpn to access the vpn:([10:51:03] <tsoome> why they have to make life hard by forcing to use stupid shit like openvpn etc... :/[10:51:28] *** raichoo has joined #openindiana[10:52:54] <lblume> run it in a Linux in VBox?[10:54:33] <tsoome> well, the thing is the fully open access to our company network is provided by openvpn, which is nice ofc. but you need bloody openvpn client to access it, and ipad does not have it. so i cant use my sunray soft from it...[10:55:44] <lblume> ... Run VBox in a VNC you access on the SunRay you display on your iPad?[10:56:16] <Gugge> just jailbreak the ipad :)[10:56:34] <tsoome> well, i wanna test the app itself, I have OVDC on my openvpn client host anyhow, so accessing the server isnt the issue;)[10:58:30] *** Whoopsie has joined #openindiana[10:58:30] *** ChanServ sets mode: +v Whoopsie[11:03:08] <tsoome> ya well jailbreaking damn thing is one option, but i really dont wanna go that path:)[11:07:23] <lblume> Of course not. Apple knows what is good for you. You wouldn't disappoint them, would you?[11:09:03] <tsoome> :P[11:09:31] <tsoome> nah, well, yes and no.[11:10:46] <tsoome> it can be disturbing, but then again, I wont change that for android, where i need to have an virus scanner;)[11:11:17] <tsoome> everything has its price:([11:19:20] *** xmikus01 has quit IRC[11:20:43] *** merzo has quit IRC[11:24:05] *** held has quit IRC[11:27:42] *** dijenerate has quit IRC[11:29:16] *** laen_ has quit IRC[11:34:45] *** laen_ has joined #openindiana[11:36:26] *** dijenerate has joined #openindiana[11:42:17] <DontKnwMuch> what whould you say if I were to use a single SSD for slog[11:42:43] <DontKnwMuch> am I looking for trouble?[11:42:50] *** mikw has joined #openindiana[11:43:04] <tsoome> well, in some extent[11:43:27] *** held has joined #openindiana[11:43:34] <DontKnwMuch> Will the pool keep working even if it fails, I mean, has anyone experience with 28 and slog fail?[11:43:53] <tsoome> the worst case scenario is if you had crash/poweroff and your ssd will die before pool completes its import.[11:44:15] <tsoome> that means the log on ssd is gone and that much of data is lost.[11:44:52] <DontKnwMuch> but the pool will still import?[11:45:16] <tsoome> never done that, cant tell for sure;)[11:45:27] <DontKnwMuch> ;)[11:45:39] <quasi> http://mirrorshades.net/post/1485951163 is a good example of what can go wrong[11:45:47] <tsoome> well, you can force it to rollback to last know good state, but i have no idea how foolproof that is in such case[11:46:09] <DontKnwMuch> mirrorshade I read :) but it was on a old version I beleive[11:46:35] <tsoome> old u7, yes[11:46:49] <DontKnwMuch> could work better now[11:47:21] <DontKnwMuch> The problem is that I can not phisically fit two ssd in the case right now... I have them on the table and look around for some nasa tape ;)[11:47:35] <DontKnwMuch> one fits great[11:47:38] <DontKnwMuch> lol[11:48:45] <tsoome> if it will die during normal job, its just an perfomance imapct and before downing the system, its good idea to remove it from the pool[11:49:08] <brandini> how do I get rtorrent on oi148?[11:49:20] <tsoome> pkg install rtorrent[11:49:56] <brandini> there it is![11:50:05] <brandini> been months since i've touched that box :)[11:50:23] <DontKnwMuch> I will stuff two in, will sleep better[11:51:16] <DontKnwMuch> sadly just for slog, I read it is not recommended to use it for boot/cache too at the same time (partitioned and stuff)[11:53:12] <kinesis> if cache ssd die in pool, what happens with data?[11:53:57] <tsoome> nothing, it was your cache what did die, not data[11:54:48] <kinesis> i dont need to mirror cache device in pool, right? instead of log.[11:54:52] <DontKnwMuch> If I have for example 48Gb of ram, I do not need ssd cache, or will it still do something for me[11:55:09] <tsoome> tht depends;)[11:55:12] <tsoome> that*[11:55:23] <DontKnwMuch> everything depends, ;) argh[11:55:32] <tsoome> if all the ram is for zfs only, you have max 47GB in arc.[11:55:49] <DontKnwMuch> this is enough for me[11:55:52] <tsoome> but if you are running apps, all the game is changed[11:55:53] <DontKnwMuch> :)[11:56:03] <DontKnwMuch> nothing else, just zfs[11:56:12] <tsoome> another thing is , how big is the actively used data[11:56:42] <tsoome> if you have 100TB of data, 1TB in active use, then sure, the l2arc will help[11:57:59] <DontKnwMuch> true. got it now. Another one, slog will not help with smb shares, they are async, right[11:58:02] <tsoome> luckily for you, there are already systems around which can fit 1TB of ram;)[12:02:42] *** Whoopsie has quit IRC[12:16:21] *** |AbsyntH| has quit IRC[12:21:08] *** dijenerate has quit IRC[12:23:56] *** staticnote has joined #openindiana[12:28:47] <DontKnwMuch> hm.. are not two same SSDs in mirror going to 'fail' / wear out at the same time more or less... will mlc drive survive for long.[12:30:37] *** nbprashanth has quit IRC[12:33:25] *** sivanov has joined #openindiana[12:33:38] *** DeanoC_ has quit IRC[12:34:10] *** Micr0mega has left #openindiana[12:36:06] <sivanov> iperf gives me 860Mbit/s between OI and windows 7 host, i get 350-400MB/s dd'ing files (movies) from zfs filesystem,however, i get only 35MB/s copying those files to NUL on windows PC[12:36:21] <sivanov> what to do to get more from the CIFS share?[12:40:10] <Warod> sivanov: Win7?[12:40:17] <sivanov> yep[12:40:39] <sivanov> used to be 60MB/s+ when i used samba on freebsd[12:41:10] <Warod> well, I've had very close to 1 Gbps speeds with OI cifs[12:41:15] <Warod> to win7 computers[12:41:43] <sivanov> without any tuning?[12:43:27] <DontKnwMuch> Full glan speed on win7 from OI here too, like 95+/s[12:43:36] <DontKnwMuch> and no tuning[12:44:08] <tsoome> sivanov: that speed was from first run, run second time as well[12:44:12] <tsoome> ?[12:44:57] <sivanov> i test on big files, like 25GB[12:45:02] <sivanov> no improvement[12:45:29] *** staticnote has quit IRC[12:46:44] <sivanov> the NIC is bge[12:47:45] <tsoome> did you test iperf on both directions?[12:47:50] *** dijenerate has joined #openindiana[12:48:20] <tsoome> someone had an issue where data stream to OI was ok, but from OI was terrible[12:49:28] <sivanov> oi->windows get 860MBps[12:49:40] <sivanov> windows->ou ~450MBps[12:50:19] <sivanov> however, i get terrible speed from oi to freebsd 8.2, like 200MBps[12:50:42] <sivanov> tried tuning tcp window sizes, no effect[12:51:12] <sivanov> the switch is dell powerconnect 2724[12:52:16] *** staticnote has joined #openindiana[12:53:41] <kinesis> sivanov, i have OI cifs server. my speeds to/from win7 is up to 95/s[12:54:02] <sivanov> what hardware?[12:54:17] <kinesis> no tuning, but i have 4 nics in aggr[12:54:26] <kinesis> hardware of server?[12:54:31] *** staticnote has quit IRC[12:54:36] <sivanov> yes, of server[12:55:13] <sivanov> and btw, 4 nics in aggr won't give you more mbps to single client than single link[12:55:44] <kinesis> some mb with phenom x3, 6 hdd in raidz (onboard ahci)[12:56:06] <kinesis> i know, but i have many win7 users )[12:56:18] <kinesis> intel nics[12:57:21] <kinesis> and HP procurve 2810-48g switch[12:58:18] <sivanov> my server is two dualcore opteron 285 at 2 dot 4Ghz, 16Gb ram, 15 hdds in 3 groups raidz and broadcom 5704c bult-in nic[12:58:37] <kinesis> but i have same problem on linux clients. from linux to OI i have 95/s, but from OI to linux only 35/s[12:59:41] <kinesis> try to use jumbo-frames if your switch support it.[13:00:23] <kinesis> i think intel nics it the best choise )[13:03:12] <sivanov> just today i brought a usb hard disk in our server room with centos 5.2 just to see if there is any problem with throughput of ethernet connection from oi:[13:03:44] <sivanov> the connection speed is 985Mbit/s between io and centos in both directions[13:04:03] <sivanov> 9000 mtu though[13:04:44] <sivanov> this time oi is on ibm x3650 server[13:06:09] <sivanov> with this server i tried to troubleshoot ESX 4.1u1 connection to iSCSI with 4x link MPIO[13:06:56] *** keremet has joined #openindiana[13:07:28] *** kart_ has joined #openindiana[13:08:39] <sivanov> got more problems than answers: this server gives only up to 67MB/s read from iscsi lun on 6disk raidz10 to the ESX host[13:09:30] <sivanov> the 67MBps are evenly spreaded between the gigabit ethernet links[13:10:01] <tsoome> 128k records or default?[13:10:04] <sivanov> yes[13:10:44] <tsoome> uhm, yes to which one?:P[13:11:06] <sivanov> 128k[13:11:51] <tsoome> comstar?[13:12:00] <sivanov> the test was iozone in throughput mode: 8 processes, 64 queue depth, 1M record[13:12:08] <sivanov> yes, comstar[13:12:55] <sivanov> no errors on the switch or whatever[13:14:04] <sivanov> just no difference, 1 link using vmware fixed path = 67MB/s, or 4 links using round-robin = same 67MB/s[13:14:25] <tsoome> thats ... curious case:)[13:15:54] <nettezzaumana> i don't understand why you still advance the same ;) .. every little kid knows that oi (sunos) is degrading i/o compared to *anything else on same hardware[13:16:12] <tsoome> if you try download over say, ftp, what troughput you will see?[13:16:38] <sivanov> trying to troubleshot another case: we got infortrend a16e-g2140 iscsi raid shelf. 16x 2GB hitachis in raid6. cant get above 106MB/s sequental read[13:18:09] <sivanov> get about 200MB/sec to a linux with 4 links and multipathd runnning[13:19:26] <sivanov> though what i really want is to get 95MB/s from cifs share at home[13:19:42] <sivanov> to watch BD movies without frame loss :D[13:21:11] *** DrLou has joined #openindiana[13:21:11] *** ChanServ sets mode: +o DrLou[13:23:25] <nettezzaumana> sivanov: http://sunaytripathi.wordpress.com/2010/03/25/solaris-10-networking-the-magic-revealed/#mozTocId400304.125[13:23:29] <tsoome> I'd still suggest to test with simple data stream from filesystem first, such as ftp transfer, to se if the normal data transfer is with reasonable speed. once thats cleared, then its point to go to next layer[13:24:06] <nettezzaumana> sivanov: unfortunatelly solaris netstack has too robust default settings .. turning off some features will greatly improve performance[13:24:16] <tsoome> you should get quite reasonable results *without any tuning*[13:25:46] *** merzo has joined #openindiana[13:30:46] *** miine has joined #openindiana[13:36:48] <lblume> I thought bluray was doing like 50 Mbps? Any half-decent fastethernet should be able to handle it :-D[13:41:42] *** dijenerate has quit IRC[13:44:17] *** held has quit IRC[13:46:28] <kinesis> i'm waiting for a new server with 31 disks for data. i think to make pool of 15 vdevs of 2-way mirrors, or 6 vdevs of 5 disk raidz. which configuration will be faster?[13:46:50] <tsoome> faster for what?[13:47:09] <kinesis> faster for reading. it will be cifs server[13:48:34] <kinesis> i cant decide.[13:48:55] <tsoome> for pure sequential reads the wide vdevs are better.[13:50:04] <kinesis> thanks! sure it will not be a sequental reading.[13:50:26] <kinesis> + 2 ssd for zil, and 1 ssd cor l2arc[13:53:25] <kinesis> i will not have enough time for testing.[13:53:53] <tsoome> ou, thats perfect case for fails then;)[13:54:00] *** gea_ has joined #openindiana[13:54:13] <kinesis> ))[13:54:36] <kinesis> i mean test both configurations.[13:54:52] <dkeav> screw it, just deploy[13:54:53] *** dijenerate has joined #openindiana[13:54:55] <dkeav> what could go wrong[13:54:58] <dkeav> :D[13:55:43] <DontKnwMuch> dkeav: this is like me talking.. lol[13:56:15] <dkeav> sounds like me in meetings where i'm being pressured on time constraints[13:56:15] <dkeav> :D[13:56:33] <dkeav> i usually throw a lot of "well in theory"'s in there too[13:56:34] <dkeav> hehe[13:56:58] <kinesis> theory... )[13:57:40] <tsoome> well, if its raidz versus raid10, if you set up raid10 you will just loose disk space, thats all, from perfomance point of view its ok for basically all cases[13:58:09] <dkeav> especially with caches[13:59:33] *** Meths_ has joined #openindiana[13:59:52] <tsoome> wide raidz can be better with streaming IO, but as you told there is none, thats not the case then.[14:00:28] <kinesis> thanks![14:00:51] <kinesis> will try to get more time for tests.[14:01:18] <tsoome> but, if for price/perfomance the raidz would be ok as well - based on your information - cant tell.[14:02:00] <tsoome> how big disks btw?[14:02:05] <kinesis> 2tb[14:02:18] *** Meths has quit IRC[14:03:34] <tsoome> well, with those, the question is, what is the chance you will loose the second half of the mirror during the resilver;)[14:04:06] <kinesis> i need more experience. working with solaris for half-year )[14:04:36] <tsoome> that last one has nothing to do with solaris[14:05:50] <dkeav> nothing at all[14:06:57] <kinesis> why i lose half? didnt understand.[14:08:03] <dkeav> if a drive dies, and a big drive at that, when you replace it and start rebuilding (resilver) the mirror, it takes a long time to copy over 2tb of data[14:08:18] <dkeav> you are vulnerable to another disk failure during that time[14:08:41] <kinesis> understood )[14:08:43] <tsoome> 1. disks do die. 2. the larger disk, the more time will it take to resilver the data, 3. during the resilver the disks are on more stress as they need to server normal IO and deal with resilver - and with more IO stress, its a bit more likely another disk will die[14:09:33] <dkeav> as long as you have good backups and downtime is not an issue if this were to happen, meh[14:10:00] <dkeav> but if high availability is required at all times, it could be an issue down the road[14:10:18] <kinesis> now i think vdevs of raidz will be better choise )[14:10:38] <kinesis> yes, high availability required.[14:10:41] <tsoome> raidz1 is just as vulnerable[14:10:52] <dkeav> yep[14:11:02] <tsoome> you can loose exactly 1 disk[14:11:06] <dkeav> you would need to do say 3 vdevs of raidz2[14:11:07] <tsoome> from vdev[14:11:36] <dkeav> that would be the most robust to failure[14:12:02] *** POloser has left #openindiana[14:12:07] <tsoome> as you see, its going to get way more complicated;)[14:12:30] *** lkocman has quit IRC[14:12:39] <tsoome> with more parity, the question of perfomance will kick in again;)[14:13:09] <dkeav> how many actual data disks do you have to work with, no ssd caches or OS, but your 2tb array disks?[14:13:33] <dkeav> tsoome: yea but with striping them and ssd caches, should be plenty fast[14:13:58] <tsoome> true[14:14:13] <kinesis> 31 disk at all. for data.[14:14:13] <dkeav> and sounds like his use case requires more data security/availability than raw performance[14:14:29] <dkeav> brb, phone[14:15:10] <tsoome> also, if the raidzX perfomance is not ok even with small vdev, there is always case of 3-way mirror[14:15:11] <man_u> you can have triple mirror, best for speed and reliability[14:16:16] <tsoome> OR. forget 3-ways at all, build normal 2-way mirror, get all the speed and quite reasonable reliability, and build second array with replicated data for fast recovery in case the ....[14:16:23] <kinesis> thanks for your suggestions ))[14:16:55] <tsoome> sometimes the rabbit is just the rabbit, no need to overthink.[14:18:05] <kinesis> need time to think.[14:18:32] <tsoome> the best suggestion still - try to take some time for tests.[14:19:09] <dkeav> yea i would either do large mirror set or 3 raidz2 vdevs of 10 drives, with 1 hotspare[14:19:24] <dkeav> stripe across the raidz2's[14:20:01] <dkeav> that gives you what, about 48Tb of storage[14:20:30] <dkeav> can lose 2 drives in any of the 3 vdevs, and have a hotspare online[14:21:31] <kinesis> dkeav, you read my mind now )[14:21:43] <kinesis> just want to ask it[14:21:46] <dkeav> would want plenty of cpu and memory[14:22:43] <tsoome> parity based raids do like ram[14:22:46] <kinesis> in will be xeon e5620 with 24gb of ram[14:23:02] *** dijenerate has quit IRC[14:23:14] <dkeav> with ssd caches too, that should be fine[14:24:19] *** movement has quit IRC[14:24:47] *** bens1 has quit IRC[14:24:56] <kinesis> thanks, guys![14:28:36] *** obit_sweden has joined #openindiana[14:28:47] *** obit_sweden has left #openindiana[14:29:09] *** ohub has quit IRC[14:33:15] *** DeanoC_ has joined #openindiana[14:33:15] *** ChanServ sets mode: +o DeanoC_[14:35:02] *** dijenerate has joined #openindiana[14:36:53] *** movement has joined #openindiana[14:44:10] *** held has joined #openindiana[14:44:12] *** hajma_ has joined #openindiana[14:44:49] *** hajma has quit IRC[14:49:07] *** Whoopsie has joined #openindiana[14:49:07] *** ChanServ sets mode: +v Whoopsie[14:49:19] *** CoilDomain has joined #openindiana[14:58:56] *** dekar has joined #openindiana[15:04:54] *** Phenox_ has quit IRC[15:07:22] *** heldchen has joined #openindiana[15:08:56] *** DeanoC_ has quit IRC[15:11:21] *** held has quit IRC[15:13:14] *** raichoo has quit IRC[15:13:26] *** raichoo has joined #openindiana[15:16:23] *** merzo has quit IRC[15:16:32] *** Phenox_ has joined #openindiana[15:18:09] *** bens1 has joined #openindiana[15:20:56] *** DanaG has joined #openindiana[15:27:47] *** |AbsyntH| has joined #openindiana[15:30:26] *** DanaG has quit IRC[15:43:59] *** keremet has quit IRC[16:01:17] *** Reepicheep has quit IRC[16:02:50] *** hsp has joined #openindiana[16:05:21] *** akamit has quit IRC[16:05:44] *** lkocman has joined #openindiana[16:06:09] *** Whoopsie has quit IRC[16:11:52] *** skeeziks has joined #openindiana[16:13:17] <Infin1ty|work> i'm buliding a storage server, going to use openindiana, the main goal of this server is to be a dedicated NFS server only, the root pool is a vdev made of 2x500GB 7.2k disks (mirrored), the nfs zpool is made out of 4 vdevs, each vdev is a raidz1 that has 4x1TB 7.2k disks, i'm also adding a mirrored log device (ZIL) made of 2x40GB Intel X-25E Extreme SSD Drives, and for the L2ARC i'm going to use 3x40GB SSD Drives (same model as t[16:13:17] <Infin1ty|work> he ZIL cache), this should give me 15TB for the NFS, is there any reason not to use 4 vdevs with raidz1 configuration?[16:14:41] <tsoome> you can have as many vdevs as you like. the more the merrier;)[16:15:05] <dkeav> yea that should work just fine[16:17:14] <sickness> 2 broken drives in 1 vdev can make you loose ALL the data... with a raidz3 you could loose whichever 3 of those 16 disks and not loose data...[16:17:33] <sickness> or you could mirror 2 vdevs with 8x1tb raidz2 drives[16:19:49] <Infin1ty|work> sickness, i know, i took this into consideration[16:20:16] <Infin1ty|work> sickness, i don't have spare room for more disks[16:20:37] <dkeav> just setup a proper backup system and you'll be fine[16:21:38] <Infin1ty|work> thanks :)[16:26:46] <DontKnwMuch> hm... I was deciding between a 3x 3 drive raidz and 8 drive raid10, iops ARE way higher in raid 10...[16:34:51] <h____> is it possible to extend a raidz1 zpool with extra disks yet?[16:34:58] <h____> i.e. i'm currently running 4+1 1.5tb disks, I'd like to extend that to 5+1[16:36:13] *** ohub has joined #openindiana[16:38:15] <tsoome> no[16:38:15] *** ohub has quit IRC[16:38:52] <tsoome> you still can add another vdev in 4.1 config tho.[16:39:19] <h____> before the whole oracle thing happened someone said it was on the roadmap[16:39:34] <h____> is it ever gonna happen?[16:39:38] <dkeav> http://blogs.oracle.com/ahl/entry/expand_o_matic_raid_z[16:41:13] *** ohub has joined #openindiana[16:42:30] <h____> thanks dkeav I hadn't seen that page even though its 3 years old[16:42:33] <tsoome> the thing is, you dont wanna make your raidzX too wide anyhow. while the need to expand is quite understandable for home users, its still quite the corner case....[16:43:30] <h____> if I added another 4.1 vdev to the zpool, and either one of the vdevs gets corrupted, the whole zpool will be gone too though[16:44:14] <h____> though I guess I'd go for 2tb disks now anyway not 1.5s[16:44:43] <dkeav> yea you could always just replace the disks in your current zpool[16:44:46] <tsoome> true. but making wider raidz vdev doesnt help about that;)[16:44:59] <dkeav> if you are needing more storage that is[16:45:09] <h____> maybe add a 2.1x2tb to my 4.1x1.5tb[16:45:34] <tsoome> the thing with wider vdev is that writes are still done with speed of single disk (unless its streaming write)[16:53:33] *** Nitial has quit IRC[16:56:51] *** axisys has joined #openindiana[16:57:42] *** raichoo has quit IRC[16:59:10] *** kinesis has quit IRC[16:59:37] *** Nitial has joined #openindiana[16:59:43] *** mnaser has joined #openindiana[16:59:45] *** DontKnwMuch has quit IRC[17:06:44] *** ilc has joined #openindiana[17:14:43] *** Thrae has quit IRC[17:15:40] *** Thrae has joined #openindiana[17:22:45] *** miine has quit IRC[17:27:22] *** gea__ has joined #openindiana[17:29:23] *** miine has joined #openindiana[17:40:11] <wonko> ok, doing the Solaris -> OI "upgrade"[17:40:14] <wonko> let's see how this goes[17:41:14] *** Infin1ty|work has quit IRC[17:43:40] <Triskelios> probably won't, if you're using S11X's zpool version[17:44:21] <wonko> S10[17:45:28] *** akamit has joined #openindiana[17:48:12] <lblume> there's a way to upgrade S10 to OI?[17:49:12] <wonko> http://blogs.oracle.com/bobn/entry/solaris_and_opensolaris_coexistence_in[17:49:17] <wonko> kinda[17:49:22] <wonko> by leveraging LU/BE[17:51:23] *** Naresh has quit IRC[17:51:57] <lblume> Ugh. Kinky.[17:52:26] <dkeav> masochist[17:55:50] *** longcat has quit IRC[17:57:46] *** DanaG has joined #openindiana[17:57:56] <DanaG> Say, how do I make openindiana install the 64-bit kernel?[17:58:07] <DanaG> I'm running the install itself within Virtualbox, with 1.5GB RAM.[17:58:20] <DanaG> But the final target will be the Microserver, with 5GB.[18:02:48] *** heldchen has quit IRC[18:02:52] <lblume> there is no way to NOT install the 64 bit kernel.[18:04:00] <DanaG> Cool. So does it do "fat binaries" like OS X?[18:04:48] <DanaG> I can't imagine any other way to have mixed 32/64 on same install.[18:04:55] <Triskelios> no, we use isaexec and multiple subdirs[18:06:14] <DanaG> Say, is a Broadcom 5723 likely to work well, or is an Intel 82574L likely to work significantly better? I know the Broadcom is fine in FreeBSD, but less fine in Linux.[18:06:53] <lblume> Only NeXTstep and descendant used the fat binaries way. Others found it too clumsy.[18:07:19] <DanaG> I know those fat EFI binaries are annoying... if you want to load it on any other EFI system, you have to split it.[18:07:53] <dkeav> fat binaries are annoying period[18:08:17] <dkeav> and you will want the intel nic[18:08:29] *** lkocman has quit IRC[18:08:47] <dkeav> bge is fine for light duty, but i wouldn't use for any serving roles[18:08:48] <Triskelios> the Broadcom NIC should work fine, with the exception of oi_148 having a slightly busted bge driver on the DVD[18:08:54] <Triskelios> so you may have to disable it for the install[18:09:25] <DanaG> I'd be using it for torrents and for syncing a big dir with tons of pictures.[18:10:07] <dkeav> if its working for you, by all means use it, just be forewarned[18:10:14] <lblume> just make sure the 5723 is indeed supported. Broadcom is a little messy.[18:10:29] <dkeav> little?[18:10:57] *** mnaser has quit IRC[18:11:00] <DanaG> All of HP's AMD-based business systems use Broadcom; not one Realtek or Marvell.[18:11:25] <DanaG> I wonder if that means Marvell is worse. I'd already imagine realtek is definitely worse.[18:11:25] <dkeav> thats a good think[18:11:28] <dkeav> *thing[18:11:49] <dkeav> intel>broadcom>marvell>realtek[18:12:41] <dkeav> err i mean intel>broadcom>marvell>*>realtek[18:14:07] <lblume> I've had less issues with Realtek than Marvell[18:14:11] *** hajma_ has quit IRC[18:15:06] <dkeav> honestly i havn't used a marvell chipset on a solaris system before[18:15:14] *** mit__ has joined #openindiana[18:15:17] <dkeav> but i have realtek and it is unusable[18:15:47] <dkeav> if marvell is worse then apparently you can't even boot a system with a marvell :D[18:16:12] <DanaG> My netbook has a Realtek wifi card, and the Linux driver is a total failure. as in, "Initialisation is failed!!"[18:16:17] <lblume> Marvell drivers are not included, so if you hit an issue, you're SOL[18:17:00] <Triskelios> we include yge actually, for Yukon II[18:17:01] <DanaG> I'm trying to figure out whether to just try it with Broadcom for now, or just buy the Intel card now. 25 bucks is not all that much.[18:17:07] <dkeav> easier solution, buy mobo's with intel nics or buy intel pci/-e nics[18:17:51] <lblume> Triskelios: I was talking about regular manufacturer support, but right :-) I need to try rge, actually.[18:18:00] <Triskelios> DanaG: that NIC is supported[18:18:15] <lblume> dkeav: It has become very hard to know what chipset is present on consumer level mobos[18:18:54] <Triskelios> I've use rge in laptops, not that it means it'll meet your needs[18:19:15] *** akamit has quit IRC[18:21:11] <DanaG> Now, if I were to have a pair of mirrored drives on-site, and a third for backup off-site, is there any point to rotating out which drive is offsite?[18:22:47] <Triskelios> I imagine that's much more physical stress for the drives, and a net loss in reliability[18:23:18] *** ravenslay3r has joined #openindiana[18:24:27] <DanaG> eh, I guess I might as well get the Intel card. The only other card I have on hand, besides the built-in Broadcom... is another Broadcom. 5721.[18:24:52] <ravenslay3r> Has anyone reported the Gnome bug in OI_148?:[18:24:53] <ravenslay3r> After clicking 'save as' in any program, if a 'new folder' is created in the 'save as' window, the application will crash prior to saving the document.[18:26:59] *** held has joined #openindiana[18:30:29] <dkeav> lblume: i usually purchase mobo's from newegg, majority of the time they will list the chipset manufacturer in the description for the lan chipset[18:31:26] <lblume> Ok, because manufacturers themselves sure don't do it very much those days.[18:31:50] <dkeav> no, not very well[18:31:58] <dkeav> usually if its not listed on newegg its realtek[18:32:08] <dkeav> or safe to assume[18:32:13] <dkeav> especially on cheap boards[18:32:28] <dkeav> http://www.newegg.com/Product/Product.aspx?Item=N82E16813121387[18:32:44] <dkeav> if you look in the description under lan chipset, intel pro 82578DC[18:37:36] *** mnaser has joined #openindiana[18:41:43] *** gea__ has quit IRC[18:43:49] *** raichoo has joined #openindiana[18:46:28] <DanaG> heck, I could even get two offsite drives.[18:47:48] *** ravenslay3r has quit IRC[18:48:39] <wonko> hmmmm[18:48:45] <wonko> wonko@zaphod$ sudo zfs receive SYS0/ROOT/oi-151 < /export/oi/zaphod.zfs[18:48:45] <wonko> cannot receive: stream is unsupported version 17[18:48:47] <wonko> foo[18:50:23] *** ravenslay3r has joined #openindiana[18:56:18] *** wonko has quit IRC[18:57:41] *** gea__ has joined #openindiana[18:58:31] *** |AbsyntH| has quit IRC[18:59:08] *** sergefonville has joined #openindiana[18:59:14] <sergefonville> good evening[19:04:01] *** fossala has joined #openindiana[19:04:12] *** dijenerate has quit IRC[19:05:34] <fossala> Is there anything like ltsp (linux terminal server project) for OI/OpenSolaris?[19:06:34] <Triskelios> there's Sun Ray...[19:06:42] <alanc> not specifically ltsp, but most of the software ltsp uses is either provided or buildable, so you could build your own[19:06:45] <alanc> and that too[19:20:09] *** wonko has joined #openindiana[19:22:04] <infinity_> I have inheritance "fd" setup on the parent directory. When files are moved on the server, they don't take the directories permissions. They keep their original permissions. Is there a way to get around this and have the file/dir inherit its new parent directory permissions? This is on oi_148[19:30:44] <sergefonville> could you post your /usr/bin/ls -lav output for the relevant directory[19:31:10] *** fossala has quit IRC[19:32:09] <sergefonville> also what is your aclinherit of the filesystem...[19:32:49] *** yalu has quit IRC[19:34:36] <sergefonville> acls are only applied on creation[19:34:56] *** gber has quit IRC[19:34:57] <sergefonville> when inheritance is on[19:35:06] <sergefonville> if you move a file, acls are maintained[19:36:38] *** DontKnwMuch has joined #openindiana[19:37:01] *** yalu has joined #openindiana[19:37:02] <infinity_> k. let me pull up the info[19:38:15] <DontKnwMuch> I have somehow fuxated my netowrk, I am getting network/physical:default misconfigured transitioned to maintenance after the /lib/svc/method/net-physical failed with exit status 96[19:38:26] *** Botanic has joined #openindiana[19:38:27] *** Botanic has quit IRC[19:38:27] *** Botanic has joined #openindiana[19:38:28] <DontKnwMuch> It seems my second nic prevents this to work ok[19:38:48] <DontKnwMuch> failed to configure IPv4 DHCP interfaces: e1000g1[19:39:10] <DontKnwMuch> how can I disable it completely[19:39:18] <DontKnwMuch> unplumb did not help[19:40:08] *** mnaser has quit IRC[19:40:48] <infinity_> http://pastebin.com/Ug0EBVTW[19:41:09] *** Nitial has quit IRC[19:41:15] <infinity_> sergefonville: http://pastebin.com/Ug0EBVTW[19:41:38] <DontKnwMuch> this network setup in oi is not very intuitive..[19:42:02] <infinity_> DontKnwMuch: working with network interfaces on opensolaris has always been confusing. i never could find the correct answers[19:42:18] <infinity_> DontKnwMuch: i have development "my" method :)[19:42:30] *** Nitial has joined #openindiana[19:43:24] <infinity_> DontKnwMuch: if i'm not using dhcp i disable nwam[19:43:42] <infinity_> you might try the ipadm and dladm commands[19:44:37] <DontKnwMuch> will check it out... I also did it onece by just editing the files for static ip[19:44:38] <dkeav> DontKnwMuch: what are you trying to do?[19:44:50] <dkeav> static ip? use ipadm[19:45:16] <infinity_> DontKnwMuch: this is my static IP instruction sheet i made for myself: http://pastebin.com/KrzhR6N3[19:45:25] <dkeav> did you disable nwam and enable physical?[19:45:33] <infinity_> dkeav: yea. thats what i do[19:45:36] <DontKnwMuch> yes I did[19:45:52] <DontKnwMuch> and now it does the maintenance mode... argh...[19:46:23] <DontKnwMuch> will try your insctructions, hey look good, thank you :)[19:46:30] <DontKnwMuch> hey = they[19:46:36] <dkeav> ipadm create-addr -T static -a ###.###.###.###/netmask interface/v4static[19:47:01] <dkeav> eg ipadm create-addr -T static -a 192.168.10.10/27 bge0/v4static[19:47:11] <infinity_> DontKnwMuch: thanks :). they took me forever to develop a solution. I played with nwam for a long time and the unconfigure method. tried 'em all at one time and thats what i came up with[19:47:53] <infinity_> sergefonville: ping. looking for some tips on inhertance and permissions :)[19:48:48] <dkeav> DontKnwMuch: you still need to create hostname.interface in /etc and set your default route with route -p[19:48:58] <dkeav> iirc[19:49:13] <infinity_> dkeav: whats the correct way to setup a default route? I used /etc/defaultrouter[19:50:22] <dkeav> infinity_: use route -p[19:50:40] <infinity_> dkeav: okay. I'll change my docs to try that next time.[19:51:02] <dkeav> want a bash script to help with setup?[19:51:31] <dkeav> http://pastie.org/2178797[19:52:14] <infinity_> dkeav: lol. is it me, or the fact that you offer a bash script to configure an interface raises red flags..[19:52:34] <dkeav> yea its a virus[19:52:38] <dkeav> i'm sneaky like that[19:52:39] <dkeav> ...[19:52:56] <richlowe> dkeav: sure you don't want -p in the default route?[19:53:09] <dkeav> i work with idiots dude, its easier to replace them with small shell scripts than explain things over and over again[19:53:12] <infinity_> like... maybe whoever developed the interface configuration is an idiot[19:53:50] <dkeav> richlowe: crap, yes, yes i do[19:54:05] <infinity_> dkeav: i'm not saying its a virus. i'm saying something is wrong. it shouldn't be that hard to require a script.[19:54:35] <infinity_> and i think it's toooo hard relative to other things i've administered[19:54:43] <dkeav> its no hard per say, just sometimes its easier to remember if stuff is written down[19:54:50] <dkeav> plus ipadm doesn't have a manpage[19:55:30] <infinity_> dkeav: you script doesn't disable nwam. why is that?[19:55:31] <dkeav> i mean ideally if you have a dhcp server on your network just set the server sticky and let nwam handle it[19:55:32] <DontKnwMuch> nice bash script[19:55:44] <dkeav> infinity_: thats just to set a static address[19:55:47] <dkeav> or change it[19:55:53] *** Meths_ is now known as Meths[19:56:07] *** ChanServ sets mode: +o Meths[19:56:09] <DontKnwMuch> I was thinking the same, use DHCP to set it up, and leave it as it is, but even my DHCP does not work anymore[19:56:11] <infinity_> whats the purpose of nwam if you're using static IPs then? shouldn't it be disabled?[19:56:12] <dkeav> i guess you could add checks for the status of nwam/physical and set accordingly[19:56:56] <DontKnwMuch> second interface is making me problems[19:57:01] <dkeav> infinity_: statically assigning an IP from dhcp server is not the same as statically assigning it on the machine itself[19:57:24] <DontKnwMuch> if the machine starts before dhcp it does not get the ip[19:57:31] <dkeav> it is easier to manager your network in more central locations than have to log into each machine you set up statically whenever something needs to change[19:58:06] *** Naresh has joined #openindiana[19:58:24] <infinity_> dkeav: maybe i'm confused. doesn't you script set a static IP on the machine?[19:59:18] <dkeav> yes, cause thats what you asked for[20:00:19] *** ravenslay3r has left #openindiana[20:01:14] <dkeav> another option would be to let your dhcp server assign your hosts "sticky" addresses[20:01:39] *** bens1 has quit IRC[20:01:41] <dkeav> this keeps all your configuration dynamic at the host and centrally controlled on the network[20:04:23] <infinity_> it sounds like we're talking about different things. you're talking about machine managment of IPs and i'm talking about the purpose of NWAM when setting a static on the machine.[20:06:29] <DontKnwMuch> address object already exists... hm..[20:06:38] <DontKnwMuch> ipadm tells me this[20:10:15] <DontKnwMuch> I will try by this method I 'developed': http://pastebin.com/tKxwXU6V[20:11:35] <Triskelios> DontKnwMuch: that works if you have nwam disabled; ipadm can be used instead of editing /etc/hostname.*[20:11:41] <infinity_> DontKnwMuch: that doesn't look right. that like solaris 2.6 -ish[20:12:25] <DontKnwMuch> it works, ipadm is bugging me too much[20:12:40] <DontKnwMuch> well, at least it did when I wrote it down[20:13:44] <DontKnwMuch> is it possible that nwam in enabled for one and disabled for another interface? it seems like that for me.. hm..[20:14:45] <Triskelios> no, nwam is systemwide[20:14:53] <dkeav> huh, no the service should be disabled if you disabled it[20:15:03] <dkeav> svcs nwam[20:15:15] <DontKnwMuch> rebooting...[20:15:39] <dkeav> should say disabled[20:16:38] *** dijenerate has joined #openindiana[20:17:25] <DontKnwMuch> ok, it is disabled[20:17:41] <DontKnwMuch> now I will try ipadm for both interfaces[20:20:47] <DanaG> hmm, I installed openindiana (graphics version) in Virtualbox, and it's getting to like 20% on the progress bar, and then instantly resetting the VM.[20:23:21] <dkeav> DanaG: did you enable "Use host I/O cache"?[20:23:51] *** dijenerate has quit IRC[20:24:21] <DanaG> Yes.[20:24:55] <Triskelios> DanaG: see the http://www.illumos.org/projects/illumos-gate/wiki/How_To_Report_Problems instructions on how to remove the splash image and enable the debugger[20:24:55] <DanaG> What's more, I used dd to copy that back to a USB drive, and that's doing the same thing on a real machine.[20:25:33] <DanaG> I can't get anywhere near where I'd get a crash report.[20:25:48] <DanaG> oh, boot -vk verbose[20:28:35] *** mnaser has joined #openindiana[20:29:05] <DanaG> NOTICE: cannot read the pool label from '/pci@0,0/pci-ide@1f,1/ide@0/cmdk@1,0:a'[20:30:35] <DanaG> The physical one is booting from USB flash drive; vbox is booting from AHCI.[20:30:40] <DanaG> With cache enabled.[20:31:01] <Triskelios> uh, you need the reimport the root pool if you change its location[20:31:13] <Triskelios> boot from live media and import -f once[20:31:28] <DanaG> But the vbox is where I originally installed it to.[20:31:43] <DanaG> er, wait, maybe that was with it in ide mode. I'll try that import.[20:32:47] <Triskelios> yes, the phys path is IDE right now[20:33:32] *** blues_ has joined #openindiana[20:34:03] <DanaG> When I installed directly to the flash drive, the thing didn't create my user on the system![20:34:12] <DanaG> And it wouldn't accept any (default or my own) passwords.[20:34:19] <Triskelios> never seen that before[20:34:22] <blues_> is that intel controller card that is a rebadged LSI my best bet for a sas/sata controller for OI ?[20:34:50] <DanaG> And when I tried to mount the pool, it seemed to contain only /boot and /etc[20:35:39] <dkeav> blues_: pretty much[20:35:45] <Triskelios> DanaG: a pool contains multiple filesystems[20:35:47] <dkeav> thats your best option anyways[20:35:54] <dkeav> as its cheap and works very well[20:36:13] <blues_> If money were no object, what would you recommend?[20:36:19] <blues_> or at least, no object within reason[20:36:31] <dkeav> probably the same card[20:36:31] <dkeav> heh[20:36:44] <dkeav> it works well with cheap HP sas expanders[20:36:55] <Triskelios> any LSI SAS controller should be good, I think?[20:37:11] <dkeav> any that support IT mode[20:37:24] <dkeav> mostly the 1068 and 2008 chipset models though[20:38:05] <DanaG> Another time the "kexec"-ish quick-reboot after install resulted in a 45-minute failure to boot.[20:38:12] <dkeav> the intel version of the 1068 can be had for about 100 bucks, is pci-e x8 and has two minisas ports[20:38:50] <blues_> each mini-sas will do 4 devices wont it?[20:39:39] <dkeav> ahem sas expander[20:39:41] <mnaser> LSI 9200-8e -- I bought it to connect it to a JBOD with ZFS, I haven't personally tried it but I've seen people use it with no problems, as well as it's listed as a support device too[20:39:48] <dkeav> http://h18000.www1.hp.com/products/servers/proliantstorage/arraycontrollers/sas-expander/index.html[20:39:55] <mnaser> but then again i don't know what you guys are talking about so I'll just throw that out :p[20:45:47] <DanaG> hmm, how do you make gdm not start at boot, by default?[20:46:08] <richlowe> svcadm disable gdm[20:47:36] <DanaG> say, what do the parameters 'v' and 'k' do (as in -vk)? I want to enable boot info text, but not the debugger.[20:50:03] *** kart_ has quit IRC[20:50:57] <dkeav> blues_: http://www.newegg.com/Product/Product.aspx?Item=N82E16816117207[20:51:02] <dkeav> another expander option[20:51:22] <dkeav> that one has external power ability too, so you don't have to use a pci-e slot if you don't want[20:51:33] <dkeav> or you can even mount it in an external enclosure[20:51:52] <blues_> that would be handy.. i'll be running low on pci-e slots[20:52:42] <dkeav> you can use the molex for power[20:52:48] <dkeav> it supports 24 drives[20:52:55] <blues_> yeah..it looks rather pimp[20:53:01] <dkeav> the HP one i think goes up to 36[20:53:18] <DanaG> svc.startd[9]: svc:/network/location/default: Method or service exit timed out. Killing contract 52.[20:53:46] <blues_> ooh... it comes with cables too[20:53:48] <dkeav> you just plug your two 4 channel sas cables into that puppy and you get back 24[20:53:48] <blues_> nice[20:53:51] <dkeav> yay![20:54:41] <DanaG> http://www.big-bubbles.fluff.org/blogs/bubbles/archives/000622.html[20:54:55] <DanaG> So, it seems like the OpenIndiana USB is really slow.[20:55:24] <blues_> is flashing the intel card to IT-mode a pain?[20:55:58] <DontKnwMuch> this network setup makes me headaches: I can ping local machines, I can nslookup internet, but I can not ping anything outside of local net[20:56:08] <DontKnwMuch> 1068 is limited to 2TB drives[20:56:19] <DontKnwMuch> 2008 works with 3TB and more[20:56:39] <dkeav> DontKnwMuch: yea[20:56:53] <dkeav> not many people are onto 3tb for large vdevs yet though[20:57:21] *** smrt has quit IRC[20:57:37] <DontKnwMuch> yes, but in a year or so 3tb drives will be 60usd...[20:57:39] *** smrt has joined #openindiana[20:57:41] <DontKnwMuch> I have two network cards,on differnet subnets, does it matter? how is the default gateway in this case setup[20:58:11] <DontKnwMuch> I do not know why is it not working[20:58:23] <kimc> hi[20:58:38] <tomww> DontKnwMuch: if you watch closely the output of "netstat -nvr", which Interface and sending IP-Address if the one towards your outernet?[20:58:51] <tomww> *is[20:59:16] <DanaG> Okay, so how do I get the USB data transfer to not be slow?[20:59:33] <DanaG> It's not the drive making it slow -- it works fine with Ubuntu on the same flash drive.[21:00:42] <tomww> which filesystem is on that disk? if it is pcfs, then, well, no luck then.[21:03:24] <blues_> vmware's HCL can kiss my ass[21:15:36] <DanaG> Okay, would usb selective suspend be what's making USB so horribly slow?[21:15:58] <blues_> Oh.. hmm Dana... what motherboard are you using?[21:16:59] <DanaG> I've tried it (though it was a failed install) on an HP Microserver. I'm also now trying it on a SuperMicro H8SMI.[21:17:08] <DanaG> Failed install as in, it didn't create my user![21:17:48] <DanaG> It took like 20 minutes to fail to boot normally, and then enter maintenance mode.[21:18:00] <DanaG> This is booting from a USB stick that works fine with Ubuntu on it.[21:18:40] <dkeav> err wait, you couldn't boot it on a microserver with usb stick?[21:18:44] <dkeav> err that usb stick[21:19:03] <dkeav> seriously i installed a microserver in like 10 minutes with a usb stick[21:19:30] <DanaG> For me, the install look like 3-4 hours.[21:19:44] <DanaG> And that was the text install... that failed.[21:20:23] <DanaG> Now I redid the install onto a vbox hard drive, then used dd to copy it to the USB stick. Booting it via USB is horribly slow -- I get service timeouts.[21:20:24] <dkeav> something is very wrong with that usb stick then, at least with solaris[21:20:44] <dkeav> i used a kingston 4gb to install[21:20:48] <dkeav> like i said, 10 mins[21:21:04] <DanaG> My USB stick does 40MB/s read, 10MB/s write. 0.7ms access times.[21:21:13] <dkeav> ...[21:21:18] <dkeav> ok, its not the usb stick...[21:21:30] <DanaG> That's a benchmark under Linux, with palimpsest.[21:23:16] <blues_> hey dkeav, you have a link to that intel card?[21:23:33] <dkeav> which?[21:23:48] <dkeav> the 1680 controller, the 2008 controller or the sas expander?[21:23:51] <blues_> the rebranded lsi[21:23:56] <blues_> 2008[21:25:13] <dkeav> http://www.newegg.com/Product/Product.aspx?Item=N82E16816117171[21:26:03] <blues_> if i don't want to mount that extender in a pci-x slot.. does it have a special bracket or something i cna get ? Need another way to mount in case[21:26:28] <dkeav> you can use motherboard risers[21:26:40] <dkeav> or any pcb riser for that matter[21:27:04] <blues_> AHH ok i see... pix help[21:28:40] <dkeav> it has riser holes in the pcb so you can essentially put risers anywhere in a case and mount it[21:29:07] <dkeav> even get riser extensions that crew into risers with the motherboard between it to mount it over the top of a mobo[21:29:17] <DanaG> Interesting: the second boot was still slow, but didn't time out.[21:29:22] <DanaG> That's after disabling swap.[21:29:29] <DanaG> Odd that it says swap is /tmp...[21:29:36] <DanaG> Aren't those two separate things?[21:29:49] <DanaG> The line was something like.... swap tmpfs /tmp[21:30:10] *** ottom has quit IRC[21:32:31] <sergefonville> sorry, I was away...[21:32:36] <sergefonville> byt I'm back (H)[21:33:20] <dkeav> DanaG: swap -l will show where your swap is at[21:34:13] <sergefonville> use nwam :P[21:34:51] <sergefonville> I'd better go, I'm starting to babble...[21:34:59] <sergefonville> talk to you later ;-)[21:35:47] <DanaG> How do I get the login prompt not to do da^H and instead actually backspace?[21:35:55] *** sergefonville has left #openindiana[21:37:17] <DanaG> delete backspaces. How odd.[21:38:23] * rj mumbles about genuine sun keyboards[21:39:18] <DanaG> Oh, and home and end keys don't work with 'less'.[21:39:34] *** xmikus01 has joined #openindiana[21:42:01] <blues_> whew...ok got em ordered... with a little luck on the shipping i should be building this badboy this weekend[21:42:14] <blues_> thanks for your help dkeav[21:42:23] <DanaG> how do I disable that stupid pc speaker beeping?[21:43:06] <DanaG> On Linux, I'd unload pcspkr or snd_pcsp... what do I do in openindiana?[21:43:53] <dkeav> blues_: yep, good luck[21:44:17] <dkeav> DanaG: not make so many typos?[21:44:21] <dkeav> :D[21:46:33] <DanaG> Try tab-completing.[21:46:39] <DanaG> You get a beep every damn time you hit tab.[21:47:10] <Hedonista> i tried xset b off ,xset -b it didnt work either[21:47:31] <Hedonista> its loud on headphones[21:47:58] <rj> actually I have done this let me check[21:48:00] <DanaG> And I can't even modunload (id of pit_beep)![21:48:04] <DanaG> Device in use.[21:48:17] <rj> yeah you tell it to be ignored in /etc/system IIRC[21:48:29] <rj> (then reboot)[21:48:49] <tsoome> http://praveen.kumar.in/2009/04/01/turning-off-loud-system-beep-in-opensolaris-gnome-and-gdm/ ?:P[21:48:57] <tsoome> man google?:P[21:49:31] <DanaG> That doesn't help for console.[21:49:35] <dkeav> echo "set bell-style none" >> ~/.inputrc[21:50:12] <DanaG> How do you blacklist the whole module?[21:50:21] *** birvin has joined #openindiana[21:50:34] <tsoome> read /etc/system[21:50:42] <tsoome> it got examples....[21:51:29] <rj> ah "exclude: pit_beep"[21:51:37] <rj> that's what I shoved in /etc/system[21:51:56] <DanaG> Thanks. Google was just giving me results for Linux, and totally ignoring the word "solaris"![21:52:05] <DanaG> Damn google is really getting annoying like that.[21:52:26] <DanaG> +Now +I +have +to +put +a +plus +sign +in +front +of +every +damn +word.[21:52:38] <rj> gaaaah[21:52:45] <alanc> Welcome to Google+ ![21:53:35] <dkeav> its social![21:53:35] *** Norrin has joined #openindiana[21:53:45] <dkeav> and its not facebook!![21:53:52] <rj> hahaha[21:53:57] <DanaG> I don't care about the social; I just want it to search for what I search for, instead of whatever the hell it wants.[21:57:56] <DanaG> Anyway, now what can I do about my USB slowness?[21:58:40] <rj> all I have is a 'try not to do that' stick[21:59:16] <dkeav> send me the microserver i'll install it for ya[21:59:26] <dkeav> with my slow benchmarking kingston stick[22:02:07] <tsoome> some sticks just dont play nice in some systems....[22:05:16] <dkeav> god forbid we try a different one[22:06:10] <tomww> DanaG: only ideas, but you could plug the USB into another port (try all), eventually you get onto another usb hub chip then.[22:06:12] <DanaG> I'm running the same stick in a different system, and it's also slow.[22:06:25] <tomww> or disable all what can be disabled in the bios to get IRQs resorted.[22:06:29] <DanaG> And I've run Ubuntu on both, with the same USB stick, and it's fine.[22:08:51] <dkeav> isn't that just more evidence that you should try a different usb stick[22:09:49] *** raichoo has quit IRC[22:10:49] <DanaG> I don't have any other flash drives big enough.[22:11:02] <DanaG> And only OpenIndiana has issues.[22:11:29] <dkeav> you can't like borrow one from someone for a few minutes?[22:11:37] <dkeav> its not like it takes a very big one to do the job[22:11:50] *** dijenerate has joined #openindiana[22:13:03] <DanaG> It looks like I'm not the only one having USB boot be slow:[22:13:03] <DanaG> http://www.big-bubbles.fluff.org/blogs/bubbles/archives/000622.html[22:14:03] <dkeav> well we know that some sticks work fine, and we know that they work fine on the machine you are trying[22:15:58] <DanaG> Hmm, I'll have to ask.[22:17:13] <DanaG> hmm, as a worst-case test, I do have (though not with me at the moment) my iAudio 6. That thing has a 1.8" hard drive in it. If that's slower than this thing, I'll know something's definitely up with this stick.[22:17:59] *** dijenerate has quit IRC[22:20:13] <DanaG> Heck, I could try an old PATA hard drive on a USB converter.[22:26:51] *** mikw has quit IRC[22:35:48] *** bens1 has joined #openindiana[22:52:41] *** Coil-Domain has joined #openindiana[22:53:44] *** CoilDomain has quit IRC[22:55:03] *** Coil-Domain has quit IRC[22:56:08] *** CoilDomain has joined #openindiana[22:57:29] *** bens1 has quit IRC[22:58:46] *** Naresh` has joined #openindiana[23:00:07] *** Naresh has quit IRC[23:13:01] <DanaG> Okay, so the PATA drive on USB converter seems about the same speed.[23:15:57] *** hsp has quit IRC[23:18:32] <DontKnwMuch> why would ping not work, I can ping from all machines, just not from OI, and I can not join the domain (probably something blocked), how can I check that[23:19:07] *** kimc has quit IRC[23:30:34] <infinity_> I have inheritance "fd" setup on the parent directory. When files are moved on the server, they don't take the directories permissions. They keep their original permissions. Is there a way to get around this and have the file/dir inherit its new parent directory permissions? This is on oi_148[23:46:49] <DontKnwMuch> anyone with AD joy(ning) experience. I am getting KPASSWD failed and Failed to establish NETLOGON credential chain... what could this be?[23:54:54] <DanaG> So, how do I make this flash drive not be so slow with OpenIndiana? Ubuntu + flash drive == fine. OpenIndiana + flash drive == slow as molasses. OpenIndiana + PATA drive on USB adapter == fine.[23:56:38] <DanaG> Interesting... removing graphics console improves it a bit.[23:59:41] *** skeeziks has quit IRC