DottorZero: nope, unfortunately

ok

sstallion yawns

I don't understand why the debian guest starts with the / in readonly

i blame linux

but that's not a bad default to those sorts of questions

"why does Linux do <some retarded thing>?" "because it's linux, it's retarded like that"

;)

shuddup e^ipi

:P

the same could be said for linux able to read zfs

:)

the code is out there, some of it is GPL too

That's not retarded, that's just NIH syndrome

btrfs right.

I don't understand advantage on using zfs for guests

i meant the GRUB implementation of zfs

using files it's more compatible

some day, someone who won't care will just add it

the install guys hope to have some degree of ext3 support long term

DottorZero: snapshots. checksums & associated self-healing.

compression

(so at least it can detect ext3 filesystems during install, which is a step in the right direction)

then it's cool zfs

DottorZero: are you talking about zfs zvol-based vs. filesystem based guests (for say, xen?)

i have an xvm domU that I use for testing stuff, because I can snapshot it, boot it, BFU it, and then rollback when i'm done playing

timsf, yes

in which case, the answer is "one less layer to go through".

- if all you're doing is block operations, you don't need a filesystem.

[ and you still get all the cool stuff e^ipi pointed at ]

the world is readonly

Use a larger calibre.

hi

i get this bad pbr sig when i boot off my computer

how do i fix it?

it's a desktop this time

new 2005.08 install

er 05

good night

the disk is serial ata does that have anything to do with bad pbr sig?

no, it means the installgrub got messed up somewhere down the line

is there a "solaris for linux admins" site anyone can point me to?

http://bhami.com/rosetta.html

Wasn't there an immigrants page on os.org?

and http://docs.sun.com/app/docs/coll/47.16

thanks...

siezer: The Admin Guides are quite good. Definitely worth reading.

e^ipi so it's not hardware issues, just installgrub bug?

i didn't say that

just that for whatever reason it didn't take

could be a bug, could be hardware

it is a mystery

anybody know if there's a package for screen?

YES I AM WORKING ON IT OKAY GOD YOU PEOPLE

haha

so about that screen package

e^ipi: pam or kerberos for screen? :)

and how do I compile the postgres module for python?

nachox: neither

?

SUNWgcc 0/19 281/3101 59.80/175.97pkg: An unexpected error happened during installation: timed out

The Boot Environment opensolaris failed to be updated. A snapshot was taken before the failed attempt and is mounted here /tmp/tmpVjqFnr. Use 'beadm activate opensolaris_static:-:2008-08-27-01:47:27 and reboot if you wish to boot to this BE.

wtf?

I keep getting that message

welcome to the cursed world of pkg

nachox: everything's managed through ~, as per ARC

i remember reading the case in opensolaris-arc, i dont know what happened in the end though...

hmm

I don't know why it's timing out

i got busy, and then had to spend some time with family

hehe

I'm a solaris noob

Blackknight: what version of OpenSolaris are you using?

Blackknight: i assume it's 2008.05? without any modifications?

yeah

I just installed from the cd

Blackknight: read http://www.opensolaris.org/os/project/indiana/resources/rn3/#Update_Inst

you might want to get a newer pkg

you need to do the first 3 commands first as listed in that first box

ok

then it'll fix the timeout issues probably

unfortunately it's a known bad bug

we hope to have updated 2008.05 isos available soon to fix this

Gman_: any eta on that?

jbk: hopefully next week

bit slower than we hoped

hmm, just installing gcc worked

Poll: The SFW Build System: Black Magic, or Voodoo?

heh

I blew up my first install so bad it wouldn't boot

Blackknight: you came across another bug

just install solaris 10 and be done with it :)

aren't they the same?

needing to run update-grub after you image-update

of course not

sort of

and does 10 have zfs?

yes, but not as the default filesystem

yes, but you cant boot from zfs

opensolaris is pretty nice once you can avoid those bugs

yeah, I'll stick with it

it's just a few steps you need to do to get from 2008.05 to a newer build

be nice if there was a minimum install option

i'll find you a mail so you can follow the steps

and remove that stupid path choise, but i agree in general, it is nice

the default bash prompt sucks btw

http://mail.opensolaris.org/pipermail/indiana-discuss/2008-August/008289.html

is a newer iso

there's a link to it from genunix.org

build95 iso, it'll get you past the bad bugs

dropping time-based releases in favour of random releses then, are we?

;)

schedules are made to be broken

heh, no.

the nice thing about deadlines is they make a pleasing whizzing sound as they pass by

:)

Gman_: it would make for an innovative new release schedule

jbk: hahaha

(i'm quoting or probably paraphrasing, so it's nothing original of my own :))

oh, and WHY does nssswitch have dns disabled by default?

resolv.conf needs to be set too

not "when it's done" or "every 6 months"... but rather "you'll get it... we don't know when, you don't know when, and it has nothing to do with what's in there"

chances are, people who can find #opensolaris are the people who will likely want to go onto the development train

server is kind of useless without dns

Blackknight: because indiana uses nwam

what's that?

an automatic network configuration thingie

http://opensolaris.org/os/project/nwam/

I couldn't even get pkg search -r to work

it should be nice once it's done

http://blogs.sun.com/observatory/entry/local_hostname_resolution might help you

hmm

I got it to work, just used to centos

it's a bug, another known one

thinking about putting the install on a real server

anything special I should do when compiling python modules?

cassini is NOT supported for IP Instances, correct?

probably not until they put in their hacks to work around not porting it to nemo

or until me & sstallion finishes our work

ld.so.1: ld: fatal: libld.so.4: version `SUNWprivate_4.2' not found (required by file /usr/bin/ld)

I installed SUNWtoo

sstallion wakes up

Blackknight: if you want to compile things, you'll need studio-dev or gcc-dev... i reccomend studio-dev because the compiler is just better

jbk: ?

cge

ahh

there was a question about ip instance support w/ cassini

hile_: like jbk said, likely not since cge implements DLPI directly

s/cge/ce/

e^ipi, does studio-dev need a gui?

I believe IP instances requires GLDv3 support at a minimum

good question, no idea

the compilers don't

I totally removed the desktop

jbk: speaking of which, I've finally puzzled out the DMA issues on ne, and came across something funny as hell in the specs

that seems a silly thing to do, you know you can just disable it right?

ne2000 devices should be able to support vlans... cracks me up.

yeah but it wastes space

heh

I'm used to servers with ssh and that's it

at 12 cents per gigabyte, who cares?

well maybe he's using EMC disk

which is probably still around $30/gb

jbk: done much with scatter/gather ?

nah, just vmware

we do have coraids at work

not really, but am familar with it

and congrats on being the only OS that supports iscsi 3

jbk: I'm in the same boat. Have worked with it a bit in the past, but never wrote it from scratch. Wondering if I can make use of it for this.

sstallion pokes gdamore

ok, img-update wants to update stuff that's not even installed

"for i in `pkginfo|grep staroffice|cut -d" " -f2 -`;do y|pkgrm $i;done"

coffman waits

ok

so i raised a CR for the <math.h> header files and it will be fixed

s/files/file/g

can you get a complete netstat connection count for global + zones?

netstat -an won't display connections open in a zone if done from global

Where is a good place to read what I am missing by going "open"

what?

do i miss the performance/goodies

if i go opensolaris rather than solairis

you miss some legacy stuff, mostly

so is there any way to get a complete connection list from netstat for all the zones+global?

'morning to all...

victori_: are you trying to kill off old tcp connections or something?

codestr0m: no just to get a view of how many connections open total

since I have web/mail and global segmented

is this to just collect data or solve a problem?

just to collect data, overview of what is open

well. if you're running two front-end nodes are reverse proxies to a large batch of java backend nodes I could understand, but otherwise.. not sure your topology or why care.. let me know if you figure it out though

(as a side question.. are you going to kick the collected data off to a wicket app for presentation?)

Coffeeeeee

victori_: iirc you're using haproxy.. you may look there to get the stats you need since that may be where most of the connections are coming from.. (a lot of guessing on my part)

codestr0m: well I wanted to generally see how many tcp connections open from the OS's perspective.. postfix, spamd etc into account

imap*

codestr0m: not using ha-proxy but perlbal with patches

codestr0m: what data do you need?

aha.. ok. I was mistaken.. imap is evil and if a lot of users you may want to consider a failover situation.. spamd (I'm not sure this will be a remote connection) I'm still poking around with 2008.05 so not sure where all the things are hidden..

in linux I'm trying to remember, but I think this is kept in proc as a worst case..

not sure if netstat would give what you need

codestr0m: I am going to do a write up on what I learned. The Java way of hosting all your applications on one application server is stupid. JVM crashes or you get rocked by heap collection. Best to run 3 jvm instances and just load balance with a smart proxy. Have a background script kill any of the 3 jvm instances if they get past x amount of RSS ram

victori_: Uhm. Wouldn't it be better to limit the RSS with resources management?

trochej: ah not a limiting issue but to avoid full GCs, it is more efficient to just kill the JVM instance

victori_: umm.. I'd counter this, but only just a little.. and I haven't ran any big production apps on a node with zfs yet, but my gut feeling is I'd still prefer to run all bl nodes diskless and pxe boot

bl nodes?

victori_: if you're hitting full gc and it's causing a world stopper on your application that's 100% something you should 1) possibly tune jvm options for 2) look at fixing your app 3) increase the amount of heap available

bl == business logic

victori_, -Xmx512m doesn't do what you want?

oh and externalizing all your cache and sessions into memcache is a big win. You can do restarts without giving your users expired screens

smtms: he's using wicket.. and does server side page state == lots of ram

we run ZFS quite happely on our UAT systems

smtms: ya it gets limited correct, however once your tenured generation gets filled up, GCs are expensive and it is just easier to kill and restart the JVM then let the GC wrestle with the garbage at that point

WickedWicky: how much disk activity and regardless.. for large clusters the power savings to pull the disks is always worth it

specially when your getting hammered on the web app

victori_: something isn't correct in your app if I had to make a blunt guess

we're running Weblogic + Oracle on it, with 1,5TB+ data

victori_: RM serves exactly that reason: to better distribute resources among applications and to avoid killing machine by a rouge application WITHOUT killing it.

rm?

Resource Management

victori_: Read on it.

In Solaris you can control more things than with ulimits on Linux

Much more

trochej: oh no it has nothing to do with the system stalling to the halt but the JVM instance due to garbage collection when the tenured generation fills up

trochej: with rm you still can't magically expand the java heap and that's what he's working around

codestr0m: Uhm. Isn't jvm supposed to take care of this?

yu dont /cant always simply extend the heap size

codestr0m: Yes, I didn't catch the beggining of the thread, my fault.

oncwe you've reached -mxmax you're pretty much stuck with gc

victori_: what's the max size your tenured generation will grow to? are you (ab)using hibernate.. ?

codestr0m: I looked around, I have a file httpsessionstore plugin for jetty written, all my sessions are small. However once you get hammered with a lot of requests the GC just can't keep up. Best to run multiple JVMs

anyway if you were to have a single application server instance, you miss out on rolling updates and such. Taking down the service for a minor update sucks.

victori_: umm.. multiple jvms across multiple physical machines and I fully agree.. same box and I say something isn't designed correctly or possibly can be fixed

victori_: sounds like you're gonna need more servers with rmi enabled :P

WickedWicky: rmi is evil.. yuck

WickedWicky: it is holding up well, 3 jvm instances on a single machine

it is but it can save lifes when you set it u properly

yea but during garbage collection the JVM wont serve requesta, how is this "holding up well" ?

while this is interesting.. victori_ you still missed my question. of what's your max tenured size? are you using hibernate? (as in have you ran jstat and watched things.. ?

codestr0m: under steady traffic the jvm can be up forever; the GC can keep up. It is just during digg effects where it becomes problematic.

256k thread size/ 64meg permgen / 1024meg heap (when running single instance)

now I just do 3 instances with 384meg heap

victori_: have you pinpointed what's causing the rapid growth? I mean otherwise. you should have scalable or predictable growth.. as in if a single machine can't handle it. it simply can't handle it w/o adverse effects..

victori_, I can't imagine how 3 JVMs fare better than one

he's trying to work around the JVM not serving requests when GC kicks in

correct

and I get the benefit of rolling updates

victori_: there's more than just permgen and heap.. run jstat. give whatever gc that's filling up some ridiculous amount and profile your app locally if you can..

perlbal keeps track of response times and sends clients to the fastest possible instance.

no sticky-session hacks

victori_, killing a JVM has no bad effects for your users?

everything nicely distributed

correct

victori_: you've moved all state into memcache?

app cache/hibernate 2ndlevel cache and httpsessionstore all in memcache.

pretty much, works wonderfully

if your sessions are larger than 500kb your not doing it right in wicket

that's good to hear.. I suggested almost exactly this and got stupid arguments over a year ago

I absolutely love the setup I have now, before this I did not know if I could scale out site

victori_: well. by /doing it right/ I would generally agree.. however, one of the clients allowed sessions to never time out with a js screensaver.. the sessions would grow and grow for days. it was evil

victori_: btw.. did you document this and is all the bits in public or are some of your patches just local?

perlbal = best proxy I have used so far, the fact it keeps track of response times and sends requests down to the fastest backend. It is also able to retry on http 500 responses.

codestr0m: not documented anywhere. I will have a write up + patches posted on my blog

jetty-memcache will be contributed to jetty7's trunk

cool. if you can ping me with the link that would be most appreciated

is this wicket 1.3 or 1.4.. ?

1.3

victori_: doesn't haproxy do that too?

or am i on crack?

does it? I though it was just sticky sessions with cookies or round robin, with fail support if a backend is down

even if it does, perlbal is configurable via plugins and it has internal redirects.

*shrug*

i know very little about webby stuff

I played with haproxy, perlbal gave better results in my testing

Coffee

hi all! Does anyone here know if Solaris10 Zones support ipv6 correctly?

Does anyone here know if Solaris10 Zones support ipv6, i.e if there are any major issues with IPv6 support in solaris containers?

bishamonten: I'd suspect so

trochej, I would think the same too. I'm currently having an argument with a colleague who insists that it's 'broken', but won't specify how. As far as I can see it's fine.

But, deeper research always helps.

bishamonten: solaris won't assign a link-local address to the zone automatically

bishamonten: this breaks router discovery in a rather non-obvious way

Hi is there any leader of the PKG project?

comay?

e^ipi: that was for me?

_mary_kate_: Thanks. Would the workaround for this be to statically assign the link-local address? If not, is there another more acceptable workaround?

bishamonten: that is the workaround

well, i don't think it's a workaround so much as the way you're meant to do it

_mary_kate_: * whew * Thanks :-) That helps!

Has anyone here run Linux in Solaris BrandZ virtual machines? Is this reasonably functional at this point in time?

it does everything you would expect from RHEL 3 with a 2.4 kernel

(there's some 2.6 work going on, but not useful yet, i don't think)

_mary_kate_: Would I be able to install a debian (or debian-like) linux in a BrandZ zone, or am I limited to a RHEL?

3.1 works, 4.0 might work, testing/stable/the next release definitely won't

er, testing/unstable

_mary_kate_: that's good enough for me. I'm intending to install an Ubuntu version in BrandZ on Solaris10.

only a very old ubuntu would work

i think they added the 2.6 dependency before debian

bishamonten: You'd get more luck with stable DEbian.

_mary_kate_: that's a little concerning, not a major train smash though.

brb.

trochej: This would be an option.

bishamonten: brandz emulates linux syscalls. It pretty well emulates 2.4 kernel line, but 2.6 is a work in progress IIRC.

trochej: would you say there are functional 2.6 kernels with a degree of buginess or there are no functional 2.6 kernel lines supported?

hello all, anybody knows if blastwave users list is active yet?

bishamonten: IIRC, you'd have to get sources for the 2.6 brandz branch and compile it yourself

gerard13: ask in #blastwave - prolly more answers

goodmorning

\_o<

Coffeee

hallo

hi. i'm having an error building a particular C++ file, which refers to stdexcept, can I ask for help or this is too off topic?

i'm using forte and sun studio

Mojo_risin: if this is studio, make sure you use CC -library=stlport4 to get a conformant C++ library

(but be careful, this changes the C++ ABI)

the error is this one: http://rafb.net/p/wnAGvJ17.html

_mary_kate_: i don't know if I can do that because this has to be compatible with software compiled by 3rd party

_mary_kate_: if i don't use -library=iostream i get this error:

http://rafb.net/p/Be87Ls50.html

trochej: thanks for the clue.

Np

One more question, and I appreciate that this one should be backed by heavy research on my part, but: Would you know if opensolaris/solaris is anywhere close to a coherent package management system in the vein of apt-get or the freebsd ports system?

bishamonten: Depends on what you're looking for.

bishamonten: IE. patching and backing out patches works in Solaris/OpenSolaris using sysvr packages

apt-get doesn't do that

bishamonten: On the other hand, ips is a work in progress to bring goodies od automatic deps tracking, online mirrored repositories and such.

trochej: The first requirement would be for security updates and patches provided by opensolaris community, Solaris sysvr provides that, as you say. The other, lesser re requirement is dependency management of mainstream packages, in the way apt-get does it.

trochej: "ips" ... I will look that up.

bishamonten: IPS is the package management from OpenSolaris 2008.08 and upcoming 11 (previously project indiana).

trochej: So, pretty much still beta. I guess I can't bank on it just yet.

bishamonten: I wouldn't either.

trochej: * sigh * :-(

is it recommended practice to update from 2008.05 b86 to a more recent build if i just wanted to run a headless box serving nfs? how's the reliability/stability of the new builds compared to the initial one?

trochej: do you have an idea of how long Solaris Virtualisation has been around versus somethign like VMware ESX?

bishamonten: do you mean zones (containers) or xVM (xen)?

_mary_kate_: Containers.

since solaris 10

_mary_kate_: So, around 2 years?

3

for customers, longer in development and beta testing

_mary_kate_: Even better. I'd estimate 5 to 6 years total. how long would you guesstimate VMware ESX then?

i don't know anything about vmware

Anyone know if SXCE b97 is comming out in the next 24h before I make my trip ?

I want to download the latest version before I make my tip and 97 has some changed I need

hello

i am assuming there is a pain free method of upgrading between sxce builds?

if so, can someone point me in the right direction

i dont want to have to download another dvd if i can avoid it :)

botox:Hi, check out live upgrade

liveupgrade

is that just a command i can run or something larger?

you still have to download the dvd, but you don't have to do a proper install

ok

lu*

thanks for that

i have another odd issue. my mirrored zfs pool (2 70gb fc-al drives).

now,

one of them failed

weirdly, it lists 2 of the same drive in the zpool status

there is only currently one drive plugged in

`pfexec fmadm faulty`

a replacement is coming in the post.. i assume i can just scrub the pool, then add the replacement drive as part of the mirror?

Should tell you what to do

evocallaghan: that for me?

yes

botox: you shouldn't need to scrub. just replace the disk and use 'zpool replace'

it'll resilver automatically

_mary_kate_: will this also remove the fact that it has the current working drive listed twice?

i don't know why it would do that. can you paste your zpool status at rafb.net/paste?

_mary_kate_: sadly not at the moment, having trouble getting into the box.. just tried.. in a few hours i can though so i will try to beckon you later ;)

I guess i'll start my download of SXCE 'just' a few hours before I leave. With some hope, 97 will come out tonight. fingers crossed !

if i remember rightly, the pool is simply c01... listed.

rpool

MIRROR

but the same drive listed twice, with one of the drives marked as

degraded

but, will paste when i get access

do we know when another release of opensolaris might be announced?

not the sxce's.. another actual release..

Hello, is it possible to install OpenSolaris to VmWare? http://dlc.sun.com/osol/opensolaris/2008/05/os200805.iso doesn't work and outputs "bad PBR image"

yes it is

http://opensolaris.org/os/article/.../

botox, ok, so what could be wrong when I get bad PBR image?

check vmware version

you need

Now check "Sun Solaris" and select "Solaris 10 (experimental)" or if you are lucky enough to have a 64-bit machine "Solaris 10 64-bit (experimental)".

to select that.

botox, thank you

my pleasure

good morning

ImportError: ld.so.1: python2.4: fatal: libpq.so.5: open failed: No such file or directory

anybody know how to fix that?

nevada?

open solaris, I just compiled the pygresql module

pkg search -r libpq.so.5

postgres is installed

problem is LD_LIBRARY_PATH doesn't find it

ahm

trygvis: pfexec crle -l /lib:/usr/lib:/usr/postgres/8.2/lib

oops

Blackknight: try that

bah

don't do that

rebuild the library

and add '-R /usr/postgres/8.2/lib' to the link command

or

now that i think about it

see if elfedit exists

on your system

if so

lemme dig up the command line

run 'elfdump -d /path/to/python.sofile | grep RUNPATH'

if anything shows up, take that value and append ':/usr/postgres/8.2/lib' to it

(think of 'RUNPATH' as a library-specific LD_LIBRARY_PATH that's part of the shared library file, so you want to take the existing value (if any) and add '/usr/postgres/8.2/lib' to it

take that new value

and run elfedit -e 'dyn:runpath NEWRUNPATHVALUE' /path/to/python/sofile

copy the original version first of course

http://blogs.sun.com/ali/entry/avoiding_ld_library_path_the has a more involved discussion

fyi, the way to avoid having to do all that is adding the -R /library:/search/path option to the link command -- it'll combine all the -R values into the runpath

(that also works on linux, just usually needs a few more flags, so you don't have to muck with /etc/ld.so.conf)

ok, thanks

I'm not sure how to change the link command, it's part of the setup.py

/usr/lib/python2.4/site-packages/pgdb.pyc is the module

Blackknight: it should create a .so file

somewhere in there

hi all

can anyone tell me how to add a user to sudoers ?

is there a file named sudoers to edit ?

did you ask google?

yeahp

[root@opensolaris] ~ > elfdump -d /usr/lib/python2.4/site-packages/_pg.so | grep RUNPATH

[4] RUNPATH 0xb4b /usr/ccs/lib:/lib:/usr/lib:/usr/sfw/lib

there it is

but /usr/postgres/8.3/lib isn't there, hehe

so run elfedit -e 'dyn:runpath /usr/ccs/lib:/lib:/usr/lib:/usr/sfw/lib:/usr/postgres/8.3/lib' /usr/lib/python2.4/site-packages/_pg.so

so SunTzuTech ?

and if you run elfdump again, and it should be good

i tried googling but i didnt get it on the first page

that worked, thanks

and am really lazy to go beyond that

question is why do I need to change it in the first place?

i can't have a Sun M9000 server ( x4 cores, x2 threads per core, sparc64 vii ) ):

guess that's what happens with packages in non-standard locations

im_alone: nice boxes.

oussema, visudo?

rmesta is back (gone 06:16:26)

kind of wish I could get sqlite to work

BlackKnight : it didnt work

hmm

I know in ubuntu all I had to do was edit that file

does apache have a file for SMF?

is it normal that opensolaris does not add a new home when i create a new user ,

oussema, a new user doesn't implicate to have its home

useradd with the -m flag will make the directory for you.

# mkdir /home/<user> ; chown <user>.<group of user> /home/<user> it's all

RESOLVED thank you

Hi, is there a way to disable swap entirely during the installation? i.e. I want opensolaris to work without a swap or dump volume and to use a read-only flash storage for the system. It looks like this is not possible according to http://opensolaris.org/jive/thread.jspa?messageID=262129 but I was wondering if things changed since then

Hmmm also, where can I get SXCE96? In the topic I read that it is available but I could only find the build 95

nevermind,found build 96.... sorry

uski`work, Try the swapoff command

Hi. I'm trying to install a minimal opensolaris server, and I'm not quite sure what package list that amounts to. I don't see anything like solaris10's core package. Is there a meta package that would do what I want? (I'm installing into /a on a nevada machine, mostly cribbing from http://blogs.sun.com/edp/entry/moving_from_nevada_and_live and pkgcreatezone)

slim_install sounds right, but I think it's got a lot of stuff I don't want in it. (like X, this is for a headless server)

good evening every1

any body knows how to assign sudoing right to a user ?

evening pavel

oussema: why not use pfexec?

sudo suxxx!!! use rbac!

hi th :-)

hey Berny

an easy way to learn how is www.c0t0d0s0.org

hmm, I'm getting a "connect retrieve manifest error" when issuing the third command.. pfexec pkg install entire@0.5.11-0.86

hm?

rorx:Upgrade your ipkg version

actually the error is "cannot retrieve manifest entire@0.5.11%2C5.11-8.86%3A20080604T12543Z

pvelj: i dont have sudo

evocallaghan: isn't that the second command?

and i couldnt figure out how to use pfexec

if i want to create file /test

isn't rbac already inside?

should i execute : pfexec -P ALL touch /test root ?

what's rbac

am totally new on solaris

am used to linux

obviously.

:-)

to be honest ,i am just using opensolaris to try the zfs and zone feateures

it's much easier to handle RAID5 with opensolaris

through ZONE

try to to run the roleadd

delewis:Hey!

Have not seen you for ages !

and RAID6?

dude

hey ppl, i've just discovered the SunVTS

anyone with experience using it?

anyone?

pavelj: yes, it has been around for a very long time... we use it in the factory

well, timing is everything...

and patience

I was gonna give seph the command to remove gnome, hehe

1) Is it possible to setup an iSCSI target in OpenSolaris?

2) I already have a ZFS CIFS share - can I make THAT my iSCSI target too?

hello

RavenSlay3r, it's easy, to create a big file filled of zeros and to mount loop it for iSCSI target

RavenSlay3r: yes, and not afaik

hmmm.... might not solve my problem then..

need vBox to have FAST access to shared data...

nothing about vbox is fast

e^ipi: I finally got Solaris file-sharing working - thanks for all your help with that! :) , the vbox networking is slow though.

Is it possible to use ZFS CIFS stuff without an AD on your network?

e^ipi: compiling my project on the laptop take 10 min., in vbox on the Ultra-40 it took 3min. - now if it could do that with data on the LAN i would be VERY happy!

postwait: yes - it works with Windows Workgroups nicely, once you beat it into submission :)

postwait: http://blogs.sun.com/timthomas/entry/solaris_cifs_in_workgroup_mode