Switch to DuckDuckGo Search
   June 18, 2018
< | 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

bottom
[00:40:38] *** v_a_b1 is now known as v_a_b
[00:48:34] *** wiedi <wiedi!~wiedi@ip5b405882.dynamic.kabel-deutschland.de> has quit IRC (Quit: ^C)
[01:12:06] *** wiedi <wiedi!~wiedi@ip5b405882.dynamic.kabel-deutschland.de> has joined #omnios
[01:12:34] *** natkin <natkin!~textual@cpe-76-185-178-35.satx.res.rr.com> has joined #omnios
[02:19:20] *** v_a_b1 <v_a_b1!~volker@p57A27EA4.dip0.t-ipconnect.de> has joined #omnios
[02:21:02] *** v_a_b <v_a_b!~volker@p57A272A9.dip0.t-ipconnect.de> has quit IRC (Ping timeout: 276 seconds)
[02:46:27] *** natkin <natkin!~textual@cpe-76-185-178-35.satx.res.rr.com> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[02:59:49] *** natkin <natkin!~textual@cpe-76-185-178-35.satx.res.rr.com> has joined #omnios
[05:25:01] *** natkin <natkin!~textual@cpe-76-185-178-35.satx.res.rr.com> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[06:46:48] *** bolt <bolt!~r00t@unaffiliated/bolt> has quit IRC (Quit: telnet irc.freenode.net 6667)
[07:46:41] *** fanta7531 <fanta7531!~fanta7531@p200300F76BC267018DABCAA7CC91B948.dip0.t-ipconnect.de> has joined #omnios
[08:07:46] *** v_a_b1 is now known as v_a_b
[08:56:36] *** alhazred <alhazred!~Alex@dsl-lprbng11-54f881-150.dhcp.inet.fi> has joined #omnios
[09:10:14] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[09:14:08] *** jimklimov1 <jimklimov1!~jimklimov@31.7.243.238> has joined #omnios
[09:14:08] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[09:34:04] *** p7mo <p7mo!~pmo@ricasa01.informatik.uni-bremen.de> has quit IRC ()
[09:36:54] *** p7mo <p7mo!~pmo@ricasa01.informatik.uni-bremen.de> has joined #omnios
[09:37:36] *** bens1 <bens1!~bens1@161.122.2.81.in-addr.arpa> has joined #omnios
[10:12:56] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has joined #omnios
[10:33:45] *** andy_js <andy_js!~andy@94.14.33.204> has joined #omnios
[11:27:49] *** jimklimov1 <jimklimov1!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[11:28:32] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[11:33:56] *** psychicist__ <psychicist__!~psychicis@5356A22B.cm-6-7c.dynamic.ziggo.nl> has joined #omnios
[11:56:19] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[11:57:02] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[11:58:01] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[11:58:36] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[11:59:52] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:00:34] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:02:13] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:02:35] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:04:08] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:04:47] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:05:19] <jimklimov> I am not sure what to ask for now :) Perhaps one question is whether the box can trust both OmniOS CE and original OmniTI CA for packaging? Another is whether anything in CE succeeded those repos (is "extra" it)?
[12:06:22] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:06:47] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:06:53] <jimklimov> FWIW, it seems nothing is currently installed from perl.omniti.com, but a number of things come from ms
[12:07:46] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:08:14] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:09:31] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:10:02] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:11:02] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:11:32] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:12:16] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:12:33] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:16:57] <jimklimov> pre-empting my question, seems extras provides the first several of the random selection of those older package names I tossed at it, so should be the upgrade path from MS at least
[12:17:57] <jimklimov> I'll see how it goes in practice
[12:20:57] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:21:38] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:22:16] <jimklimov> it seems complicated - the version on bloody repo is 151027, and the machine is at 151025 :^\
[12:22:47] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:23:08] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:23:36] <andyf> Hi Jim... bloody-to-bloody upgrades aren't something I'd expect to be smooth..
[12:23:40] <andyf> but I'm sure we can work through it
[12:24:09] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:24:37] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:24:58] <andyf> You can trust the Omniti CA by adding it to the filesystem in /etc/ssl/pkg..
[12:24:59] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:25:16] <andyf> wget -P /etc/ssl/pkg https://raw.githubusercontent.com/omniosorg/omnios-build/r151022/build/ca-bundle/files/OmniTI_CA2_OmniOS.pem
[12:25:42] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:25:45] <andyf> (alternatively set the publisher property to ignore certificates)
[12:25:52] <andyf> odd thing is, bloody isn't signed
[12:26:01] <andyf> but ms is I think
[12:26:03] *** dmip <dmip!~Thunderbi@95.143.115.50> has joined #omnios
[12:30:49] *** fanta7531 <fanta7531!~fanta7531@p200300F76BC267018DABCAA7CC91B948.dip0.t-ipconnect.de> has quit IRC (Quit: ...)
[12:31:36] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:32:04] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:37:39] <jimklimov> arrgh, where is the IRC banner link to echelog when you need it? ;)
[12:38:39] <jimklimov> andyf : thanks... my previous experience was that the upgrades were rather smooth - in fact, the issue here is rather with some ms/extras package(s) that hop several releases (some seem as old as 151006, mostly 014)
[12:39:39] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:40:04] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:40:33] <jimklimov> at least, so goes my experience from around where bloody started and machine stayed on it layer after layer ;) and up into 123=>123ce=>125ce...
[12:41:33] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:42:04] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:43:14] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:43:37] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:43:56] <jimklimov> in the end I used the files from an older (omniti) BE, but thanks for encouraging that this should work and suffice :)
[12:45:08] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:45:38] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:47:04] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:47:42] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:49:02] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:49:37] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:50:31] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:51:06] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:51:27] *** dmip <dmip!~Thunderbi@95.143.115.50> has quit IRC (Ping timeout: 245 seconds)
[12:56:57] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[12:58:06] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[12:59:13] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[13:00:06] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[13:03:21] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[13:03:49] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[13:04:58] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[13:05:19] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[13:12:16] *** dmip <dmip!~Thunderbi@95.143.115.50> has joined #omnios
[14:19:37] *** gizah623 <gizah623!~gizah623@ip4da19d56.direct-adsl.nl> has joined #omnios
[14:19:44] *** gizah623 <gizah623!~gizah623@ip4da19d56.direct-adsl.nl> has left #omnios
[14:20:30] *** gizah623 <gizah623!~gizah623@ip4da19d56.direct-adsl.nl> has joined #omnios
[14:21:20] <gizah623> Hi, perhaps someone has some suggestions as to what we might be missing; we have (re)created our zpool; consisting of 18 vdevs consisting of 2 SAS drives of 1,8TB each. The drives are a mix of 512b and 512e, so we opted to format all vdevs as ashift=12; While writing, using either bonnie or dd we get the performance we expect: around 1,4GB/s, however reads get stuck at a stammering 300meg/s (dd) or about 600 meg/s (bonnie); scrubs do run at t
[14:21:21] <gizah623> he expected speed: around 2GB/s according to zpool status, or 4GB/s according to iostat -x; what might we need to tune to get our real read speed as expected?
[14:25:19] <ptribble> Is that just a single instance of dd, or multiple instances in parallel?
[14:27:56] <ptribble> If you're just running one copy, then it'll be sequential, and will read a block
[14:28:15] <ptribble> from one disk, then a block from the next disk, and so on,
[14:28:48] <ptribble> so you will only get the read bandwidth of a single drive
[14:29:01] <ptribble> unless you can trigger readahead
[14:48:09] <gizah623> ah, yes it was running single dd's
[14:49:23] <gizah623> would the same then be true for running 1 instance of bonnie++?
[14:51:13] *** andy_js <andy_js!~andy@94.14.33.204> has quit IRC (Ping timeout: 248 seconds)
[14:52:03] *** andy_js <andy_js!~andy@94.14.33.204> has joined #omnios
[14:52:13] *** jeffpc <jeffpc!~jeffpc@josefsipek.net> has quit IRC (Ping timeout: 256 seconds)
[14:53:20] <ptribble> Haven't run bonnie or bonnie++ for a while, so not sure, but I think it's single-threade
[14:53:20] *** dmip <dmip!~Thunderbi@95.143.115.50> has quit IRC (Ping timeout: 268 seconds)
[14:54:08] <ptribble> Generally, though, are bonnie and/or dd actually representative of your workload?
[14:57:07] <jimklimov> maybe run the test with large block sizes (in the 128k-fewMegs range) so a read is spread across the disks?
[14:58:23] <poige> Is there notion of GPT?
[14:59:04] <poige> Oracle's `format` seems to have gotten it in 11, but not its counterpart in Illumos
[15:00:34] <jimklimov> it believe the NOTION is certainly there since Sol10
[15:00:41] <gizah623> now retesting with iozone; which does threading, will take a bit since the machine has 128GB ram so we use a 256g testfile
[15:00:54] <jimklimov> e.g. you can make a efi/gpt layout with gparted, and format honors that
[15:01:09] <jimklimov> zpool does it internally by default, if given a whole disk (maybe over a certain size)
[15:01:32] <poige> jimklimov: first of all I can't install gparted :)
[15:01:49] <poige> 01:45:50 <poige>  pkg update parted == pkg update: 'parted' matches no installed packages
[15:01:49] <poige> 01:46:02 <poige>  pkg install parted == No updates necessary for this image.
[15:02:43] <jimklimov> pkg:/system/storage/parted at 1 dot 8.8-0.151027
[15:02:53] <poige> Yeah I know
[15:02:59] <poige> search tells the same
[15:03:15] <poige> ah, nope
[15:03:15] <poige> pkg.fmri   set    omnios/system/storage/parted                              pkg:/system/storage/parted at 1 dot 8.8-0.151026
[15:03:24] <poige> a little bit of a difference. 26 here
[15:03:39] <jimklimov> depends on distro - you must be on stable :)
[15:03:44] <poige> so ok, I got this output.
[15:03:47] <poige> What's next?
[15:03:54] <jimklimov> marked obsolete though :^\
[15:04:07] <jimklimov> so may be an empty pkg
[15:04:18] <poige> The fact is it's not getting installed
[15:04:21] *** dmip <dmip!~Thunderbi@95.143.115.50> has joined #omnios
[15:04:24] <jimklimov> just for dependencies of old setups to be happy
[15:05:20] <poige> I'm not sure what you mean. The only thing I know for sure it couldn't be installed with `pkg install parted`
[15:06:23] <poige> But anyways, I've prepared test disk with Linux and attached it to OmniOS. `format` has no notion of GPT's as I can tell
[15:07:01] <jimklimov> in OI, the tool is still there... I think it was expelled from illumos-gate at some point
[15:07:15] <jimklimov> so some distros resumed packaging it separately, some did not
[15:07:16] <poige> what is OI?
[15:07:20] <jimklimov> OpenIndiana
[15:07:22] <poige> Ah
[15:07:50] <poige> Well, so basically I don't see anything helpful to get to the disk's content
[15:07:58] <jimklimov> as for format, I think if you go to p/p to "p"rint "p"artitions, it should show the list of GPT partitions
[15:08:12] <poige> nope
[15:08:14] <jimklimov> showing like 10 entries
[15:08:15] <poige> It doesn't
[15:08:24] <jimklimov> while for legacy slices it would have 8
[15:08:34] <poige> ah, lemme check once again
[15:08:58] <poige> nope. Just 7 (0..6)
[15:09:57] <poige> I've noticed that ls -l /dev/rdsk/c7t0d0* gives even /dev/rdsk/c7t0d0s15 for e. g.
[15:10:31] <poige> but generally from what I'm seeing I'd say that working with hard disk in OmniOS is limited to default ZFS installs
[15:10:45] <poige> Anything not that trivial is ergh non-trivial
[15:10:59] <jimklimov> huh, at the system I'm looking at, "fdisk" says the disk has one 100% EFI partition
[15:11:19] <jimklimov> note to use rdsk and p0 in the device path, e.g. fdisk /dev/rdsk/c0t4d0p0
[15:11:38] <gizah623> the 100% efi partition 'might' be the protective MBR that an GPT formatted disk is supposed to have
[15:12:04] <poige>  Is there notion of GPT? -- back to the q-n
[15:12:11] <jimklimov> so I guess ZFS did create an EFI partition, and filled the whole drive with it, and sliced inside for legacy defaults  compatibility, and used one big slice as a vdev for the pool
[15:12:37] <jimklimov> so showing that the higher level has no place for other (non-sliced) data is redundant and so skipped
[15:13:25] <poige> In my case I've prepared GPT's partitions with Linux and attached disk to OmniOS
[15:14:10] <gizah623> not sure if you create your pool that way the whole disk flag will be set on the devices
[15:14:33] <jimklimov> on my dualboot laptop, there are partitions for NTFS and "Solaris2", the latter is seen as having slices numbered 0..9 and in those I have several non-zero sized ones for different pool components
[15:15:03] <jimklimov> but again, seems "format" indeed only shows data of interest to the OS - which it can manage
[15:15:43] <poige> Well, I can just repeat.
[15:16:10] <poige> Using only OmniOS you can't do anything beyond trivial ZFS installs
[15:16:18] <poige> (using GPT)
[15:16:20] <jimklimov> gizah623: the whole disk flag is not a hard requirement for goodness :) I believe it only triggers the caching policy by default, which you can set with an initscript or alike.
[15:16:40] <gizah623> true ;)
[15:16:48] <jimklimov> poige: well, now that I look at it, probably so
[15:17:00] <gizah623> also; creating the pool under another OS will set the flag; and the flag survives import afaik
[15:17:13] <jimklimov> and why would you need any other, when making a server system capable of hosting the storage and VMs natively?
[15:17:15] <jimklimov> ;)
[15:18:08] <poige> You said VMs. What do you mean -- zones?
[15:18:11] <gizah623> I would not know ;) I would not even know why one wants GPT partitions when one can just give an entire disk
[15:18:13] <jimklimov> so I guess nobody had a use-case in OmniOS to make this
[15:18:27] <jimklimov> zones, KVMs, bhyves, virtualboxes, you name it
[15:18:37] <poige> gizah623: people often think to narrow, agree
[15:18:49] <jimklimov> LX zones...
[15:19:06] <poige> say you want to use 2 SSDs: part of it for system and for ZIL
[15:19:13] <poige> other part -- for cache
[15:19:34] <jimklimov> not sure if bhyves landed yet, but there was work in progress some time ago
[15:19:44] <poige> now go do this with why one wants GPT partitions when can hust give an entire disk (c)
[15:19:45] <jimklimov> poige: easy with slicing
[15:19:57] <gizah623> yes it's possible, and I do this on my home lab, but I don't think it is 'very' important to support such an setup poige
[15:20:00] <poige> Easy?
[15:20:07] <poige> r u f-n kidding? :)
[15:20:09] <jimklimov> resize/remove the s2 slice
[15:20:15] *** natkin <natkin!~textual@cpe-76-185-178-35.satx.res.rr.com> has joined #omnios
[15:20:29] <poige> have you ever tried tinkering with slice size, start, end?
[15:20:29] <jimklimov> and make several others
[15:20:36] <jimklimov> for decades
[15:20:46] <poige> making sure it's all aligned properly for nowadays SSDs
[15:20:56] <poige> I can't it's easy
[15:20:57] <poige> It's no actually
[15:21:00] <jimklimov> that is more complicated, but doable )
[15:22:06] <jimklimov> https://wiki.illumos.org/display/illumos/ZFS+and+Advanced+Format+disks
[15:22:11] <poige> I saw this
[15:22:13] <poige> this is awful
[15:22:28] <poige> It doesn't even have parted, jzzz
[15:22:33] <jimklimov> well, with parted much of this is easier
[15:22:56] <jimklimov> run an OI USB, prepare the pool, install OmniOS to use it :)
[15:23:06] <poige> that sucks
[15:23:12] <jimklimov> nag OmniOS devs to do have parted in the system
[15:23:17] <poige> plainly does suck
[15:24:39] <jimklimov> maybe it is professional deformation that comes over time, people think binary, speak assembler, see the matrix and can't fugure out why others can't :\
[15:25:26] <andyf> fmthard and prtvtoc work with GPT partitions
[15:25:37] <jimklimov> after some time doing disk recovery with a hex editor, having these tools makes it feel trivial
[15:26:17] <jimklimov> which makes really important the complaints like yours, that in fact something is lacking for a newcomer
[15:26:30] <jimklimov> maybe docs, maybe tools - something
[15:27:06] <poige> complete b-shit
[15:27:23] <andyf> and maybe `format -e`
[15:27:27] <poige> go install 15 boxes with hex editor and don't show off
[15:27:36] <poige> andyf: nopr
[15:28:46] <andyf> been out for lunch, just catching up...
[15:29:09] <andyf> The default OmniOS install is for GPT-parititioned disks these days, along with a reserved partition for ESP
[15:29:10] <poige> andyf: prtvtoc doesn't work with GPT's 3 partitions I've created
[15:29:18] <andyf> although we don't do anything with that yet
[15:29:25] <andyf> prtvtoc definitely works with GPT..
[15:29:45] <poige> ok, it doesn't show 3 GPT partitions I've created )
[15:31:08] <poige> plus there's no parted available in despite it's found with pkg search
[15:31:10] <andyf> That is odd...
[15:31:30] <andyf> no, parted is a deprecated package marked obsolete, from before my involvement
[15:31:52] <andyf> what prtvtoc command are you running?
[15:33:06] <poige> https://www.irccloud.com/pastebin/nqWNNqxX/
[15:33:38] <andyf> any different with p0 on the end?
[15:34:13] <poige> same
[15:34:22] <andyf> The whole disk/partition management thing is a bit of a mess unfortunately
[15:34:35] <poige> that's the thing
[15:34:45] <andyf> it doesn't help that fmthard/prtvtoc try to present the same interface regardless of whether the disk is GPT or MBR + slices(VTOC)
[15:37:39] <andyf> I'm used to prtvtoc being able to display the EFI/GPT partitions though - https://paste.ec/paste/RcxvVPIv#hIgsYvsivfg1rIEWxlTBpnkoB1aGh-9sGr/hsLOWrzT
[15:38:36] <jimklimov> fwiw, https://pastebin.com/ChKxSiHN
[15:39:21] <poige> so what about parted?
[15:39:51] <jimklimov> from the looks of it, parted is currently not part of omnios - so someone has to make a recipe and put it there
[15:40:19] <poige> well, I just have no idea how otherwise to show what layout I have
[15:42:16] <jimklimov> and here is one from the dualboot laptop, without GPT/EFI I guess : https://pastebin.com/ChKxSiHN
[15:42:22] <jimklimov> but with many partitions
[15:43:18] <poige> c7t0d0     c7t0d0p1   c7t0d0p3   c7t0d0s0   c7t0d0s10  c7t0d0s12  c7t0d0s14  c7t0d0s2   c7t0d0s4   c7t0d0s6   c7t0d0s9
[15:43:19] <poige> c7t0d0p0   c7t0d0p2   c7t0d0p4   c7t0d0s1   c7t0d0s11  c7t0d0s13  c7t0d0s15  c7t0d0s3   c7t0d0s5   c7t0d0s8
[15:43:37] <poige> obviously s11, 13 15 aren't there just for kicks
[15:44:18] <jimklimov> for the latter disk (many partitions), fdisk displays the two windows ones and one for Solaris overall - under which there are several slices for different zpool components : https://pastebin.com/ChKxSiHN
[15:46:16] <jimklimov> i guess most of those tools do skip the layers of whatever could contain the slices
[15:47:11] <poige> thanks god there're hexeditors!!!!11
[15:50:22] <jimklimov> I think it boils down to what was said earlier: few if any people in the community had a real need to coexist with other OSes
[15:50:41] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)
[15:51:03] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #omnios
[15:52:17] <andyf> It will get easier once we have EFI boot
[16:01:33] <jimklimov> It was not so long ago (really not long) that new installs from CD/USB could only be done by wiping the disk - not upgrading to what the new media carried, or at least installing into another dataset so you could later migrate the data. Getting that bit into many distros is already a leap forward <grin>
[16:06:00] <poige> ah, BTW, jimklimov ! You said "can be done w/ slices"
[16:14:25] <jimklimov> yes, for the full game you 1) fdisk the raw disk to create a Solaris2 partition (100% or less if coexisting with someone, bootable or not, efi or mbr - many nuances depend on particular machine setup); 2) format the resulting disk to edit "partitions" that are in fact slices, or use prtvtoc/fmthard to set them up
[16:15:02] <jimklimov> in the end you have arbitrarily sized sliced to use as zpool components for data, zil, l2arc, perhaps some spare space for better SSD lifetime, etc
[16:16:35] <jimklimov> equivalently, on x86 you can give at least MBR (maybe efi?) partitions completely as one zfs vdev, I thikn; but I don't have an example under hand to copy-paste a layout
[16:18:26] <poige> jimklimov: nope
[16:18:42] <poige> Solaris2 would imply 2 TB limit IIRC
[16:19:52] <jimklimov> ok, then you create an "EFI" partition... looking back at my screenshots...
[16:20:09] <poige> I have my own screens actually.
[16:20:22] <poige> I've told you already I've created GPT with 3 partitions
[16:20:30] <poige> With slices I've done the thing yesterday
[16:20:48] <poige> But I'm pretty sure GPT partitions are better for that task
[16:25:16] *** Kurlon <Kurlon!~Kurlon@98.13.72.207> has joined #omnios
[16:31:19] <poige> BTW, is it possible to run installer when booted with PXE?
[16:32:45] <andyf> What does `fdisk -W - /dev/rdsk/c7t0d0p0` show?
[16:33:20] <andyf> yes, it's just /kayak/installer/dialog-install
[16:33:30] <andyf> (assuming r151026)
[16:34:41] <poige> hm
[16:34:53] <poige> but there's HTTP request for config file
[16:35:23] <poige> how to make it skipping that stuff?
[16:35:33] <poige> as to fdisk
[16:35:35] <poige> fdisk: Cannot open device /dev/rdsk/c7t0d0p0.
[16:36:13] <poige> https://www.irccloud.com/pastebin/lFkxhIRp/
[16:36:21] <poige> works that's way though
[16:38:54] <andyf> try `devfsadm -vC` just to check that all the device links are ther properly
[16:39:22] <poige> I've just booted
[16:39:26] <poige> how could that be
[16:39:41] <poige> # devfsadm -vC
[16:39:41] <poige> #  fdisk -W - /dev/rdsk/c7t0d0p0
[16:39:41] <poige> fdisk: Cannot open device /dev/rdsk/c7t0d0p0.
[16:49:57] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Ping timeout: 240 seconds)
[17:37:09] *** gizah623 <gizah623!~gizah623@ip4da19d56.direct-adsl.nl> has quit IRC (Quit: gizah623)
[17:38:59] *** xmerlin_ <xmerlin_!~xmerlin@87.19.30.113> has joined #omnios
[18:08:53] *** sac is now known as sac|afk
[18:29:18] *** ldepandis <ldepandis!~ldepandis@unaffiliated/ldepandis> has joined #omnios
[18:36:07] *** dmip <dmip!~Thunderbi@95.143.115.50> has quit IRC (Ping timeout: 276 seconds)
[19:00:05] *** bens1 <bens1!~bens1@161.122.2.81.in-addr.arpa> has quit IRC (Quit: bens1)
[19:14:06] *** dmip <dmip!~Thunderbi@95.143.115.50> has joined #omnios
top

   June 18, 2018
< | 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