July 30, 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 30, 2011 [00:00:15] *** CoilDomain has quit IRC[00:16:23] <az_ordog_maga> anybody using oi with tohshiba nb200?[00:17:45] *** Alasdairrr is now known as AlasAway[00:18:06] *** Botanic_ has quit IRC[00:20:08] *** Worsoe has quit IRC[00:21:30] <n2deep> phretor_ last time I checked the si3124 driver was in really bad shape, and needed a lot of work.[00:22:36] <phretor_> n2deep: and so why is it listed as supported?[00:23:46] <phretor_> it would be the second PCIe SATA board that I buy after checking the HCL, and which turns out unsupported. Is there any explanation about that?[00:24:01] <n2deep> I really don't know. Last time I messed with it was in the winter. I kept having kernel panics on b147, whenever I had the card using taht driver connected.[00:24:37] <n2deep> I'm sorry you are having issues.[00:24:39] <phretor_> I bought a JMB363-based card because it was listed as supported, and it didn't work. Then I bought a SIL3132 because it was listed as supported, and now you tell me that it's not working.[00:24:50] <phretor_> no, seriously, who is in charge of updating the HCL?[00:24:52] *** Worsoe has joined #openindiana[00:26:14] *** phretor_ is now known as phretor[00:26:18] <n2deep> it's a community HCL?[00:27:18] <n2deep> the problem I was having at the time is anytime I would try to move a file more than ~2GB to storage using that driver, the machine would panic.[00:27:34] <phretor> n2deep: this is for OI_148[00:27:36] <phretor> http://wiki.openindiana.org/pages/viewpage.action?pageId=4885461[00:27:56] <phretor> n2deep: Silicon Image SiI3124/SiI3132 -> Supported: si3124[00:28:05] <phretor> no notes, no drawbacks reported.[00:28:47] <n2deep> so I gave some of the devs output from my log files and detailed accounts of my troubles.[00:29:04] <n2deep> it's possible that fixes went into b148?[00:29:23] <phretor> I have no idea. I am just a OI user.[00:29:38] <n2deep> .. and perhaps you are seeing another, seperate issue.[00:30:20] <phretor> n2deep: what I'm seeing is money blowed out thanks to that HCL, so far :)[00:32:13] <n2deep> I feel your pain, I too wasted money on this.[00:32:30] <phretor> btw, does anybody know whether I can boot from a root mirrored zpool?[00:32:50] <phretor> (via USB preferably, so I can re-sell my eSATA card I bought hoping it worked)[00:34:11] <n2deep> my rpool is mirrored on several machines, no issues for me.[00:34:44] <phretor> n2deep: but it'd be unusual if it's via USB too[00:34:52] <phretor> well, I'd be happy in that case.[00:35:07] <n2deep> hmm. via USB, I don't know about.[00:36:04] <phretor> it'd also involve the mother board in that case I gues[00:36:25] <n2deep> I ended up bying a motherboard with 6 sataII ports and 4 SAS ports[00:36:46] <phretor> n2deep: I've got a mini-itx scenario[00:37:12] <n2deep> oh, yeah that makes it harder.[00:38:18] <phretor> n2deep: I don't think anything with such many ports exists for mITX. I bought a mobo with 4SATAs and added a PCIe-to-2eSATA card.[00:39:05] <miine> phretor: supermicro has a 6 port sata atom mini-itx board. but expensive...[00:39:19] <phretor> miine: atom, does it play nice with OI?[00:39:25] *** Worsoe2 has joined #openindiana[00:39:46] <miine> phretor: sure. have one running on oi148 using since osol 08/10 ...[00:40:12] <phretor> miine: this guy here? http://www.supermicro.com/products/motherboard/ATOM/ICH9/X7SPA.cfm?typ=H[00:41:04] * cfs shudders at the thought of supermicro[00:41:15] <miine> phretor: that should be fine. the only thing about it I don't know if the two nics are correctly handled... but you should be able to disable one in case it is not...[00:41:21] *** dekar_ has joined #openindiana[00:41:45] <phretor> miine: do you by any chance remember the model of your mobo?[00:42:14] <miine> phretor: I have a intel one (D945CLF2 or so)[00:43:10] <phretor> miine: and that works fine with OI?[00:43:47] *** Worsoe has quit IRC[00:43:47] <phretor> miine: you said you had a supermicro one.[00:43:49] <miine> but I had to add a pci sata controller. using a SIl 3112 (or something like that). the trick for me was to use a Macintosh card so the bios isn't recognized...[00:44:11] <miine> phretor: I said that I have running oi on an atom ...[00:44:53] <phretor> I see if I can boot via USB and make rpool mirrored and then boot from those.[00:45:01] <miine> phretor: but if I had known the trouble of finding a working sata card - I've either bought the supermicro board or a uATX solution...[00:45:54] <miine> ok. the supermicro board was too expensive anyway. it was just that I fell in love with that 4 HD chembro mini-itx case...[00:46:05] <phretor> miine: I've been having troubles even getting the BIOS recognizing an HD attached to the on-board-eSATA port. So, I guess it's the cable.[00:46:10] <miine> as it is a real task to find good looking cases....[00:47:16] <phretor> When I loaded the si3132 driver in OI it didn't complain. It just said that the sata1/0 and sata1/1 ports (those eSATAs where I attached the wannabe rpool drives) were empty. So I guess it's really the cable.[00:48:32] <miine> phretor: maybe. depending on the board many things can be messed up in bios... (my favorite are those ram timings... don't know who needs that)[00:50:02] <miine> a good rule I learned using oi/osol is to disable everything in bios not needed and enable if later if the setup works...[00:50:30] <phretor> miine: for instance, now only one USB disk is seen by OI's installer.[00:51:01] <miine> what board are you using?[00:51:15] <phretor> Intel dg45fc[00:51:20] <phretor> ICH10R[00:51:58] <miine> look if they have a bios update...[00:52:12] <phretor> I just flashed it to the latest one.[00:52:22] <phretor> miine: should I disable Core Multiplexing Technology?[00:52:38] <miine> don't know. even what that IS...[00:52:59] <miine> I really like those bios settings :-)[00:53:03] <phretor> yeah[00:55:00] <miine> what happens if you set the drives from native sata to ide in bios? maybe its a ICH10R issue...[00:55:07] <az_ordog_maga> anybody using oi with tohshiba nb200?[00:56:01] <phretor> miine: they will appear as connected, but who knows whether it's the eSATA-to-SATA cable which is faulty or the eSATA port which isn't handled correctly?[00:56:13] <phretor> so, I'm sure the disks work fine.[00:56:53] <cfs> si3124 driver doesn't support port multipliers, fwiw[00:56:56] <phretor> now I've disabled most of the weird settings and rebooted[00:57:07] <cfs> or shouldn't[00:57:09] <miine> so get another cable or try to use the internal ports (I think that eSATA maybe needs a converter board...)[00:57:31] <phretor> cfs: I've got one on-board eSATA and two over-PCI-e eSATA ports.[00:57:46] <cfs> ah so no multipliers[00:58:22] <phretor> non of them work, that's why I'm guessing it's the cable. The PCIe card is sil3132 (via si3124 should be supported) and the on-board eSATA is via ICH10R[00:58:34] <phretor> s/non/none[00:58:36] <cfs> the onboard is the ahci driver[00:59:24] <phretor> yeah, but neither the BIOS nor OI's installer see any disk attached to it. So, again, it's either the port or the cable. cfgadm -l | grep sata sees the ports as empty and configured.[00:59:37] <cfs> what size are the drives you are trying to boot off of.[00:59:47] <cfs> rather, install on[01:00:04] <phretor> they're a couple of wd3200bpvt 320GB[01:00:10] <cfs> hrm.[01:00:46] <cfs> you should see what cfgadm says.[01:00:58] <cfs> (sorry if you did that, my scrollback is limited)[01:01:03] <phretor> the issue now it's not about booting, it's rather on getting them recognized via sil3132 and this would be the card I used http://www.rosewill.com/products/1131/productDetail.htm[01:01:21] <phretor> (sure, np)[01:01:51] <cfs> if you got a mac SiI3132 based card there is no BIOS, thus no love from the system[01:01:57] <cfs> of course you won't get to boot off of them.[01:02:11] <phretor> cfs: why not?[01:02:57] <miine> phretor: mac bios means that the pc bios doesn't know about the card and its connected drives...[01:03:02] <cfs> sorry, I was in 2004 for a moment.[01:03:26] <miine> phretor: but oi should find the card on the PCIe bus and attach the correct driver...[01:03:37] <phretor> miine: you mean at boot?[01:03:44] <phretor> (i.e., GRUB)[01:03:53] <cfs> are you getting to grub?[01:04:06] <miine> phretor: no. after the kernel is loaded. you;ll see that in the logs...[01:04:06] <phretor> well, yes, the installer loads grub[01:04:17] <cfs> no after you install[01:04:20] <cfs> can you boot into grub?[01:04:28] <cfs> from the HD?[01:04:29] <miine> phretor: nope. bios load grub. grub loads kernel. kernel loads drivers...[01:04:37] <phretor> oh I see, so, is there ultimately a way to "boot" off from a mirroed rpool?[01:04:44] <phretor> I didn't even install OI yet.[01:04:48] <cfs> ah[01:05:02] <miine> phretor: of course you can boot from a mirrored rpool. but not from raidz yet...[01:05:15] <cfs> installer won't let you install a raidz pool anyways.[01:05:23] <az_ordog_maga> guys youre better than csarli seen[01:05:24] <az_ordog_maga> ahahah[01:05:33] <az_ordog_maga> :)[01:05:40] <miine> phretor: as there is no raidz support in grub...[01:05:54] <phretor> I'm now in front of the installer (I temporarily connected the drives via USB: only one of them is listed in the installer's screen)[01:06:15] <cfs> that's fine, just install on one disk and attach the second one into the pool later.[01:06:18] <phretor> I don't really care about raidz for system pool. I only want it mirrored (i.e., RAID0 style)[01:06:39] <miine> phretor: and the other ones are missing? not good...[01:07:10] <phretor> I've got 4 2TB disks (now disconnected for testing) for data and 2 320GB (now over USB).[01:07:47] <miine> phretor: so the one 320 gb not showing up is bad (its case/usb-sata converter)[01:07:56] <phretor> only one over-usb, 320GB disk is listed[01:08:28] <miine> phretor: if you have attached TWO, there should be TWO. either its the disk, the cable or the board.[01:08:43] <phretor> well, actually, if I cfgadm -la | grep usb-storage, I see two of them. Now, I'm new to OI, so I don't think if there is a way to thoroughly test whether they're both recognized fine.[01:09:41] <miine> phretor: so it might be an installer issue. I suggest always to use the live (gui) cd and check via device driver utility if everything is fine...[01:09:45] <phretor> usb11/5 and usb11/6 usb-storage connected configured ok (from cfgadm -l)[01:10:16] <miine> phretor: so just go ahead and try :-)[01:10:38] <phretor> miine: OK, these days I've toasted a dozen of osol/oi/freebsd/whatever-nfs distro flavors. I don't minding burning another DVD.[01:10:57] *** DontKnwMuch has joined #openindiana[01:11:09] <phretor> also, I feel guilty because I run out of CD-Rs so I'm wasting DVD-Rs for burning CD-sized ISOs. Bad phretor, bad.[01:11:10] <miine> phretor: thats the right mindset :D[01:11:13] <DontKnwMuch> hi, can this be ignored? http://www.pastie.org/2292206[01:11:46] <phretor> DontKnwMuch: I'm tempted to click on it, but I bet you deliberately used that premise to invite us to click :) so I won't.[01:11:55] <phretor> sorry, nice try.[01:12:36] <DontKnwMuch> well I just did not want to paste more than three lines to the channel, that is ;)[01:13:05] <phretor> then you could have just pasted the link and added a nice screaming emoticon next to it. I like emoticons.[01:13:11] <miine> phretor: oi/osol/solaris is a beast at first. but the reward for handling it is also big...[01:14:11] <phretor> miine: I'm tempted at selling my Intel board and getting that 6-ports SuperMicro. But of course I'd prefer getting this working. I actually don't care about speed as these 2 additional SATA drives are just for the OS to boot and load.[01:15:02] *** phretor has quit IRC[01:15:30] <miine> phretor: look if that old Dell i3 T110 would serve you better. put in an intel e1000 card for 20 bucks and you can't beat them on value/money...[01:16:12] *** phretor has joined #openindiana[01:16:20] *** CVLTCMK0 has joined #openindiana[01:16:24] * phretor is back.[01:16:33] <miine> phretor: I repost it for you: look if that old Dell i3 T110 would serve you better. put in an intel e1000 card for 20 bucks and you can't beat them on value/money...[01:17:32] <phretor> miine: well it's quite expensive. I spent about $550 for my stuff.[01:17:37] <miine> phretor: they are selling at discount (maybe its over and they have only new model ones), and they work fine. even the power consumption is low. but with only one hd - the carriers and cables for up to 4 interal hds are included...[01:17:38] <phretor> http://configure.us.dell.com/dellstore/config.aspx?oc=becm42y&cs=04&dgvcode=ss&c=US&l=EN&m_3=4G4U3S&CAWELAID=582104003&dgc=SS&cid=27721&lid=628326[01:17:47] <phretor> oic[01:18:11] <miine> phretor: some days ago it was about 500 $ less...[01:18:48] <miine> phretor: so they sold all old ones...[01:18:50] <phretor> miine: this is my case http://linitx.com/product/12789 and this is my board http://www.intel.com/products/desktop/motherboards/DG45FC/DG45FC-overview.htm[01:19:10] <phretor> plus this PCIe 2-eSATA card I bought for the system drives http://www.rosewill.com/products/1131/productDetail.htm[01:19:55] <phretor> it's a pretty, nice-looking home NAS actually. So, I'd prefer to keep it this compact if I can manage getting it working :)[01:20:23] <phretor> but yeah, I agree with you that that Dell is worth a look.[01:21:35] <miine> phretor: in your case I would use 3 disks for storage and one smaller for the system...[01:21:47] <miine> phretor: so you've already a spare in case you need it.[01:22:23] <phretor> miine: I don't want to waste one 2TB disk for system. And I've bought these two nice 320GB 2.5'' disks; they fit nicely in the case.[01:22:49] <miine> phretor: just try first to use the INTERNAL ports. If you don't get them working reliable, you won't be able to use your setup at all...[01:23:56] <miine> phretor: THEN decide if you stick with the interla-ports only solution or try to get the pcie card working...[01:24:12] <miine> phretor: otherwise you have to many unknown variables to solve at once.[01:27:11] *** Botanic has joined #openindiana[01:27:11] *** Botanic has quit IRC[01:27:12] *** Botanic has joined #openindiana[01:29:32] <DontKnwMuch> connecting through putty to a remote (ipsec) oi, makes me wait 20 seconds or so after entering username/password. After that it is fast. it has something to do with DNS I beleive, did anyone exerience this too?[01:31:27] *** gnomad has quit IRC[01:33:48] *** laserbled has quit IRC[01:34:14] <phretor> miine: you mean use the internal ports to do some tests such as installing OI on a mirrored rpool made out of the two 320GB?[01:34:45] <miine> phretor: exactly. but try un-mirrored first ;-)[01:35:22] <miine> phretor: don't mind to take "baby steps"[01:36:14] <az_ordog_maga> first put to the one 320 gigs drive[01:36:44] <az_ordog_maga> when this is working env stick it other services and drives[01:37:07] <az_ordog_maga> my next step in this situation first install internal[01:37:30] <az_ordog_maga> then probaly this WORKING system on the external eSATA port[01:39:03] *** az_ordog_maga has quit IRC[01:43:41] *** InTheWings has quit IRC[01:53:10] <phretor> miine: yeah, I'll give it a shot with that.[01:53:37] <phretor> miine: so, internal SATA ports, one drive, unmirrored.[01:54:11] <phretor> miine: what kind of tests do you suggest to perform, just install it and see how it goes, or what else?[01:55:12] <miine> phretor: after you;ve got it installed and booting, put in three data disks and see if they work correctly and also your networking is fine etc.[01:56:07] <miine> phretor: as you might find out that setting up oi and its services might be tricky enough, I wouldn't wonder if you end up with using just the internal ports...[01:56:29] <miine> phretor: but that depends on what you're going to run on that machine.[01:56:34] <phretor> miine: I'd prefer to have a 4-disks RAIDZ[01:56:44] <phretor> but I agree with the baby-steps approach[01:56:50] <miine> phretor: i would prefer a STABLE machine :D[01:57:28] <phretor> also blowing money isn't satisfactory.[01:57:40] <phretor> but yeah, data has no price.[01:58:25] <miine> phretor: just had to ditch a pcie graphics card I didn't get working. I might get it working later, but for now I'm sticking with a working on-board intel graphics... (which is surprisingly fast)[01:58:36] <RoyK> if you data has no price, stripe them :P[01:59:07] <phretor> RoyK: that's what I'm planning to do. My goal would be a RAIDZ[01:59:40] <RoyK> just don't do like this one guy on the zfs list did - 35 drives in a RAIDz1[01:59:43] <RoyK> one VDEV[02:00:18] <phretor> RoyK: oh the guy who run OI/OSOL over VirtualBox running on windows?[02:00:25] *** chototsu has quit IRC[02:00:27] <RoyK> something like BASE jumping with an umbrella :P[02:00:34] <miine> phretor: you can also use RAID-Z with 3 disks. and then you have one spare "Ready". look at googles analysis: when one disk fails in a RAID chances are good that another fails in the near future (or even while resilvering)[02:00:46] <RoyK> phretor: no, I think this guy used hardware[02:01:00] <phretor> how many vdev should I use RoyK?[02:01:07] <RoyK> how many drives?[02:01:46] <phretor> miine: oh you mean keep one spare ready in my drawer w/o having to run to the shop to get a new one.[02:02:12] *** axisys has quit IRC[02:03:12] <phretor> RoyK: the goal would be to have 4 RAIDZ for data and a mirrored rpool made from 2 smaller disks. But I've been having trouble getting the two additional eSATA ports over PCIe working. So miine was suggesting using only those four ports and make a 3-disk RAIDZ. Unfortunately, in that case my rpool won't be mirrored anymore.[02:03:20] *** axisys has joined #openindiana[02:03:42] <miine> phretor: yep. thats a good idea - if you can live with 4 TB disk space..[02:03:48] <RoyK> phretor: how many data drives?[02:04:05] <phretor> RoyK: 4[02:04:28] <RoyK> then a single RAIDz1 VDEV should do well[02:04:33] <miine> phretor: you can copy the rpool onto the data disks and restore it later if needed. see zfs send | zfs receive ...[02:04:37] <RoyK> or striped mirrors if you want it quick[02:05:13] <phretor> also, I think I have to cope with 4K blocks issue also. The data drives are WD**EARS[02:05:35] <miine> phretor: don't mind. just try and test...[02:05:38] <RoyK> miine: that shouldn't be necessary - just backup the useful things to the data pool once in a while[02:06:05] * RoyK doesn't know how the rpool likes to get a zfs receive[02:06:40] <phretor> so, let's suppose something goes wrong. One data disk goes to heaven. I get my spare and do resilvering.[02:07:01] <miine> RoyK: doesn't need to as the restore is made booted from live medium (which has no rpool)[02:07:10] <RoyK> phretor: no problem unless you find bad data on another drive during resilver[02:07:10] <phretor> so, let's now suppose that the rpool disk goes to heaven (not at the same time, just for the sake of simplicity). What do I do?[02:07:38] <phretor> RoyK: yeah but if I have a spare ready...[02:07:40] <phretor> or not?[02:07:59] <miine> phretor: change rpool disk. boot from live cd. do a zfs send | zfs receive onto the new disks. then do a install grub (like you would do when mirroring)[02:08:08] <miine> phretor: ready you are...[02:08:12] <RoyK> phretor: if you, during resilver, find bad data on another drive, there'll be data corruption[02:08:30] <phretor> RoyK: and what do I do in that case? Besides sucicide I mean.[02:08:45] <RoyK> just restore the bad files from backup[02:08:56] <phretor> RoyK: wait, which backup?[02:09:11] <RoyK> heh[02:09:23] <phretor> we haven't spoken about backups so far.[02:09:54] <RoyK> phretor: RAIDz1 is good so long as n-1 drives are good. if you lose a drive and you find bad sectors on another, you lose data[02:10:05] <miine> phretor: get yourself familar with zones. you can run any non-kernel services from them. their data could completly reside on the raid-z . so the need for a mirrored rpool is not that big...[02:10:06] <RoyK> phretor: that's why I always use RAIDz2[02:10:47] <phretor> but in that case I'd end up with as low as 2TB of daa[02:10:48] <phretor> data[02:11:01] <RoyK> less but safer, yes[02:11:04] <phretor> what do I do with the other 6TB of disks lying around my desk?[02:11:04] <miine> phretor: its as always: trading space vs. security.[02:11:17] <RoyK> (or safety)[02:11:27] <phretor> or I could sell them and buy 2 3TB drives.[02:12:07] <RoyK> phretor: you can replace them with 2TB drives when they grow full[02:12:13] <RoyK> one by one[02:12:29] <phretor> I don't follow.[02:12:31] <RoyK> given autoexpand=on, the pool will grow once all drives are replaced[02:12:52] <RoyK> phretor: you have 4 1TB drives now, right?[02:13:09] <phretor> no, 4 2TB[02:13:27] <RoyK> so with a RAIDz2, that'll be 4TB[02:13:52] <phretor> RoyK: if I could get the damn system disk recognized by at least one of the damn eSATA ports :)[02:14:03] <phretor> because I only have a 4-bay case :)[02:14:11] *** CoilDomain has joined #openindiana[02:14:20] <RoyK> phretor: use some USB plugs for the root[02:14:30] <RoyK> the rpool is hardly accessed anyway[02:15:29] <phretor> RoyK: I do have them: I used two 320GB spare disks. They're both recognized by the BIOS by OI's installer (non GUI) only sees one of them. Now I'm going with the GUI installer and see what happens.[02:15:57] <phretor> I would be more than happy to have very safe data, of course, I use it to store pictures.[02:16:05] <phretor> and pictures are priceless.[02:16:11] <RoyK> the only time the rpool gets excessive traffic is either a process going bad and starts logging or if you run out of memory and the swapper starts hichuping[02:17:05] <RoyK> phretor: are those SATA or ATA drives?[02:17:12] <phretor> SATA[02:17:17] <phretor> WD20EARS[02:17:39] <RoyK> you want to use 2TB drives for the rpool????[02:18:10] <phretor> RoyK: ah no, sorry. Those are WD3200BPVT 320GB SATA[02:18:12] <phretor> 2.5''[02:18:22] <RoyK> k[02:18:29] <RoyK> I have some of those[02:18:33] <RoyK> works well[02:18:46] <phretor> yeah, the only thing is the cable, I guess.[02:19:00] <DontKnwMuch> can any of you just for a reference please run powertop and tell me how many wakeups/s do you have if you systems are more or less idle please :)[02:19:09] <phretor> because neither the on-board eSATA nor the PCIe eSATAs ports work. But if I connect the same disks via SATA they work.[02:20:10] <RoyK> phretor: most SATA boards with eSATA have jumpers - external or internal - may that be it?[02:22:41] *** redgone has joined #openindiana[02:22:53] <phretor> RoyK: mh, I have to check about the board (Intel DG45FC) but for sure this guy here hasn't http://www.rosewill.com/products/1131/productDetail.htm[02:22:57] <phretor> any jumper[02:23:02] <phretor> at least that I know about[02:24:02] <phretor> here is the specs http://www.rosewill.com/Mgnt/Uploads2/AttachmentForProduct/UserManual%20for%20RC-219%20v1.0.pdf[02:31:32] <RoyK> plain eSATA boards shouldn't need jumpers[02:31:44] <RoyK> but then, dunno if that chipset is supported in OI[02:39:50] <phretor> RoyK: it is, via si3124 on OI_148 http://wiki.openindiana.org/pages/viewpage.action?pageId=4885461[02:39:59] <phretor> RoyK: it indeed sees the two extra sata ports[02:40:12] <phretor> sata1/0 and sata1/1 via cfgadm -l | grep sata1[02:42:44] <miine> have to say that my puppet zone provider is going to look really good: if zone creation fails it will roll back automatically to it's original state (even if that means deleting it...) :D[02:43:39] <phretor> how should I partition the system disk?[02:43:54] <phretor> actually, the installer doesn't let me partition it.[02:43:55] <miine> I'm going to have a bootstrap parameter: a path to a directory which is copied into the zone and then a 'bootstrap' binary is called inside the zone...[02:44:33] <miine> first class oi/solaris zone support in puppet...[02:44:52] <miine> can't wait to finish it and release it.[02:45:48] <phretor> miine: I can't wait to finish installing OI on the 320GB drive and see how it goes.[02:46:31] <miine> phretor: tell us how the pcie works. its always good to know which hw works or not...[02:46:42] *** elgar has joined #openindiana[02:46:46] <miine> phretor: later of course...[02:46:48] <phretor> miine: I have to buy another cable.[02:47:11] <phretor> it's 2:45am here :)[02:47:21] <miine> phretor: here too :D[02:48:26] <phretor> where do you live miine[02:48:34] <miine> germany[02:50:11] <phretor> or maybe I can cut the eSATA port away and solder a working SATA cable directly on the PCIe[02:51:10] <miine> phretor: if you have an anti-static workplace/tools and fun with such thing...[02:51:36] <phretor> I don't have an anti-static workplace unluckily[02:52:10] <miine> I just tried to clip of a pin of a 3114 SiL chip to disable the external bios which could not be flashed to non-raid...[02:52:23] <miine> didn't work. but it SHOULD have. :D[02:52:33] <phretor> I can just break a SATA connector and make it fit into the eSATA; they are really close. I don't see the point of making the eSATA when it's exactly identical to the SATA.[02:54:07] <miine> phretor: why not drill a hole into the card's shield and put the eSATA cable thru it? so no soldering needed...[02:55:14] <phretor> well, the physical space is not an issue. The card comes with a lower profile and lower back shield so there is already a hole there thru which the I put the cables.[02:55:31] <miine> phretor: looks not so good - but who cares? at least its easier to do without anti static equipment because the shild could be taken away from the card for drilling. BTW: always ground yourself when doing stuff with the mobo etc.[02:56:05] <miine> phretor: so stick with that. just buy new cables. always take the easy and safe route...[02:56:12] <phretor> what I wanted to do, instead of waiting until monday for some stores to open, was to stick a SATA male into an eSATA female by modifying the male such that it fits.[02:56:39] <miine> waiting was invented by devil :D[02:56:44] <phretor> eSATA is really pointless.[02:56:48] <phretor> true[02:57:33] <miine> but it's there to resist. just learn the oi basics like zfs, zones etc. . you'll be amazed how fast it is monday :-)[02:58:58] <phretor> miine: would I really need zones for a soho nas?[02:59:12] <dkeav> no[02:59:19] <miine> why not?[02:59:25] <phretor> just to give you an idea of what is my goal, I wanted to install NexentaStor in principle.[03:00:04] <phretor> so, I really need something that simple to manage. Then I saw a little bit of performance comparison and decided to switch for OI + Napp-it, but just for the sake of speed, not b/c I'm masochist.[03:00:11] <dkeav> why *would* you need zones just for a nas??[03:00:13] <miine> depending on what services your nas will provide besides filesharing (kernel based cifs/nfs) zones might be a good idea.[03:00:42] <phretor> I indeed need it for TimeMachine shares, CIFS, AFS and maybe some music/movie sharing services.[03:00:44] *** yourself22 has joined #openindiana[03:00:48] <phretor> ah, and torrent.[03:00:54] <phretor> torrent client: I need that as well.[03:00:59] <miine> it might also be a good idea NOT to use kernel based cifs but samba within one or two zones. depending if you can connect to the shares from outside...[03:01:44] <phretor> no, I can connect to the box via SSH only.[03:01:48] <phretor> remotely, I mean.[03:02:10] <phretor> within my home network, I will use CIFS/AFP/TimeMachine/iTunes kind of stuff.[03:02:11] <miine> phretor: ok. USE zones. its better security wise. you don't know if a virus etc. might try to hack your NAS...[03:02:38] <dkeav> o.0[03:02:40] <phretor> zones remind to me of jails[03:02:48] <dkeav> they should[03:02:50] <miine> and as zones are easy to setup: netatalk doesn't mind if it runs in the global or another zone...[03:02:54] <phretor> I think, however, that it's quite an overkill.[03:03:21] <miine> phretor: nope. on my atom 330 I have >5 zones running...[03:03:50] <miine> zones have "near no" overhead.[03:03:53] <phretor> not a performance overkill, but a setup/management overkill.[03:03:55] *** yourself22 has quit IRC[03:04:03] <bdha> phretor: Zones are a huge management win.[03:04:21] <bdha> phretor: Especially if you automate them with Puppet/Chef/whatever.[03:04:52] <bdha> My previus job had 12 Solaris boxes and ~150 zones. Very easy to see:[03:04:56] <bdha> 1) What a system is actually doing[03:05:16] <bdha> 2) What service (which will be a collection of many processes) is consuming, in terms of resources[03:05:31] <miine> phretor: the managment overhead is also near none: you;ve to setup netatalk etc. anyway...[03:05:34] <bdha> When you abstract services to little boxes, it makes them easier to manage, migrate, etc...[03:06:07] <bdha> Not to mention that you are segregating services that don't need to talk to each other, which is at least a small security win.[03:06:16] <dkeav> while i agree with everything you said there[03:06:20] <dkeav> for a simple nas??[03:06:22] <miine> bdha: I think its a BIG security win.[03:06:23] <dkeav> really[03:06:32] <miine> dkeav: sure[03:06:37] <bdha> dkeav: Yeah, I wouldn't bother for a home NAS. :)[03:06:46] <dkeav> kinda what i was thinking[03:06:47] <bdha> Single-purposes systems I run in the gz.[03:06:53] <dkeav> aye[03:07:04] <miine> why not use freebsd or linux then?[03:07:09] <bdha> Not many single purpose systems in production, though. :)[03:07:18] <phretor> so, from my uderstanding, I should be using zones to jail service processes and then exploit zfs send | receive to backup the system pool onto the raidz or raidz2 data pool.[03:07:32] <bdha> miine: "You aren't using the CD player in this Masserati, why not drive a FIAT."[03:07:35] <bdha> wtf :P[03:07:37] <miine> come on - its make NO difference to setup netatalk in its own zone or in the global zone.[03:07:46] <bdha> If it makes no different, why do it?[03:07:50] <bdha> difference[03:07:55] <phretor> well, these are two separate things, but I bet they're quite good practices.[03:08:03] <phretor> BTW, what does Chef/Puppet does?[03:08:13] <bdha> phretor: They are configuration management tools.[03:08:15] <miine> bdha: difference like: additional work.[03:08:29] <bdha> miine: There is a cost/benefit to everything.[03:08:33] <bdha> Learning is a huge benefit.[03:08:51] <phretor> http://projects.puppetlabs.com/projects/1/wiki/Puppet_Solaris ?[03:09:03] <bdha> Aye, that's Puppet.[03:09:19] <bdha> Wow, that's a long document.[03:09:29] <bdha> haha[03:09:30] <miine> phretor: forget about puppet. FIRST learn to setup them by "hand" and how they work. puppet is a complete beast on its own...[03:09:30] <bdha> 23) That is it.[03:09:50] <phretor> OI_148 finished its installation.[03:09:52] <phretor> reboot[03:09:56] <miine> phretor: also you've to learn what services are, packages etc. etc.[03:10:08] <phretor> well I come from a linux/bsd background.[03:10:16] <bdha> zonecfg -z foo -b create ; set zonepath=/zones/foo ; add net ; set physical=igb0 ; end ; set autoboot=true ; verify ; commit ; exit[03:10:17] <miine> phretor: maybe even how to compile and package yourself...[03:10:24] <phretor> the lingo here is very very diffrent.[03:10:25] <bdha> zoneadm -z foo install && zoneadm -z foo boot[03:10:27] <bdha> welp.[03:10:27] <bdha> :P[03:10:38] <bdha> Also, prstat -Z for the win.[03:11:30] <dkeav> phretor: not really[03:11:43] <miine> bdha: :D[03:12:04] <dkeav> most unix principles apply to other unices, yea the projects may not call things the same name, but the concepts still apply[03:12:22] <bdha> If you understand the primitives, the rest is gravy.[03:12:26] <dkeav> as a 70% bsd 25% solaris/oi user[03:12:27] <bdha> "I know this! This is UNIX!"[03:12:35] <dkeav> heh[03:13:02] <phretor> bdha: no, that was a quite hard way to browse a filesystem :)[03:13:07] <dkeav> bdha: fsn![03:13:15] <miine> oh BTW: would it be sufficient to add the zones fqdn/ip into it's /etc/hosts so that sysidcfg gets fooled (meaning no dns entry needed elsewhere)?[03:13:46] <dkeav> hah i used fsn on irix 6.3/6.5 when i was doing molecular modeling[03:13:56] <dkeav> it was more annoying than anything else[03:14:02] <phretor> dkeav: I was about to ask[03:14:35] <dkeav> it was more of a gimmic thing than actually meant for real use[03:15:02] *** Botanic_ has joined #openindiana[03:15:04] *** Botanic has quit IRC[03:15:31] <dkeav> http://fsv.sourceforge.net/[03:15:33] <dkeav> :D[03:15:48] <phretor> I went to vizsec2011 and the closing keynote speaker commenced his rant-talk with that scene from the movie; the whole talk was about the lack of really useful visualization systems today.[03:16:31] <dkeav> its just very difficult to make generic vis systems[03:16:33] * phretor is logged into this solaris box[03:16:53] <phretor> he concluded his talk with that MRTG is the best vis system by far[03:17:08] <dkeav> purpose built interfaces for certain applications that don't change much, sure, but not just a real generic system[03:18:07] <dkeav> phretor: ehh, rrdtool is better for that stuff[03:19:05] *** master_of_master has quit IRC[03:19:45] <dkeav> which mrtg and rrdtool are good examples of generic vis frameworks, but you still have to set them up for whatever your actual task is[03:20:11] <phretor> true[03:21:02] *** master_of_master has joined #openindiana[03:21:33] <phretor> ok, now that I have OI installed, I am going to proceed and make a RAIDZ2 off of 2 of the 3 big disks.[03:21:42] <phretor> I'm following this guide http://www.solarisinternals.com/wiki/index.php/ZFS_Best_Practices_Guide[03:27:50] *** DerSaidin has quit IRC[03:28:18] *** DerSaidin has joined #openindiana[03:28:18] *** DerSaidin has joined #openindiana[03:28:21] <phretor> miine: so, were you suggesting RAIDZ2 or mirroring?[03:28:35] <phretor> because I'm reading about them right now.[03:29:56] <miine> good question. three-way mirror should be faster. AND it should be no problem to detach a disk for an backup...[03:30:14] <phretor> miine: Z3 != 3-way mirroring?[03:30:40] <dkeav> no[03:30:48] <phretor> ah ok, so they're the same thing?[03:31:12] <miine> RAID-Z2 means that you could loose two disks. same on a three way mirror. Asuming that you'll use three disks in total.[03:31:46] <dkeav> guess it depends on how many discs you intend to use[03:31:48] <miine> but with RAID-Z2 the data is stored a bit different and possible more secure.[03:31:55] <dkeav> if just 3 discs, then yea just do a 3 way mirror[03:32:04] <dkeav> you only get the capacity of one disk though[03:32:39] <phretor> If I manage to have them all connected, I can get up to 2 equally-sized small disks + 4 equally-sized large data disks.[03:32:55] <dkeav> nice thing about mirrors is you can alway split a drive off it and add a 4th disk to create an array of mirrors[03:33:12] <miine> but I'm running here a RAID-Z1 with 4 disks (only 600 GB drives)... so it really depends on the type of data you want to store.[03:33:40] <dkeav> aye i'm running two raidz1's in a mirror[03:33:42] <miine> as for real secure storage also a UPS and (even more important) ECC-RAM is needed...[03:33:54] <phretor> it's movies, files, backups of other machines, pictures.[03:34:00] <phretor> no, I can't afford ECC now.[03:34:08] <phretor> I can afford a UPS yes.[03:34:17] <phretor> and I'd definitely buy one soon.[03:34:28] <miine> phretor: so even a three disk RAID-Z1 would be fine in your case.[03:34:37] <dkeav> yea[03:34:47] <miine> phretor: it justs depends on the space needed.[03:34:50] <phretor> miine: you mean the three + spare[03:34:51] <dkeav> raidz1 would be fine, your not doing anything "mission critical"[03:34:57] *** DrLou has quit IRC[03:35:20] <dkeav> if your confident you can replace a failed drive quickly, then you can skip the spare even[03:35:24] <dkeav> and do a 4disk array[03:35:32] <miine> phretor: yep. and you've to decide if you want ONE or TWO disk's space usable...[03:35:48] <dkeav> but i would suggest at some point maybe getting an external usb drive or something with a large drive or drives as a backup device[03:36:13] <phretor> well if I loose movies then who cares, but I don't want to lose backups. Actually, if I loose the backup, there is the original data in the client computer.[03:36:34] <phretor> dkeav: here I am missing something. What do you mean with "backup" here? Should I backup snapshots there?[03:36:35] <miine> the problem is: when a lightning strikes your home - all machines connected to anything will be gone...[03:36:54] <phretor> no way, I'll have a UPS for that :)[03:37:20] <dkeav> phretor: yea an external device that can mirror your live array too for offline storage[03:37:31] <dkeav> like a weekly backup or the daily or whatever[03:37:33] <dkeav> but a backup[03:37:48] <phretor> dkeav: you mean zfs-level backup, i.e., snapshots, right?[03:37:48] <dkeav> your nas is not a backup, its just network storage[03:38:00] <phretor> but it's backed up in some sense.[03:38:02] <miine> phretor: sure? If the data is that important: do a two way mirror and backup to an eSATA/USB two-way mirror. Then disconnect those drives and store them safely...[03:38:50] <phretor> miine: if I'm a home user like you are, then I'm probably OK with RAIDZ1 and some offline backup every month.[03:39:00] <dkeav> phretor: yea you can store snapshots, you can mirror, you can do actual file copies to a different filesystem, options are pretty open to whatever you want[03:39:09] <miine> some days ago I had to setup my system because the non-ECC RAM had two bit error resulting in ZFS mount errors. not that funny...[03:39:29] <dkeav> no[03:39:33] <phretor> miine: and how did you deal with that?[03:39:36] <dkeav> i get ecc mobo's whenever i can[03:39:44] <dkeav> ecc memory is pretty cheap these days[03:40:01] <miine> phretor: reinstall :-( . data was backuped on a different machine :-)[03:40:02] <phretor> dkeav: mITX ECC mobos are not that easy to find though.[03:40:07] <dkeav> i built a hp microserver a few weeks ago for my garage, and shoved 8gb of ecc in it for less than 100 bucks iirc[03:40:10] <phretor> miine: backed up?[03:40:26] <phretor> dkeav: how many disks you have and how large is the backup?[03:40:34] <dkeav> on that machine?[03:40:45] <dkeav> i have 4 1.5Tb discs in raidz1[03:40:47] <phretor> dkeav: sorry, that was for miine[03:40:48] <miine> phretor: mostly. it was my build machine. so it's really important data (sources) where in svn on my nas...[03:40:59] <dkeav> but this is for my home movie storage[03:41:07] <phretor> and backup[03:41:20] <dkeav> at work we have a much larger setup and its mirrored to a identical vdev[03:41:27] <phretor> I have four computers at home and want to backup them[03:41:27] *** CoilDomain has quit IRC[03:41:32] <dkeav> but thats using sas expanders and what not[03:41:58] <miine> the really important data isn't that much. maybe 80 gigs (email, sources, music etc.) ...[03:42:12] <miine> the rest is "just stuff".[03:42:55] <phretor> I agree, my picture folder is around 150GB[03:43:33] <phretor> so, the whole backup of this one important machine is 750GB, but only 200GB is very important.[03:43:36] <miine> but in your case: when you have already 4 x 2TB drives, it might be a good idea to use some of then for an offline backup...[03:44:12] <phretor> miine: offline you mean, physically outside the box or simply on a weekly/monthly basis?[03:45:03] <miine> phretor: "offline" i mean: connect, backup, disconnect and put it somewhere... I wouldn't trust that ups as an insurance against lightning etc.[03:45:28] <miine> it is even more likely that you might delete yourself something important away...[03:45:35] <phretor> I can always backup over network[03:45:47] <phretor> but I don't know if that's efficient.[03:46:06] <phretor> I'm telling this because my NAS, once placed, won't be easy to reach.[03:46:14] <phretor> Yes, maybe on a monthly basis I can do it.[03:46:33] <dkeav> over the network is fine if you have a reasonably fast network[03:46:44] <phretor> dkeav: I've got a gigabit network.[03:47:07] <dkeav> then it would be fine, especially incremental backups with say rsync[03:47:26] <dkeav> i would schedule that for offpeak network times[03:47:31] <dkeav> eg the wee hours[03:47:41] <miine> phretor: just run oi an a mac in vbox and connect the backup disks via usb (or something like that). zfs send | zfs receive the differences via snapshot. (if it is that hard to go to the nas and connect the backup disk)[03:48:09] <phretor> now I'm going to bed. Just to conclude: is there any post-installation checks/tests that I need to do in order to ensure that everything is working fine? At the moment I've installed OI on a separate 320GB drive.[03:48:32] <phretor> and I haven't done anything further.[03:48:49] <miine> phretor: or maybe buy TenComplement's ZFS. MacZFS has only zpool version 8 ...[03:49:17] <miine> phretor: just install everything (netatalk etc.) and do a real world test...[03:49:39] <phretor> miine: so I'd have to make the zpool either raid or whatever[03:49:52] <phretor> I mean raidzK or K-way mirror[03:49:57] <miine> phretor: yep. and transfer some data via network onto it...[03:50:44] <phretor> K-way mirror should be faster, simple and therefore more solid I guess. But even with a 3-way mirror I'd get a 2TB drive. Right?[03:50:45] <miine> phretor: to test the disks a simple cat /dev/zero > might be ok... will need some time to fill them up :-)[03:51:13] <miine> phretor: 3 drives, 3 way mirror = only one drive usable space[03:51:29] <miine> phretor: same with RAID-Z2.[03:55:50] <phretor> it's really hard to choose; I'll sleep on it.[03:56:15] <phretor> well, thank you very much for chatting about this guys. It was a real pleasure.[03:57:13] <dkeav> g' night[04:12:40] *** baitisj has joined #openindiana[04:15:06] *** mnaser_ has joined #openindiana[04:15:27] *** odyi has quit IRC[04:15:27] *** dkeav has quit IRC[04:16:12] *** gea__ has joined #openindiana[04:16:41] *** mnaser has quit IRC[04:16:41] *** StyleWarz has quit IRC[04:16:42] *** StyleWarz has joined #openindiana[04:16:59] <baitisj> Wow. Between Virtualbox 4.1.0 USB driver, and the nvidia driver, my machine is very unstable.[04:17:15] *** gea has quit IRC[04:17:29] *** gea__ is now known as gea[04:17:32] *** trn has quit IRC[04:18:00] <miine> baitisj: vb's 4.1 snapshots won't work for me (oi 148)[04:18:28] <miine> baitisj: think that they are testing on sol11xp first (only?)[04:18:33] <baitisj> miine, did you experience some corruption of your state file?[04:18:42] *** trn has joined #openindiana[04:18:45] *** odyi has joined #openindiana[04:18:45] *** dkeav has joined #openindiana[04:19:03] <miine> baitisj: yep. had to delete/create a vm because of that...[04:19:13] <baitisj> miine, I had to manually edit my machine's XML spec file after I upgraded from 4.0.x in order to get 4.1.0 snapshots to work[04:20:01] *** phretor has quit IRC[04:20:17] <miine> baitisj: interesting. maybe it was because it was an existing 4.0 vm...[04:20:18] <baitisj> I noticed that the most recent SUNWcs and friends from OpenIndianaDev don't seem to play well with VBox AT ALL[04:20:57] <baitisj> miine, same thing happened here, too. I also found that exporting an appliance and re-importing it seemed to clean up a lot of problems at some point...[04:21:14] *** hunter has joined #openindiana[04:21:45] <miine> vbox 4.1 has a new crossbow compatible network driver. but only for sol11xp... hope that we will get that too (soon)[04:21:59] <baitisj> I just rolled back the latest OpenIndianaDev stuff. We'll see if the machine state improves.[04:22:43] *** hunter has quit IRC[04:22:45] <baitisj> yeah, that sounds nifty. My latest bitch about OI network adapter stuff is that ipadm can't seem to modify persistent settings correctly.[04:23:02] <baitisj> i.e. ipadm set-ifprop -p seems completely broken[04:23:12] <miine> I even didn't use usb 2.0 after experiencing that vm corruption (don't need it).[04:23:49] <miine> my problem with vbox and network was, that oi as client sometimes can't create zones with exclusive nics ...[04:24:13] <miine> depends on host machine and vbox versions.[04:24:43] <miine> don't know what they have done that it is not that "virtual"...[04:24:44] <baitisj> miine, I actually found that the USB 2.0 support was very useful in creating a print server for a Canon printer/fax/copier that has no opensource scanning driver support. it seemed to work well until I updated OIdev a couple days ago...[04:25:19] <miine> baitisj: I consider using OIdev experimental anyway :-)[04:25:33] <baitisj> it is experimental by definition :-)[04:25:40] <baitisj> that's why I'm not really that upset...[04:26:18] <baitisj> i'm actually surprised at OI-151 stability. previously, the only thing that would crash frequently was the x server[04:27:08] <baitisj> miine, I don't know if this helps, but have you tried creating a vlan within a zone, and then creating vnics in bridged mode in virtualbox?[04:28:23] <baitisj> i found that creating one vnic and setting it as the router/default gw for all the other vnics that virtualbox uses seemed to work particularly well. but i don't know very much about networking within non-global zones.[04:29:00] <baitisj> (i run all my vbox nics in bridged mode)[04:29:47] <miine> most times I use a brigded nic on host. but I still don't get it how that setting would affect the client...[04:30:28] <miine> would like to have a real crossbow bridge between oi host and oi client in vbox ...[04:30:39] <baitisj> yeah, it would solve a lot of problems.[04:31:10] <miine> maybe its exactly that what they have done for sol11xp ??[04:31:26] <baitisj> the reason why I found the bridged host nic problematic was because of the issues with mac addresses.[04:31:45] <miine> why? duplicates?[04:31:57] <baitisj> correct[04:33:45] <baitisj> bridged mode necessitated setting the mac address of the vnic and the vbox interface to the same mac address, so when it was time for incoming packets to access my vbox instance, they would get lost.[04:34:01] <miine> what I also miss in vbox is to be able to boot from an usb stick... (while I'm writing it: I should try if it is possible to use that .usb as an image...)[04:34:42] <miine> but anyway: vbox is great for it's price point :-)[04:34:50] <baitisj> indeed. no complaints here![04:34:50] <dkeav> exactly[04:35:05] <dkeav> can't argue with free[04:51:21] *** gea_ has quit IRC[05:30:21] *** robinsmidsrod has quit IRC[05:36:53] *** robinsmidsrod has joined #openindiana[06:03:01] *** ianh1 has quit IRC[06:19:16] *** redgone has quit IRC[06:28:03] <miine> anyone here having experiences with ruby and pseudo-terminals on oi?[06:28:28] *** flyz has joined #openindiana[06:31:38] *** flyz_ has quit IRC[06:39:48] *** CVLTCMK0 has quit IRC[06:42:37] <richlowe> bdha[06:42:40] <richlowe> if bad experiences count.[06:44:05] <miine> hmm. have to find either a working ruby version or a ruby only workaround. but maybe oi148 stock ruby is flawed...[06:44:51] <miine> as it doesn't get all lines of output when executing another tool which does pty (from puppet).[06:45:07] *** CVLTCMK0 has joined #openindiana[06:47:14] <miine> is there a tool on solaris which calls another tool opening an pty before and piping that back?[06:47:55] <alanc> I don't understand what you're asking for, but "expect" sounds like a likely answer[06:49:37] <miine> alanc: is that within a standard installation? or does it need an additional pkg to be installed?[06:49:58] <alanc> it's in the package repository[06:50:13] <alanc> "standard installation" varies so much as to be mostly meaningless[06:50:43] <alanc> GNOME & X are in the "standard installation" if you use the LiveCD, but not the text installer or auto-install[06:50:59] <miine> alanc: thanks. didn't think of that "classic" tool. will try it. having another dependency within puppet is better then a not-so-reliable solution...[06:51:22] <alanc> (plus, I really don't know if it's in any of the install sets, since I've not needed to know)[06:52:26] <miine> or maybe I should check first if zlogin -C needs a pty at all...[06:55:45] <richlowe> If you're trying to drive sysidcfg with expect, or something else, you really don't want to[06:56:38] <richlowe> and sysidcfg is the only reason I can think of you'd need the zone _console_[06:56:43] <miine> richlowe: no no. I just need to check for "[Connected" and "console login:". After the last one sysidcfg is ready.[06:56:57] <miine> ready -> done.[06:57:35] <miine> if it hangs (no dns entry etc.) it should time out and my puppet provider will roll back the zone to its initial state...[06:58:34] <richlowe> Suspect a better way would be to wait for a given zone process.[06:58:53] *** kart_ has joined #openindiana[06:58:59] <richlowe> perhaps the console ttymon?[06:59:13] <miine> richlowe: thats a good idea. or checking if /etc/.sysIDtool.cfg has everything set to '1'...[07:02:09] <miine> console ttymon? its only running when sysidcfg did finish. that would be to cool...[07:03:31] <miine> but thats a good idea to go away from that zlogin / expect stuff.[07:06:06] <miine> my new puppet zone provider now has also bootstrapping support. just a bootstrap parameter giving a path to a directory which is copied to the zone's tmp and then a 'bootstrap' executable is called within the zone. so everyone could put into that directory / bootstrap file what they want.[07:06:48] <miine> there is also a always_bootstrap. if set to true that procedure is done on every boot, otherwise only if the zone didn't finish it's sysidcfg ...[07:10:10] *** Thrae has joined #openindiana[07:32:35] *** elgar has quit IRC[07:37:13] *** elgar has joined #openindiana[07:38:33] *** elgar has quit IRC[08:22:59] *** McBofh has quit IRC[08:31:39] *** McBofh has joined #openindiana[08:44:02] *** chototsu has joined #openindiana[08:44:02] *** ChanServ sets mode: +v chototsu[09:11:28] *** gea_ has joined #openindiana[09:26:43] *** chototsu has quit IRC[09:29:30] *** gea_ has quit IRC[09:48:51] *** gea_ has joined #openindiana[10:00:13] *** flyz has quit IRC[10:00:26] *** flyz has joined #openindiana[10:38:21] *** az_ordog_maga has joined #openindiana[10:38:27] <az_ordog_maga> aloha[10:41:39] *** Botanic__ has joined #openindiana[10:41:40] *** Botanic_ has quit IRC[10:45:06] *** Worsoe2 has quit IRC[10:47:43] *** kart_ has quit IRC[10:52:39] *** bens1 has joined #openindiana[10:53:10] *** Botanic__ has quit IRC[10:55:57] *** Botanic has joined #openindiana[10:55:58] *** Botanic has quit IRC[10:55:58] *** Botanic has joined #openindiana[10:56:59] *** DrLou has joined #openindiana[10:57:00] *** ChanServ sets mode: +o DrLou[10:57:59] *** InTheWings has joined #openindiana[11:04:55] *** gea_ has quit IRC[12:00:42] *** CVLTCMK0 has quit IRC[12:02:56] *** CVLTCMK0 has joined #openindiana[12:33:35] *** spanglywires has quit IRC[12:34:01] *** Botanic has quit IRC[12:34:41] *** movement has quit IRC[12:39:13] *** AlasAway is now known as Alasdairrr[12:39:25] *** monta-` has quit IRC[12:48:14] *** gea_ has joined #openindiana[13:00:36] *** Botanic has joined #openindiana[13:01:25] *** DontKnwMuch has quit IRC[13:03:15] *** movement has joined #openindiana[13:27:05] <EisNerd> moin[14:15:04] *** dijenerate has quit IRC[14:16:15] *** phretor has joined #openindiana[14:21:27] <az_ordog_maga> phretor, whats happened with the oi and esata drivers?[14:21:35] <az_ordog_maga> what is the end of the yesterday story?[14:24:30] *** elgar has joined #openindiana[14:31:18] *** phretor_ has joined #openindiana[14:34:37] *** phretor has quit IRC[14:47:46] *** idnar has joined #openindiana[14:49:58] *** gea_ has quit IRC[14:53:12] *** wonslung has quit IRC[14:53:26] *** wonslung has joined #openindiana[14:55:43] *** DrLou has quit IRC[14:55:56] *** raichoo has joined #openindiana[14:56:15] <raichoo> hi everyone[14:58:55] <joffe> hello[15:15:19] *** jedi4ever_ has joined #openindiana[16:00:18] *** McBofh has quit IRC[16:18:27] *** dijenerate has joined #openindiana[16:19:47] *** ihptru has quit IRC[16:25:39] *** McBofh has joined #openindiana[16:34:55] *** phretor has joined #openindiana[16:35:07] *** phretor_ has quit IRC[16:55:42] <ianj|arc> is there some issue with configuring printers in oi_151?[16:56:09] <ianj|arc> the GUI printer utility can't connect to CUPS, and the CUPS web interface keeps throwing 500 errors[17:18:17] *** heldchen has joined #openindiana[17:19:41] *** hajma has quit IRC[17:20:21] *** held has quit IRC[17:27:39] *** hajma has joined #openindiana[17:39:04] *** miine has quit IRC[17:45:26] *** miine has joined #openindiana[17:48:33] *** raichoo has quit IRC[17:58:40] *** dijenerate has quit IRC[18:01:33] *** mnaser_ has quit IRC[18:11:09] *** dekar_ has quit IRC[18:18:10] *** dijenerate has joined #openindiana[18:20:38] *** DrLou has joined #openindiana[18:20:38] *** ChanServ sets mode: +o DrLou[18:35:51] *** dijenerate has quit IRC[18:49:58] *** InTheWings has quit IRC[18:56:57] *** jedi4ever_ has quit IRC[18:59:06] *** jedi4ever_ has joined #openindiana[19:01:41] *** az_ordog_maga has quit IRC[19:08:09] *** Worsoe has joined #openindiana[19:09:40] <phretor> miine: I think there is something wrong. I pulled out all the disks and zpool status says they are all ONLINE[19:10:42] *** raichoo has joined #openindiana[19:13:14] <RoyK> phretor: try to start a scrub, that'll probably show some facts ;)[19:15:58] <phretor> RoyK: yeah, now the new fact is that cfgadm -lav | grep sata shows one of the 2TB disks as "unusable"[19:16:21] <phretor> RoyK: more precisely http://paste.pocoo.org/show/449466/[19:17:42] <phretor> "/pci@0,0/pci8086,5003@1f,2/disk@1,0 (sd2) i/o to invalid geometry"[19:18:09] <phretor> that happened probably when I pulled out the disks.[19:21:05] <RoyK> does the controller/driver support hotplug?[19:22:13] <tsoome> you did cfgadm -c unconfigure on disk?[19:23:33] <phretor> RoyK: well, I hope it does. It's AHCI ICH10R[19:23:36] <phretor> tsoome: no I haven't[19:23:42] <tsoome> sata disks?[19:24:04] <tsoome> always use cfgadm before pulling disks[19:24:26] <phretor> ok[19:24:46] <phretor> now I want to test the throughput over network[19:24:47] <tsoome> at a disks don't care much about that, but data framework needs it[19:24:55] <tsoome> sata*[19:24:57] <phretor> no, first I better test it locally[19:25:26] <tsoome> before going to net, apply those: http://slaptijack.com/system-administration/solaris-tcp-performance-tuning/[19:26:11] <phretor> now I'm going with a sudo dd if=/dev/zero of=/tank/test.dd[19:28:41] <phretor> looks like the disks themselves are slow[19:28:50] <tsoome> .oO[19:28:55] <phretor> 13238kps on one of the two disks making up the mirror[19:29:00] <phretor> iostat -e -d says[19:29:20] <tsoome> that dd is good if you wanna see whats your 512B blocks streaming IO will look like[19:29:40] <phretor> tsoome: should I increase the bs?[19:29:48] <tsoome> somehow i doubt you gonna see such IO in real life[19:30:09] <lblume> Use a tool that is actually made to benchmark disk I/O?[19:30:31] <tsoome> well, dd can be nice tool if it will simulate the real life workload;)[19:30:31] <phretor> it is my very first storage server; do you have any suggestion?[19:30:43] <phretor> I don't need anything complex, just plain I/O stats.[19:30:44] <tsoome> storage server for what?[19:30:53] <lennard> add a bs=4M to the dd command[19:30:54] <phretor> tsoome: (so)HO scenario[19:30:57] <lblume> bonnie++ is nice then[19:30:58] <tsoome> large file moving?[19:31:01] <lennard> at least that'll make *some* sense :P[19:31:05] <phretor> tsoome: movies and streaming[19:31:35] <tsoome> in that case the dd can be just fine, except use larger bs:D[19:31:59] <lblume> It's not like nothing better than dd was made in the last 20 years to do benchmarks and provide synthetic, readable results.[19:32:12] <tsoome> true that:D[19:32:14] <phretor> as far as I understand from a quick skim through bonnie++'s man page, it looks like it can produce better results.[19:32:49] <lblume> Anything can produce better results. dd is not a benchmark tool (unless maybe if you live in 1985).[19:33:13] <phretor> I was actually born near 1985 :)[19:33:20] <phretor> and I had a C64 at the time[19:33:29] <lblume> You probably didn't need dd at that time,then.[19:33:35] <phretor> true[19:34:20] *** sponix has joined #openindiana[19:34:40] <lblume> It seems to many people, dd is a kind of semi-religious artefact they inherited from their grandfather who gave it to them on their deathbed and made them swear to never use anything else.[19:37:15] <tsoome> well, the benchmark tool is just as good as good you are in presenting test environment and reading result. also there is little point to shoot sparrow with cannon;)[19:38:05] *** gea_ has joined #openindiana[19:39:26] * phretor is bombing his first ZFS mirror with bonnie++ -s 8G[19:39:50] <lblume> you shouldn't even need a parameter, it adjusts itself to the availablememory[19:40:17] <tsoome> mirror? you have 2 disks?[19:40:20] *** russiane39 has joined #openindiana[19:40:23] <lblume> tsoome: It's difficulr to find a tool both more used and misused than dd.[19:40:36] <tsoome> lblume: true;)[19:41:31] <phretor> tsoome: I've 4x2TB drives, but I'm using only 3 now (actually 2 in a mirror, 1 is unused) because I am fighting with the eSATA controller. So, I don't have enough free ports for the system drive, rpool, now attached to the onboard SATA.[19:42:31] <phretor> my ideal setup would be 2x320G mirrored rpool and 4x2TB either mirrored or RAIDZn. Still have to make up my mind about the data part.[19:43:13] * phretor taking a shower while bonnie++ writes intelligently...[20:02:58] *** DeanoC has quit IRC[20:06:07] *** phretor has quit IRC[20:14:55] *** randomuser has joined #openindiana[20:22:30] *** DeanoC has joined #openindiana[20:22:31] *** ChanServ sets mode: +o DeanoC[20:23:41] *** nikolam has joined #openindiana[20:26:38] *** bens1 has quit IRC[20:47:55] *** mikw has joined #openindiana[20:59:45] *** gea_ has quit IRC[21:10:01] *** yalu has quit IRC[21:10:31] *** vinc_ has quit IRC[21:11:46] *** yalu has joined #openindiana[21:11:56] *** gea_ has joined #openindiana[21:12:32] *** gea__ has joined #openindiana[21:14:48] *** jedi4ever_ has quit IRC[21:16:35] *** gea_ has quit IRC[21:17:49] *** jedi4ever_ has joined #openindiana[21:26:15] *** Botanic has quit IRC[21:26:57] *** az_ordog_maga has joined #openindiana[21:29:17] *** dekar_ has joined #openindiana[21:44:02] *** gwr has quit IRC[21:50:51] *** InTheWings has joined #openindiana[21:50:59] *** robinbowes has quit IRC[22:00:29] *** mikw has quit IRC[22:09:10] *** jedi4ever_ is now known as jedi4ever[22:14:06] *** mikw has joined #openindiana[22:17:56] *** laserbled has joined #openindiana[22:20:08] *** winIO has joined #openindiana[22:20:35] *** dijenerate has joined #openindiana[22:22:12] *** raininja has joined #openindiana[22:22:12] *** elgar has quit IRC[22:24:14] *** mikw has quit IRC[22:29:29] *** robinbowes has joined #openindiana[22:30:47] *** jedi4ever has quit IRC[22:33:33] *** nikolam has quit IRC[22:36:18] <ianj|arc> oh mpd[22:36:48] *** laserbled has quit IRC[22:36:48] <ianj|arc> I keep locking up the GUI in oi_151 and it just keeps on trucking when I restart gdm[22:37:00] <ianj|arc> I didn't realize going in that it would have that side benefity[22:37:08] <ianj|arc> s/benefity/benefit[22:39:42] <Meths> nvidia?[22:40:07] <ianj|arc> yeah[22:40:12] <ianj|arc> is that a known issue?[22:40:16] <Meths> Yeah, just having to restart gdm is really useful.[22:40:21] <ianj|arc> true[22:40:28] <ianj|arc> although it tends to happen like every half hour[22:40:29] <Meths> Yes, lots of us seem to be seeing these X lockups.[22:40:42] <Meths> Oh, I don't get anything that frequent[22:40:48] <Meths> Which card?[22:41:01] <ianj|arc> GeForce 430GT[22:41:22] <ianj|arc> brand new card[22:41:28] <Meths> Do you have anything that can guarentee a lockup?[22:41:46] <Meths> guarantee*[22:41:50] <ianj|arc> not any specific thing, but it always happens in the middle of a window animation[22:41:57] <ianj|arc> open, close or minimize[22:42:15] <Alasdairrr> the nvidia driver has been a bit poopy in the later versions[22:42:20] <ianj|arc> oof[22:42:22] <Alasdairrr> i wonder if they're only testing on S11X[22:42:34] <ianj|arc> I wasn't having any problems with 270 on Solaris 11 Express[22:42:41] <Alasdairrr> yeah[22:42:43] <ianj|arc> but 275 on oi_151 has been locking up a lot[22:43:08] <ianj|arc> I just moved to OI on Wednesday and that's the only issue I've been having[22:43:21] <Alasdairrr> quite a few of us are having it :-/[22:43:28] <Alasdairrr> its one of those hard to debug things[22:43:39] <ianj|arc> I wouldn't expect nvidia to be testing on OI, but who knows[22:43:58] *** raichoo has quit IRC[22:44:01] <Meths> Can you add a note to https://www.illumos.org/issues/367 giving the card and noting the frequency you see it please. How much RAM do you have?[22:44:07] <ianj|arc> will do.[22:44:14] *** raichoo has joined #openindiana[22:44:28] <ianj|arc> this has been a known issue since last October, eh?[22:45:16] <ianj|arc> hm, should I file it with this bug? it seems to be describing something a bit different (hang on statup instead of hang after running for a while~[22:45:19] <Meths> In various forms, yeah, lots of people found improvements as we've bumped driver versions and we've also bumped OI versions twice since that bug was reported I think.[22:47:00] <Meths> Yeah, there were other bugs that apparently got closed. Your issue matches the title so I'll leave it up to you, feel free to file a new bug if you think it more appropriate.[22:47:29] <ianj|arc> I'll add to this one, I suppose[22:47:32] <ianj|arc> it is similar...[22:48:18] *** bens1 has joined #openindiana[22:52:33] *** nikolam has joined #openindiana[22:55:07] *** bens1 has quit IRC[22:55:22] <ianj|arc> 8GB of RAM in my system[22:55:29] <ianj|arc> 1GB video RAM[22:59:08] <ianj|arc> added a post.[23:01:32] *** dijenerate has quit IRC[23:04:25] <herzen> I have sent an announcement of the OI SFE testing repo to oi-announce and oi-discuss: http://openindiana.org/pipermail/openindiana-announce/2011-July/000010.html[23:04:48] <ianj|arc> just got that here.[23:04:59] <ianj|arc> so more people to build the packages are needed?[23:05:35] <herzen> yes, it would be nice[23:05:41] <ianj|arc> cool[23:05:54] <ianj|arc> I remember asking you about it earlier but it was still a very preliminary thing at the time[23:06:15] <herzen> ah[23:06:35] <ianj|arc> I've been building a lot of SFE specs for myself lately, so I wouldn't mind pointing them at the public repo instead of my own[23:06:55] <herzen> I still have to create a repo at hg.oi.o containing the various scripts I use to build[23:07:05] <ianj|arc> OK[23:07:18] <ianj|arc> on my end I've just been setting up a CBE and then invoking pkgtool manuallyu[23:07:28] <herzen> cool[23:07:36] <Meths> herzen: Do SFE packages install to a certain prefix or just to the general OS?[23:07:59] <herzen> the only thing is that the OI SFE repo is just a mirror of the SFE repo: you can't add specs to it directly[23:08:01] <ianj|arc> it depends on how you set up your environment[23:08:37] <herzen> Meths: SFE specs generally avoid using an SFE-specific prefix[23:08:42] <ianj|arc> right, although I hear getting commit access to SFE is tough these days, so I'm fine with just taking what's there and building it for now[23:08:51] <herzen> they're written so that they can coexist with system packages[23:08:57] <ianj|arc> oh, that prefix. yeah, they pretty much all go into /usr[23:09:21] <Meths> Cool, okay.[23:09:47] <ianj|arc> I hate having to keep track of a multitude of prefixes[23:09:59] <herzen> g++-built libs go into /usr/g++, because you may have the same libs built with CC using stlport4 or stdcxx[23:11:01] <Meths> Do you provide a boost for each of those or just one?[23:11:02] <herzen> sometimes, as in the case of the Emacs package (which is at a newer version than the system Emacs and has sound enabled), they go in /usr/gnu, so as not to conflict with the system package[23:11:40] <herzen> well, since Alasdairrr wants to move away from Sun Studio, I decided to build Boost and Qt only with gcc[23:11:47] *** DrLou has quit IRC[23:12:20] <herzen> specs to exist for Boost and Qt built with CC/stdcxx, and those install in /usr/stdcxx[23:12:52] <herzen> in case someone wants to build Boost with stlport4, which is still more native to OI/Solaris than stdcxx[23:13:32] <herzen> in general, if something builds with CC, it will build with g++[23:13:37] *** sudosu has joined #openindiana[23:13:44] <Meths> herzen: Okay, I'm just thinking about a few packages I'd like to have that I've been ignoring as they're C++.[23:14:15] <Meths> So now oi-build and SFE are both using g++ right? So I can use that and be compatible with both?[23:14:44] <herzen> you should be able to[23:15:59] <herzen> many of the SFE contributors work for Oracle, so they like to use CC. so in many cases, specs exist both to build with CC and with g++[23:19:01] <herzen> ianj|arc: I believe that tomww can give commit access to SFE now[23:19:20] <ianj|arc> OK[23:20:33] *** JStoker has quit IRC[23:22:37] *** sudosu has quit IRC[23:23:19] <herzen> ianj|arc: thanks for your reply on oi-discuss. there are bound to be missing dependencies, at least. you've found them in the past, and yofuh found a couple yesterday.[23:23:52] *** InTheWings has quit IRC[23:24:17] *** raininja has quit IRC[23:24:30] <ianj|arc> yep[23:24:38] <ianj|arc> haven't found any others since the mplayer ones.[23:24:42] <ianj|arc> so far so good[23:27:23] *** raininja has joined #openindiana[23:27:23] *** raininja has joined #openindiana[23:27:39] <herzen> I am really happy with mplayer2. I went through a period when it would occasionally stop playing for a split second, but I think that that was because I had a zombie process running that took up a lot of CPU time, because it's stopped doing that.[23:27:49] <ianj|arc> ah[23:28:06] *** gea__ has quit IRC[23:28:11] <ianj|arc> I've been using it as my only video player for about three months and it's been pretty reliable[23:28:34] <ianj|arc> meaning the :10002 repo's build until a few days ago, and the :10003 repo's build since then.[23:29:48] <herzen> that reminds me: we don't need the :10002 repo any more[23:30:23] <ianj|arc> there are still a few things on it that haven't made it to 10003, but I assume that won't be the case for too much longer[23:31:06] *** JStoker has joined #openindiana[23:33:06] *** raichoo has quit IRC[23:36:02] *** raininja has quit IRC[23:36:19] <herzen> ianj|arc: like what? I just told Alasdairrr he can destroy it[23:36:35] *** gea_ has joined #openindiana[23:37:05] <ianj|arc> abiword is the only one I can recall off the top of my head[23:37:26] <ianj|arc> it hasn't been a problem for me because usually by the time I notice one that's missing, it shows up the next day[23:38:22] <herzen> abiword is in the new repo now. it needed a patch to build with gcc 4.6.[23:38:47] <ianj|arc> ah OK[23:38:53] *** raininja has joined #openindiana[23:38:54] <ianj|arc> oh also, I didn't see SFE's gcc in there[23:40:38] <herzen> the new one has SFE's gcc. it just comes as a single package, following the new facets "paradigm"[23:42:19] *** omido has joined #openindiana[23:42:54] <omido> hi guys . i'm a newbie . i've downloaded and inserted openindiana and i'm now in its live mode. but dont have access to wireless network[23:43:08] <ianj|arc> what's the package name?[23:43:20] <herzen> runtime/gcc[23:43:20] <ianj|arc> all I see is the runtime... or does that include the compiler too?[23:43:23] <ianj|arc> oh[23:43:26] <herzen> yes[23:43:31] <omido> openindiana 's live cd asks me password. what shoud i enter?[23:44:35] <herzen> if you don't want the development files (which obviously you do), you disable the devel facet (or whatever it's called) globally using pkg(1)[23:44:53] <ianj|arc> ah[23:45:02] <ianj|arc> clever... reminds me of IRIX packaging.[23:46:22] <Meths> omido: Try jack/jack[23:46:28] <omido> what?[23:46:34] <omido> its the root password?[23:47:07] *** raininja has quit IRC[23:47:25] <omido> Meths is openindiana suitable for newbies like me? should i be able to install my hardware like wireless and my graphic card on it?[23:47:33] <Meths> Can't remember how the live cd works. You sure it isn't a sudo passwd?[23:47:53] *** raininja has joined #openindiana[23:47:53] *** raininja has joined #openindiana[23:48:04] <Meths> Run the driver device utility and you'll see if you have supported hardware, should be an icon on the desktop[23:48:29] <omido> it asks me for password[23:48:41] <Meths> Which ISO are you running?[23:48:52] <omido> and i dont know what should i enter. i found nothing neither on live cd nor on OI's website[23:49:01] <Meths> Have you tried jack?[23:49:18] <omido> openindiana live x86 build 148[23:50:39] <omido> yes i tried[23:52:13] <omido> and drivers cannot be installed[23:52:18] <omido> it finds them but i cant install[23:54:52] *** Feherlofia has joined #openindiana[23:55:44] <Meths> omido: Sorry, not sure. Don't have a machine to boot live CD on at the moment.[23:57:58] *** az_ordog_maga has quit IRC