Switch to DuckDuckGo Search
   February 22, 2012  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | >

Toggle Join/Part | bottom
[00:08:29] <On__> tsoome did you get to check out the pstack output?
[00:12:27] *** TPickle has quit IRC
[00:12:29] <tsoome> was just looking again, but its not too helpful for me at least. maybe im too sleepy alreay, meybe its better to check with debugger
[00:13:03] <liberal> you can join a windows computer to a kerberos zone w/o joining it to a domain
[00:13:09] <liberal> s/zone/realm/
[00:13:59] <tsoome> yea but that not the point really
[00:14:18] <liberal> sorry, i haven't finished catching up on the convo
[00:15:13] <tsoome> the original issue is the bind is crashing if built to use dynamic updates from samba.
[00:15:41] <tsoome> samba4, to be exact (which in turn is alpha version)
[00:15:42] *** wiseen has joined #openindiana
[00:18:30] <On__> liberal, yeah, that would be good but the CIFS wouldn't be able to auth to the domain.
[00:18:59] <On__> however, that is useful to know
[00:19:27] <On__> I had tried to use MIT's old windows implementation before about 6 years ago or so
[00:20:18] *** hunter has joined #openindiana
[00:20:33] <On__> just to throw it out there, this wounldn't be a 64/32 bit problem?
[00:20:41] <On__> the samba library is 32bit...
[00:20:49] <On__> (why, I don't know)
[00:21:12] <tsoome> no, you would know at once if it is;)
[00:21:36] <liberal> brb, irc client is acting up.
[00:21:38] <On__> ah, I see, thanks
[00:21:41] *** liberal has left #openindiana
[00:21:50] *** liberal has joined #openindiana
[00:22:05] <tsoome> the runtime linker is very loud about 32/64 bit
[00:22:41] <On__> you also saw the bit about Samba not writing the runpath right?
[00:23:08] <tsoome> yes, but thats build scripts/system issue
[00:23:09] <On__> I ran ldd -d to find what it expected to link to, and added their directory with elfedit
[00:24:01] <On__> I don't expect that would cause a problem.
[00:24:30] <tsoome> you can also try gdb (or dbx) /usr/local/bin/named core and see what debugger thinks about the location of the crash
[00:24:33] <On__> Another point of interest is that when I build config.developer it just segfaults, no need to change the runpath.
[00:24:47] <On__> I've done that...
[00:24:54] <On__> let me see if I can get that output again
[00:24:55] *** CoilDomain has joined #openindiana
[00:27:42] <On__> #0 0xfefcce01 in rtld_db_dlactivity () from /lib/ld.so.1
[00:27:54] <On__> So there is where pstack ends also
[00:28:23] <tsoome> hm
[00:29:02] <tsoome> maybe its better idea to add correct -R to the module linkage you did fix manually and relink
[00:29:23] <tsoome> still, just wild guess
[00:29:28] *** enmand has quit IRC
[00:29:47] <tsoome> anyhow, its way too late for me, 1:29am here:D
[00:31:07] <On__> good night and many thanks
[00:41:32] <On__> hmm, it seems to be a segfault specifically in libcrypto.so.0.9.8
[00:52:05] <richlowe> well that sums up my day
[01:01:36] *** wiseen has quit IRC
[01:01:53] *** Savis has quit IRC
[01:04:07] <On__> hmmmm
[01:04:12] <On__> what if it isn't samba4 at all
[01:04:28] <On__> what if it is simply a problem between bind and openssl
[01:05:10] <On__> and samba4 is only asking bind to use openssl
[01:06:34] *** enmand has joined #openindiana
[01:10:05] *** laogrymt2_ has joined #openindiana
[01:14:13] *** e1kg has quit IRC
[01:15:28] *** InTheWings has quit IRC
[01:23:04] *** master_of_master has quit IRC
[01:24:19] *** master_of_master has joined #openindiana
[01:24:59] *** jMCg has left #openindiana
[01:30:54] *** ozquera has quit IRC
[01:40:42] *** vados has quit IRC
[01:52:19] *** joffe has quit IRC
[02:02:46] *** Seony has joined #openindiana
[02:16:41] *** tsoome has quit IRC
[02:26:52] *** ritzz has joined #openindiana
[02:27:07] *** ritzz has quit IRC
[02:27:23] *** timclassic has quit IRC
[02:28:17] *** der_adi has quit IRC
[02:40:52] *** enmand has quit IRC
[02:41:25] *** enmand has joined #openindiana
[02:41:38] *** dijenerate has joined #openindiana
[02:55:51] *** On__ has quit IRC
[03:10:16] *** jwit has joined #openindiana
[03:17:40] *** Licenser_ has joined #openindiana
[03:18:08] *** Licenser has quit IRC
[03:18:08] *** Licenser_ is now known as Licenser
[03:23:49] *** Hakon|mbp has joined #openindiana
[03:29:40] <Hakon|mbp> Hi all! I was trying to import a zfs pool created on FreeBSD today and had some problems
[03:30:02] <Hakon|mbp> this is the output of zdb on freebsd https://gist.github.com/76efc6e9f31a7adad881
[03:30:16] <Hakon|mbp> (tried to import on 151a)
[03:31:47] <Hakon|mbp> the problem (as i see it) is that one of the vdevs (children[1]) is based on raw disks
[03:32:24] <Hakon|mbp> and in openindiana it only recognizes it as unallocated space
[03:33:04] <Hakon|mbp> when i try to import the pool it finds the vdev based on partitions of the disks
[03:34:06] <Triskelios> yeah, illumos ZFS expects whole-disk vdevs to actually have an EFI label
[03:34:14] <Hakon|mbp> oh
[03:34:43] <Hakon|mbp> an identificator?
[03:35:14] <Hakon|mbp> can i trick it to create it somehow?
[03:35:32] <Hakon|mbp> or would i have to destroy the vdev to do it
[03:36:30] <Triskelios> Hakon|mbp: can you check if FreeBSD didn't label them for you?
[03:36:41] <Hakon|mbp> hmm, how?
[03:37:01] <Triskelios> I forget the name of the FreeBSD partition tool
[03:37:20] <Hakon|mbp> gpart ?
[03:37:27] <Triskelios> that's it
[03:38:49] <Hakon|mbp> hmm, it shows up as no partition table
[03:39:11] *** voidcoder has quit IRC
[03:41:14] <Triskelios> you're probably screwed
[03:41:19] <Hakon|mbp> haha
[03:41:28] <Hakon|mbp> well, my data is intact on freebsd
[03:41:31] <Hakon|mbp> so no harm done there
[03:47:22] *** Botanic has quit IRC
[03:49:57] <Hakon|mbp> thanks for the help, gotta go
[03:52:25] <Hakon|mbp> one more question, can associate the disk with the pool without that label?
[03:52:35] <Hakon|mbp> like, zpool online datar <disk>
[03:52:38] <Hakon|mbp> or something
[03:55:25] <Triskelios> I'm not entirely sure... these configurations haven't really been tested
[03:56:07] <Hakon|mbp> hmm, the zpool man page states that a whole disk must be used for it to be portable
[03:57:29] <Triskelios> a "whole disk" vdev always has an implicit label
[03:57:44] <Hakon|mbp> i might have mixed the disks that were available in illumos.. i think that means i gotta sleep some
[03:58:04] <Hakon|mbp> hmm, ah
[03:58:28] *** jwit has quit IRC
[03:58:35] <Hakon|mbp> and zfs writes that independently of the OS it's on
[03:58:44] <Triskelios> I think it's a quirk of GEOM that you can use unlabeled disks
[03:58:47] <Triskelios> yes
[03:59:22] <Hakon|mbp> they might be labeled though
[03:59:40] <Triskelios> gpart should be able to see it
[03:59:45] <Hakon|mbp> hmm
[04:00:00] <Hakon|mbp> it was imported when i checked it
[04:00:52] <Hakon|mbp> i'll check again tomorrow and give you a peep of what i find out
[04:02:29] <Hakon|mbp> i might be able to convert it to a raw-disk only pool by replacing disks and resilvering
[04:02:35] <Hakon|mbp> will take some time though ;)
[04:03:11] *** jwit has joined #openindiana
[04:09:36] *** tsoome has joined #openindiana
[04:10:07] *** Botanic has joined #openindiana
[04:10:36] *** jwit has quit IRC
[04:11:02] *** kart_ has joined #openindiana
[04:14:11] *** jwit has joined #openindiana
[04:23:42] *** Kurlon_ has joined #openindiana
[04:23:42] *** Kurlon has quit IRC
[04:34:51] *** Seony has quit IRC
[04:35:51] *** enmand has quit IRC
[04:45:29] *** Seony has joined #openindiana
[04:48:17] *** Aengus- has quit IRC
[05:23:23] *** Norrin has joined #openindiana
[05:24:50] *** Norrin has left #openindiana
[06:08:25] *** kart_ has quit IRC
[06:10:01] *** kart_ has joined #openindiana
[06:12:49] *** e1kg has joined #openindiana
[06:29:00] *** der_adi has joined #openindiana
[06:45:56] *** Zenos has quit IRC
[06:50:11] *** d_k has joined #openindiana
[07:03:21] *** jw_urodoc has joined #openindiana
[07:13:48] *** rwentzel has joined #openindiana
[07:14:13] <rwentzel> hello
[07:14:49] <rwentzel> have I have a question on a fresh install
[07:15:03] <rwentzel> running nappit on top of 151
[07:15:25] <rwentzel> I have a 128gb ssd I wanted to use for the zfs but it is only using 2gb of it
[07:15:29] <rwentzel> seems liek a waste
[07:15:58] <rwentzel> is it possible to install the OS on this ssd and use a partition on the SSD for that cache?
[07:17:41] <rwentzel> should i just use a class 10 thumb drive to do same function? and take SSD out?
[07:25:19] <rwentzel> back to work.. will try again later
[07:25:40] *** rwentzel has quit IRC
[07:30:24] *** d_k has quit IRC
[07:53:17] *** tsoome has quit IRC
[08:27:47] *** tsoome has joined #openindiana
[08:31:40] *** Licenser has quit IRC
[08:34:29] *** gatekeep has quit IRC
[08:36:02] *** gatekeep has joined #openindiana
[08:36:49] *** merzo has quit IRC
[08:39:14] *** Worsoe has joined #openindiana
[08:48:01] *** hcaulfield57 has joined #openindiana
[08:50:15] *** hcaulfield57 has quit IRC
[08:52:00] *** r4idenZA has joined #openindiana
[08:57:33] <madwizard> Coffee
[09:00:07] *** Licenser has joined #openindiana
[09:09:33] *** PiotrSikora has quit IRC
[09:10:34] *** PiotrSikora has joined #openindiana
[09:10:41] <sivanov_> btw, what Solaris Adaptec driver to use with OI: from solaris 10 or solaris 11?
[09:14:56] *** |AbsyntH| has joined #openindiana
[09:16:46] *** bens1 has joined #openindiana
[09:22:15] *** Vutral has quit IRC
[09:28:42] <tsoome> huh SRU4 is released...
[09:42:03] *** |AbsyntH| has quit IRC
[09:42:55] <madwizard> tsoome: SRU4?
[09:43:16] <tsoome> for s11
[09:43:30] <tsoome> support repository update 4
[09:44:13] <madwizard> ah
[09:44:21] <madwizard> I'm not up to date with s11
[09:45:47] <sivanov_> tsoome, btw, i'll try ibm br10i in IT mode in the backup server tomorrow
[09:46:12] <sivanov_> i think it will work better with s11&zfs than adaptec
[09:47:10] <sivanov_> i opened the server today and the sas cable is long enough to reach to a card in the raiser
[09:59:13] <tsoome> btw, i have one system with aac and disks exported to os. simple mirror and dd:2718957568 bytes (2.7 GB) copied, 21.648 s, 126 MB/s
[09:59:26] <tsoome> but iostat isnt too nice:
[09:59:33] <tsoome> 0.0 40.0 0.0 40580.9 0.0 10.0 0.1 249.9 0 100 c3t0d0
[09:59:33] <tsoome> 0.0 44.0 0.0 44677.1 0.0 10.0 0.1 227.2 0 100 c3t1d0
[09:59:42] <tsoome> not too bad either
[10:04:01] *** d_k has joined #openindiana
[10:11:13] *** Botanic has quit IRC
[10:12:15] *** Cobi has quit IRC
[10:12:47] *** spanglywires has joined #openindiana
[10:16:38] <tsoome> ahm, maybe its on writetrough mode… cba to check atm…. :P
[10:17:34] *** Cobi has joined #openindiana
[10:22:26] <sivanov_> but now im going to try centos&scst
[10:25:40] *** jswlinux has joined #openindiana
[10:28:46] *** Seony has quit IRC
[10:28:52] *** jswlinux is now known as Seony
[11:06:49] *** smrt has quit IRC
[11:07:50] *** bahamas10 has quit IRC
[11:08:07] *** papertigers has quit IRC
[11:16:12] *** Vutral has joined #openindiana
[11:28:02] *** Whoopsie has joined #openindiana
[11:28:02] *** ChanServ sets mode: +v Whoopsie
[11:34:25] *** bahamas10 has joined #openindiana
[11:35:01] *** papertigers has joined #openindiana
[11:37:24] *** vados has joined #openindiana
[11:46:22] *** smrt has joined #openindiana
[11:47:12] *** smrt has joined #openindiana
[11:52:06] *** d_k has quit IRC
[12:01:55] *** InTheWings has joined #openindiana
[12:06:42] *** dvx has joined #openindiana
[12:11:52] *** jw_urodoc has quit IRC
[12:16:10] *** voidcoder has joined #openindiana
[12:31:49] * RoyK just got his new 80TB toy
[12:37:29] *** Hakon|mbp has quit IRC
[12:37:35] <holden__> RoyK: o.O
[12:41:48] <RoyK> 80 2TB drives in striped mirrors
[12:41:57] <RoyK> 2,5TB SSD cache
[12:57:07] <cypromis> ouch
[12:57:14] <cypromis> happy running out of RAM ?
[13:03:20] *** Seony has quit IRC
[13:07:14] <patdk-lap> should be fine, with a nice 200gigs ram
[13:07:37] *** d_k has joined #openindiana
[13:08:20] *** voidcoder has quit IRC
[13:10:14] <RoyK> cypromis: 96GB
[13:10:23] <RoyK> cypromis: no dedup
[13:10:30] <patdk-lap> wouldn't that be cutting it alittle close?
[13:10:48] <patdk-lap> going use up 50gigs of ram to index your ssd's
[13:10:48] <cypromis> with 2.5TB L2arc ?
[13:10:56] <cypromis> yah
[13:11:01] <cypromis> depending on your block sizes
[13:12:10] <RoyK> to be honest, I haven't even thought of that - the SSDs were bought and added some time back to troubleshoot performance - perhaps it's too much...
[13:12:25] <RoyK> guess I can use a bit less - I really don't need all of that
[13:12:47] <cypromis> :)
[13:13:22] <RoyK> probably a few ppl around that want those SSDs for their laptops (if I don't tell them what they've been used for)
[13:13:57] <RoyK> how do I do the math for L2ARC alone, no dedup etc?
[13:22:17] *** Hakon|mbp has joined #openindiana
[13:25:07] <RoyK> so...
[13:25:17] <RoyK> if setting this up the default way
[13:25:42] <RoyK> and later swapping the drives for something with 4k sectors emulated as 512b sectors
[13:25:50] <RoyK> what would the performance penalty be?
[13:27:34] <RoyK> also - any idea how to find the average blocksize on a fs? I found some zdb magick some time back, but that just ran out of memory
[13:29:54] *** lennard has quit IRC
[13:31:55] *** lennard has joined #openindiana
[13:33:13] <cypromis> since zfs has flex block size thats tricky
[13:33:35] <cypromis> meaning every time you change it it will write the new files with the new block size
[13:33:42] <cypromis> so you can have any amount of different ones you can imagine
[13:34:57] <RoyK> so what I need to do, is just get a few terabytes of ARC and then do that zdb magick? ;)
[13:35:43] *** kart__ has joined #openindiana
[13:37:39] *** kart_ has quit IRC
[13:40:18] <cypromis> yah tht would for sure help
[13:40:40] <cypromis> on th other hand you would need to do quite some adjusting since RAMN > 144GB always starts to require some adjustments
[13:40:53] <cypromis> just so we FE do not gt bored of course
[13:40:53] <cypromis> :D
[13:48:55] *** voidcoder has joined #openindiana
[13:53:44] *** bens1 has quit IRC
[13:57:24] *** enmand has joined #openindiana
[14:10:15] *** CoilDomain has quit IRC
[14:25:15] *** Kurlon_ has quit IRC
[14:30:44] *** Guest66561 is now known as LadySerena
[14:32:51] *** joffe has joined #openindiana
[14:35:05] *** der_adi_ has joined #openindiana
[14:37:14] *** der_adi has quit IRC
[14:37:14] *** der_adi_ is now known as der_adi
[14:43:33] *** dvx has quit IRC
[14:52:55] *** bens1 has joined #openindiana
[15:13:29] *** Kurlon has joined #openindiana
[15:19:03] *** e1kg has quit IRC
[15:25:39] *** d_k has quit IRC
[15:32:17] *** tsoome has quit IRC
[15:45:33] *** dijenerate has quit IRC
[15:47:05] *** dijenerate has joined #openindiana
[15:47:20] *** |AbsyntH| has joined #openindiana
[15:51:25] *** voidcoder has quit IRC
[15:54:45] *** voidcoder has joined #openindiana
[16:02:41] *** Neddie has joined #openindiana
[16:03:10] *** kart__ is now known as kart_
[16:03:35] *** flyz has quit IRC
[16:05:28] *** limgad has quit IRC
[16:06:52] *** limgad has joined #openindiana
[16:10:53] *** flyz has joined #openindiana
[16:12:54] *** Worsoe has quit IRC
[16:21:00] *** voidcoder has quit IRC
[16:22:53] *** tsoome has joined #openindiana
[16:24:25] *** dijenerate has quit IRC
[16:28:09] *** dijenerate has joined #openindiana
[16:34:32] *** WLLem_ has joined #openindiana
[16:40:18] *** voidcoder has joined #openindiana
[16:43:42] *** WLLem_ has quit IRC
[16:44:41] *** c2n2 has joined #openindiana
[16:45:14] *** c2n2 has left #openindiana
[16:49:37] *** timclassic has joined #openindiana
[16:57:08] *** voidcoder has quit IRC
[17:01:08] *** Aengus has joined #openindiana
[17:03:04] *** loganbr has quit IRC
[17:04:59] *** mikw has joined #openindiana
[17:06:51] *** Bdragon has quit IRC
[17:09:38] *** Bdragon has joined #openindiana
[17:25:42] *** DwD__ has joined #openindiana
[17:27:29] *** axisys has joined #openindiana
[17:29:46] *** Neddie has quit IRC
[17:37:01] *** loganbr has joined #openindiana
[17:37:09] *** raichoo has joined #openindiana
[17:45:02] *** WLLem_ has joined #openindiana
[17:46:04] *** voidcoder has joined #openindiana
[17:50:10] *** nikolam has joined #openindiana
[18:03:10] <nikolam> how to make ipfilter be started after putting wireless and other connections up .
[18:03:26] *** r4idenZA has quit IRC
[18:03:37] <nikolam> how to make routing for internet use by default one adapter and not another (Lan)
[18:04:04] *** girafe has joined #openindiana
[18:06:30] *** |AbsyntH| has quit IRC
[18:07:50] *** Botanic has joined #openindiana
[18:07:50] *** Botanic has joined #openindiana
[18:14:40] <Kurlon> Remove the default route for the LAN
[18:17:11] *** mikw has quit IRC
[18:20:43] *** nevermetme has joined #openindiana
[18:23:22] *** WLLem_ has quit IRC
[18:33:25] *** e1kg has joined #openindiana
[18:36:34] *** Zenos has joined #openindiana
[18:39:19] <nevermetme> hey, can anyone help me to get a intel 82579v network adapter working?
[18:39:51] *** nevermetme has quit IRC
[18:40:05] *** nevermetme has joined #openindiana
[18:40:34] <nikolam> Kurlon, I just did. Problem is to make it use internet on Lan, once other adapter is not working and get back using other adapter by default when it is connected.
[18:46:48] *** d_k has joined #openindiana
[18:47:19] <nikolam> nevermetme, 82541PI says it is working here at me.
[18:50:14] *** kart_ has quit IRC
[18:52:01] *** kart_ has joined #openindiana
[18:52:26] <nevermetme> well, the 82579v apparently has some problems, i already did everything in here http://lalsamtech.blogspot.com/2011/11/getting-intel-82578v-to-work-with.html but the device driver utility still marks it as misconfigured
[18:53:39] <nikolam> and 82579v is kind of supported in S11 final, as per Solaris HCL, so question is for Intel to make driver available on Illumos/Openindiana if it is not already there,
[18:54:04] <nikolam> aha i see
[18:54:16] *** Whoopsie has quit IRC
[18:54:39] *** e1kg has quit IRC
[18:56:20] <Triskelios> nevermetme: just update the OS if you have some other network connection
[18:56:44] *** nikolam has quit IRC
[18:57:47] <nevermetme> that's my problem, I don't have any other NIC around... is there another way to update it manually? downloading the updates on a USB thumb drive or something like that?
[19:08:25] <Triskelios> nevermetme: you could do that, although the update is very large
[19:09:09] <Triskelios> nevermetme: the test binary in the bug should work if you installed it properly
[19:11:26] <nevermetme> as root I replaced the existing one (32bit, checked it before with file), added the line to driver_aliases and then rebooted, anything I did wrong there?
[19:12:54] *** vados has quit IRC
[19:12:54] <Triskelios> technically, you should use update_drv and you don't need to reboot
[19:13:08] <Triskelios> you have a 32-bit system?
[19:13:28] <Triskelios> that NIC is usually found integrated on newer boards
[19:13:57] <nevermetme> to be honest, I shouldn't. but that is my very first install of OI, so I'm still kinda green
[19:15:39] <Triskelios> are you sure your system is 32-bit?
[19:15:45] <richlowe> isainfo -b
[19:15:48] <richlowe> it almost certainly isn't
[19:16:06] <nevermetme> 64
[19:16:29] <nevermetme> but then, if i run file against a few binaries in /kernel/drv they all seem to be 32bit
[19:16:31] <Triskelios> nevermetme: replace /kernel/drv/amd64/e1000g
[19:17:24] <Triskelios> nevermetme: because both the 32- and 64-bit kernel are installed. the appropriate one is chosen at boot
[19:18:43] <Triskelios> update_drv -a -i '"pci8086,1503"' e1000g
[19:18:57] <Triskelios> ifconfig e1000g0 plumb
[19:19:08] <Triskelios> you should be able to use the interface
[19:21:07] *** e1kg has joined #openindiana
[19:25:36] <nevermetme> HA, thanks, seems like that was the problem, now it works. Thank you!
[19:30:29] *** kart_ has quit IRC
[19:34:26] *** Botanic has quit IRC
[19:35:52] <Triskelios> enjoy
[19:40:43] *** Licenser has quit IRC
[19:53:23] *** raichoo has quit IRC
[19:53:28] <axisys> how do I get latest ff ? It is running 3.6.2
[19:55:52] *** Botanic has joined #openindiana
[19:55:52] *** Botanic has joined #openindiana
[19:57:26] <e1kg> axisys: ftp://ftp.mozilla.org/pub/firefox/
[19:57:55] <Triskelios> there might be something in /experimental
[19:59:36] <axisys> e1kg: thanks.. not that it is a must.. but is there a ips version ? I found pkg
[20:03:50] <e1kg> axisys: no. do you worry about upgrade frequently?
[20:05:33] <axisys> e1kg: no
[20:08:02] *** raichoo has joined #openindiana
[20:10:36] *** rame has quit IRC
[20:10:56] *** Seony has joined #openindiana
[20:15:46] *** whhat has joined #openindiana
[20:16:50] *** descipher has joined #openindiana
[20:17:22] <whhat> I am trying to use logadm to rotate, gzip, and move the file in the end. The -z option for logadm gzips the file at the end. I s there no way to gzip first, and then move the file?
[20:18:48] <viridari> you need to rotate it out of the way before you gzip it
[20:19:00] <viridari> log msgs continue to come in
[20:19:39] <whhat> Oh i see, right. What i am confused about is that once I move it, the -z option doesn't know where the file is, does it?
[20:19:54] *** Licenser has joined #openindiana
[20:20:14] <viridari> are you trying to move it to another directory?
[20:20:49] <whhat> yea, to an NFS filesystem
[20:21:31] <richlowe> I would expect logadm to know where the file is, because it put it there.
[20:21:35] <viridari> whhat: I might be tempted to just let it rotate the logs in-place without gzipping or rotating them, and then have a followup cron job to handle the compression & moving them
[20:21:36] <richlowe> but it's been probably 6 years since I touched it
[20:21:46] <patdk-wk> who was saying they really wanted ksm?
[20:23:44] <viridari> whhat: for i in `find /var/log -type f -name \*.[0-9]`; do (bzip2 -9 $i && mv ${i}.bz2 /nfs/path); done ### something like that
[20:24:45] * patdk-wk just syslogs -> to remote system
[20:24:45] <whhat> i can do it thorugh cron, i was trying to se if theres a way i could do it all thorugh logadm :P
[20:24:50] <patdk-wk> and just handles it on that system
[20:25:15] <RoyK> remote syslogging is good
[20:25:27] <RoyK> and very, very easy...
[20:26:13] <patdk-wk> remote syslog -> zfs with compression on :)
[20:26:20] <descipher> viridari: Is the target NFS system zfs? If so just enable compression ...
[20:26:33] <viridari> descipher: I think you wanted whhat
[20:26:38] <whhat> It's not unfortunately
[20:26:47] <RoyK> patdk-wk: gzip is neat for syslogs
[20:27:23] <patdk-wk> I should give it a try
[20:27:31] <descipher> yes, my bad
[20:27:42] <viridari> if it were zfs you could forget about log rotation, enable zfs compression, snapshot /var/log before nulling out every log you want to rotate.
[20:27:51] <patdk-wk> right now it's like, syslog -> mysql -> ext4 -> iscsi zfs lun with lzjb
[20:27:53] <RoyK> patdk-wk: low cpu usage, very good compression
[20:27:59] <patdk-wk> wonder how gzip instead would perform
[20:28:31] <RoyK> it's rather heavy compared to lzjb, but then, what isn't
[20:29:01] <patdk-wk> well, I'm more worried about reading latency for the db
[20:29:04] <patdk-wk> than writing though
[20:29:40] <RoyK> patdk-wk: dunno about dbs, but for flat files, it's close to zero, it can even be faster reading from compressed than non-compressed
[20:30:17] <patdk-wk> I should give it a try, and use 128k blocksize :)
[20:30:25] <RoyK> yeah :)
[20:30:30] <patdk-wk> using 8k blocks now :)
[20:30:36] <RoyK> ouch
[20:30:44] <patdk-wk> not really, the system was made for 4k or less
[20:31:00] <patdk-wk> my average nfs filesize is 6k
[20:31:21] <whhat> how do you figure what's your average nfs filesize? dtrace?
[20:31:33] <patdk-wk> whhat, du
[20:31:38] * RoyK is very curious about that too
[20:31:42] <patdk-wk> just a simple perl script
[20:31:43] <whhat> du, really?
[20:31:48] <patdk-wk> that went though all the folders
[20:31:55] <Botanic> count the number of files and divide by size?
[20:32:08] <RoyK> patdk-wk: more like 'accessed files average size', I guess
[20:32:08] <patdk-wk> well, I did it differently
[20:32:14] <patdk-wk> I did, files per size range
[20:32:21] <patdk-wk> for powers of 2
[20:32:28] <patdk-wk> and stopped at 1meg
[20:32:47] <patdk-wk> 90% of my files fell in the 4-8k ranges
[20:33:01] <RoyK> what sort of files are these?
[20:33:10] <patdk-wk> websites
[20:33:14] <RoyK> ok
[20:33:31] <patdk-wk> maildir files came in at about 16k I think
[20:33:59] <patdk-wk> I'm gzip maildir files now, before they go over nfs
[20:34:05] <patdk-wk> I wish I could do that for all nfs files
[20:34:11] <patdk-wk> have the nfs client gzip/degzip
[20:53:11] *** mikaeld has quit IRC
[20:53:33] *** c2n2 has joined #openindiana
[20:56:42] <patdk-wk> going take alittle over an hour to move :(
[20:56:51] <patdk-wk> zfs gzip will only use a single cpu
[20:57:02] <patdk-wk> and a single core only nets me about 25mb/sec
[20:57:17] <patdk-wk> using pigz I get closer to 100mb/sec compression speeds
[20:57:40] <Kurlon> Tunnel the nfs connection and do the compression on the stream?
[20:58:00] <patdk-wk> kurlon, heh?
[20:58:07] <patdk-wk> but then I would have to gzip it on the nfs server
[20:58:16] <patdk-wk> I wan the nfs server to save it gzip'd :)
[20:58:35] <Kurlon> Ok, so the net connection isn't the bottleneck?
[20:58:36] <patdk-wk> put some kind of fuse filesystem ontop of nfs?
[20:58:41] <patdk-wk> kurlon, nope
[20:58:46] <patdk-wk> I tested it before
[20:59:02] <patdk-wk> with webstats atleast
[20:59:14] <Kurlon> hrmmm
[20:59:16] <patdk-wk> it takes 4hours to process webstats, and for it to make it's hundreds of html files
[20:59:28] <patdk-wk> if I tell save all html files as gzip, it only takes 3hours
[20:59:40] <patdk-wk> cause of reduced nfs traffic (avg file isze 1k)
[21:00:22] <patdk-wk> for email, same thing, email comes in, gets compressed, then saved over nfs, read over nfs, decompressed, and servced via imap
[21:00:28] <patdk-wk> just wish I could do it for other people
[21:00:50] <patdk-wk> webstats, I have apache serving the gzip files directly, and if their client doesn't support compression, then it will uncompress it first
[21:01:34] <patdk-wk> not sure what about compression helped nfs, but it does for my workload, a good amount :)
[21:01:53] <patdk-wk> I just got done checking my nic's
[21:02:11] <patdk-wk> so far, over my gigabit nic,s none of them have a max usage >90mbit
[21:02:12] <Kurlon> Can you multithread the copy to coax ZFS to use more threads for writing?
[21:02:33] <patdk-wk> Kurlon, I tried that before I believe
[21:02:40] <patdk-wk> think the zpool is singlethreaded there
[21:02:53] <Kurlon> Or setup a ZIL drive so ZFS can stream to it, then compress/dump over time?
[21:03:01] <patdk-wk> zil is setup
[21:03:34] <Kurlon> What's it showing for %busy?
[21:04:43] <patdk-wk> all drives range from 6-10%
[21:04:50] *** Savis has joined #openindiana
[21:04:56] *** andrnils has joined #openindiana
[21:06:30] <Kurlon> I'd say separate the work then, stream the files over into say, an uncompressed tarball to get them onto the target system as quickly as possible, then blow that stream down to the compressed zfs locally in one shot to max out the cpu.
[21:06:52] *** andrnils has quit IRC
[21:07:12] <patdk-wk> I do max it out locally, but only using one core
[21:08:47] *** voidcoder has quit IRC
[21:09:13] *** voidcoder has joined #openindiana
[21:14:07] *** PiotrSikora has quit IRC
[21:15:20] *** vados has joined #openindiana
[21:16:18] *** PiotrSikora has joined #openindiana
[21:21:14] *** d_k_ has joined #openindiana
[21:21:14] *** d_k has quit IRC
[21:21:14] *** d_k_ is now known as d_k
[21:25:07] *** pothos_ has joined #openindiana
[21:26:28] <pettson> Who is maintaining the IPS-packages used in openindiana? Is that OI with community, or is it oracle?
[21:26:38] *** pothos has quit IRC
[21:26:39] <pettson> And how is security issues handled?
[21:27:09] *** bens1 has quit IRC
[21:27:46] <Agnar> pettson: community. security: well...;)
[21:28:04] <Agnar> pettson: interested in a new not-well-payed-job? ;)
[21:28:08] *** pothos_ has quit IRC
[21:28:16] <pettson> Agnar: Not exactly. :)
[21:28:28] *** pothos has joined #openindiana
[21:29:08] <pettson> My current server is running sol10, and I found out that building my own solaris-packages doesn't really work in the long run (works as student, with gf an job, not so much).
[21:29:26] <pettson> So I will probabaly replace it with either openindiana of freebsd.
[21:31:08] *** pothos has quit IRC
[21:31:10] <pettson> Agnar: So, do you have some feeling for how security suff is handled? Is it like same day +/- a couple of days for larger packages, or can it go like weeks?
[21:31:29] *** pothos has joined #openindiana
[21:31:42] <Agnar> pettson: no idea actually...
[21:33:23] <pettson> okay
[21:35:32] <alanc> I think given the current state, several security fixes have gone unpushed for months, while they try to get stuff in order to have a stable release that gets more active security fixing
[21:35:50] <richlowe> sounds depressingly accurate
[21:36:02] <richlowe> though I think Meths is largely on top of it in theory.
[21:36:08] <richlowe> and... someone else
[21:37:12] <pettson> I think it's unfortune that not all "opensolaris-distros" use the same packaging. :(
[21:37:48] <richlowe> It wouldn't help.
[21:37:55] <richlowe> about 20 more people doing things would
[21:38:41] *** _Tenchi_ has quit IRC
[21:38:51] <pettson> richlowe: Well, better if the 3 projects could cooperate for such infrastructure.
[21:40:44] <alanc> that's what illumos-userland is trying to do, at least for OI & illumian/nexenta
[21:41:49] <pettson> alanc: Will nexenta switch to ips?
[21:42:11] <alanc> no, they're building debian apt packages out of the same source tree
[21:42:28] <alanc> if they switched to ips, then their distro would be utterly pointless, instead of just mostly
[21:42:30] <pettson> ah, cool
[21:42:46] <pettson> =)
[21:43:18] <pettson> Well, if there's an easy way to build both kinds from the same source its a step forward at least.
[21:43:33] *** axisys has quit IRC
[21:43:47] <alanc> since there'd be little left to answer the question "Why put in a ton of effort building a distro that's almost exactly the same as OI instead of just helping with OI?"
[21:44:02] *** axisys has joined #openindiana
[21:44:39] <richlowe> "If you build appliances, why do you need packaging and 3rd party software?"
[21:46:17] <alanc> no, that's the "SmartOS" motto
[21:46:41] <richlowe> well, not really, because they don't build appliances (exactly) and do need it (in customer environments)
[21:46:42] <tomww> well, that only applies if it is really an applicance with so little extra software on it
[21:48:03] <Savis> any of you guys ever worked with nexenta?
[21:48:15] <richlowe> the people or the product?
[21:48:21] <Savis> product
[21:48:35] <Savis> im trying to create a cifs share thats acceable to everyone with no username and password
[22:00:55] *** gwr_ has joined #openindiana
[22:01:28] <gwr_> Savis, someone said I should come look in here...
[22:01:34] <gwr_> no scroll back, sorry.
[22:01:47] <Savis> what?
[22:02:03] <richlowe> Savis: gwr knows about nexenta stuff.
[22:02:16] <richlowe> and cifs in specific.
[22:02:19] *** rame has joined #openindiana
[22:02:36] *** smrt has quit IRC
[22:02:54] *** smrt has joined #openindiana
[22:03:00] <Savis> sugestion on how to make a 100% open share ?
[22:03:04] <Savis> no username and password needed?
[22:03:53] <gwr_> zset set sharesmb=name=foo,guestok=true
[22:03:58] <gwr_> should do it.
[22:04:20] <Savis> so just login ssh and do it?
[22:04:27] <Savis> from the console with expert_mode = 1
[22:04:49] <gwr_> Oh, and the "guest" account has to exist, both in /etc/passwd and in /var/smb/smbpasswd (I think - off the top of my head)
[22:05:04] <Savis> hrmm where do i check that?
[22:05:08] <gwr_> which version do you have?
[22:05:40] <Savis> 3.1.2
[22:07:06] <gwr_> You can use the UI to create a local user "guest".
[22:07:19] <Savis> and i have that
[22:07:23] <Savis> do i want to clear out the pw
[22:07:25] <Savis> make it no pw
[22:07:37] <gwr_> Or do the expert mode thing if doing so doesn't violate your support.
[22:07:53] <Savis> naw ill use the gui if i can
[22:08:00] <Savis> so i wana blank out the pw right?
[22:08:05] <gwr_> I think when we tried it, the guest p/w didn't matter.
[22:08:14] <gwr_> because 'guest' is specia.
[22:08:16] <gwr_> l
[22:08:30] <Savis> i wana make sure to check unix user
[22:08:34] <Savis> so it creates the user
[22:08:40] <gwr_> smbd just wants to know if the guest account exists and is enabled for smb
[22:08:51] <Savis> UNIX User
[22:08:52] <Savis> If unchecked, local UNIX user will not be created.
[22:08:52] <Savis> Note: UNIX and/or LDAP users are required for integrated NFS/CIFS ACLs to work properly.
[22:09:00] <Savis> so i wana make a unix user
[22:09:04] <gwr_> yes.
[22:09:29] <gwr_> has to be local.
[22:09:34] <gwr_> not nis or ldap
[22:10:28] <gwr_> Later versions are supposed to have a check box for "allow guest", but I don't remember which release that goes into.
[22:10:54] <gwr_> hit up your support folks for that :)
[22:14:07] <Savis> ok
[22:14:11] <Savis> so i created the guest user
[22:15:19] <crsd> doesn't it work without the local "guest" account, just using the everyone@ ACE?
[22:15:53] <Savis> the thing is i dont want to have to put a user and password in
[22:16:09] <Savis> i can never get domain membership to stay for more than a couple weeks with this damn thing
[22:16:23] <Savis> so i have a device that i have pluged in to the server that we run nexenta on
[22:16:42] <Savis> and i wanted to just share the external drive with cifs and the backups could be coppied there for a lcoal copy
[22:17:03] <Savis> no one will know its there
[22:17:09] <Savis> i just need a place to store this stuff lol
[22:17:27] <Savis> itso n a external usb drive pluged in to the nexenta box and i wana share it with no restrictions on permisions.
[22:18:43] <gwr_> crsd, the authentication (as guest) fails unless the guest account is enables, so you don't get as far as evaluating an @everyone ACL.
[22:18:50] <gwr_> enabled
[22:19:15] <gwr_> what happens to the domain membership?
[22:19:20] * crsd doublechecks
[22:19:22] <gwr_> Does re-join fix it?
[22:19:34] <gwr_> is it clock skew?
[22:20:01] <Savis> no
[22:20:13] <Savis> krb5 has never worked right
[22:20:20] <Savis> i have allways had to edit the template file to make it work.
[22:21:19] <Savis> ok so
[22:21:23] <Savis> im going to create a new share
[22:21:28] <Savis> lets see what happens lol
[22:21:40] *** worsoe has joined #openindiana
[22:21:53] <gwr_> huh. the default krb5 works OK in our testing (use DNS to find the KDC, etc.)
[22:23:25] <richlowe> whether the KDC works depends how you configure it though
[22:23:50] <crsd> it works for me without creating any 'guest' accounts, -rwx------+ 1 2147483649 2147483650 0 Feb 23 01:23 New Text Document.txt
[22:25:32] <gwr_> Thanks. Savis, as it happens, we're trying to get a nexenta page going on stackexchange:
[22:25:32] <gwr_> http://area51.stackexchange.com/proposals/38495/nexenta?referrer=ywdaq4rD-DIKKSRJLXXTEg2
[22:26:01] <gwr_> If you could add a question re. guest OK, crsd and I promise to get an answer for it up there too,
[22:26:19] <gwr_> and that will further our purpose of getting the group past the initial stage.
[22:26:34] * gwr_ can't find the shift keys
[22:29:22] <Savis> bah this still isnt working :(
[22:30:05] <gwr_> crsd, you in domain mode?
[22:30:10] <crsd> yes..
[22:30:18] <crsd> that's probably why?
[22:30:53] <gwr_> It may be easier to get "guestok" configured in workgroup mode.
[22:31:23] *** DwD__ has quit IRC
[22:31:26] <gwr_> We obviously need to put better notes on this somewhere.
[22:34:44] *** bens1 has joined #openindiana
[22:36:15] *** enmand has quit IRC
[22:37:04] <Savis> hrmm no log is being generated in the messages...
[22:37:55] <Savis> its verry simple all i wana do is create 1 folder here on the box that is free to read and write from with out any account info
[22:40:41] <gwr_> When I go to debug this problem, I enable debug logging in syslog.conf
[22:40:57] <richlowe> write it up for the wiki! :)
[22:40:58] <gwr_> then look at what happens when someone tries to connect to the share.
[22:41:05] <richlowe> unless you don't have an account, in which case you're sorta screwed.
[22:41:34] <gwr_> yeah, we should do that too.
[22:41:40] <Savis> its on now
[22:42:47] <Savis> Feb 22 13:42:16 SERVERNAME smbd[19440]: [ID 812811 daemon.notice] logon[DOMAIN\guest]: WRONG_PASSWORD
[22:43:17] <gwr_> is there a domain guest account?
[22:43:20] <Patrickdk> man, it's still moving this 100gigs over to that disk
[22:43:37] <gwr_> you want it to use the local guest account, as I recall.
[22:43:38] <Patrickdk> I guess the good thing is, 50gigs of log data in mysql is compressed down to 10gigs
[22:43:42] <Savis> should be disabled
[22:45:01] <Savis> guest account is disabled
[22:54:35] *** Vutral has quit IRC
[22:56:21] *** Vutral has joined #openindiana
[22:56:30] *** bens1 has quit IRC
[22:59:05] *** MikeCronn has joined #openindiana
[23:00:17] *** DocHoliday has joined #openindiana
[23:01:00] *** raichoo has quit IRC
[23:02:43] *** raichoo has joined #openindiana
[23:02:48] *** laogrymt2_ has quit IRC
[23:03:58] *** rlaager has joined #openindiana
[23:05:58] *** worsoe has quit IRC
[23:06:28] <rlaager> I'm trying to use an FTDI-based USB-to-Serial converter in OI. It's getting misdetected and using usb_mid. I used prtconf -v to get the USB vendor and device IDs. Then I ran: update_drv -a -I '"usb403,6010"' usbftdi. I got a message about not attaching the driver. Also, it shows up in prtconf -D as using usbftdi, but prtconf -v says "driver not attached". I get no new entries in /dev. Replugging it doesn't help. Nor does rebooting. Ideas?
[23:08:52] *** DocHoliday has quit IRC
[23:11:24] <Triskelios> check dmesg
[23:11:32] <rlaager> Triskelios: For what, specifically?
[23:11:40] <richlowe> whining that probably mentions the driver name
[23:16:49] <rlaager> Now that I've mapped it to usbftdi (using update_drv), we're not getting anything that appears useful. Here is the complete output from a replug: http://www.pidgin.im/nopaste/148
[23:17:06] *** laogrymt2_ has joined #openindiana
[23:19:43] <Triskelios> I suspect the chip is providing a bit-bang interface rather than the conventional UART
[23:19:46] *** Kurlon has quit IRC
[23:20:13] <richlowe> I've also seen usb_mid being stubborn as all hell once it was hold of device
[23:20:24] <richlowe> replugging was certainly not sufficient to make it let go.
[23:31:41] *** c2n2 has quit IRC
[23:37:15] <Triskelios> ah, we also have a PID list in the driver, and a uftdi_attach_unrecognized flag (why?)
[23:39:55] <richlowe> isn't uftdi the one marsland wrote because a toy of his needed it?
[23:40:40] *** raichoo has quit IRC
[23:41:14] <richlowe> the answer is "Yes"
[23:44:44] *** timclassic has quit IRC
[23:52:52] *** laogrymt2_ has quit IRC

   February 22, 2012  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | >