IRC Logs

Network: freenode Channel: #solaris

	Enter your search terms Submit search form
	Web www.irclog.org

[19:11]<rrfv>geekboi, probably another no-knock warrant
[19:14]<ygsjdgyr>http://www.cnn.com/2006/TRAVEL/05/26/dollywood.gas.ap/index.html
[19:14]<ygsjdgyr>Dollywood pwnz!
[19:15]<ygsjdgyr>kerouac: hi!
[19:23]<rnnc2jy>galt: On a federal building?
[19:25]<rnnyn>Good afternoon. On Solaris 10, if I assign a role profile to a role/user, I can perform administrative actions directly from pfsh, but if I start bash and work from there, I get permission denied?
[19:29]<rrfv>geekboi, that's what the big stink is, the FBI did a warrant on a representative's main offices
[19:30]<rrfv>presumably, the FBI now thinks that since that was successful, they can do no-knock warrants now, which are invariable preformed by SWAT
[19:33]<rnnyn>galt: was it ?
[19:34]<rrfv>azzie, they had access to the docs for a full two days before they were sealed
[19:35]<rrfv>and if they do a no-knock on friday, that means they get the whole weekend before they're ordered not to look at the docs
[19:36]<anzjura>hi insomnia
[19:36]<ajxrpr>Hey guys!
[19:36]<ajxrpr>How would you compare Fedora Core with Solaris? Or where would you tell me to look for a comparison?
[19:38]<rnnyn>johnfg, you don't. Solaris is better :-)
[19:39]<rrfv>I wouldn't. I'd hurl fedoira core with great force from my server room
[19:40]<ajxrpr>Well, that's what I'd like to find out. I've been using FDS, and noticed that a number of other FDS users are running solaris.
[19:40]<ajxrpr>How long has solaris been open sourced? Didn't it used to cost?
[19:41]<rrfv>if he asks what linux kernel it has I swear I'm kicking him
[19:42]<ajxrpr>Now galt, don't be that touchy. I wouldn't do such a thing :-)
[19:42]<ajxrpr>I'm downloading the iso's now, as we chat.
[19:44]<ajxrpr>Are any of you using FDS?
[19:44]<ajc_z0>johnfg: Like with any OS comparisons, you should consider not only desired functionality, but also design goals and of course TCO
[19:45]<ajxrpr>Right.
[19:47]<pnzvnzb>nah, nobody uses TECO any more
[19:48]<ajc_z0>TCO includes not just hardware, software and maintenance, but also admin cost
[19:49]<ajc_z0>To save you the effort, your best choice is FreeBSD running on decent but cheap peecees
[19:49]<ajxrpr>What do you think of this quote from linuxforums: We do have option that is far superior than Linux; the UNIX Solaris 10, developed by Sun Microsystems, where its performance and reliability is unmatched. I installed Solaris 10 on my PC and never look back for Linux and I bet in the future the Solaris will gain more popularity among PC users. For now, just one obstacle for Solaris - the installation and configuration process is a bit complicated compare t
[19:49]<ajxrpr>o Linux. And I'm sure this issue will be resolved soon.
[19:49]<ajc_z0>You should ignore it along with all other heresay
[19:50]<rrfv>see what letting X86 into the mainstream solaris does?
[19:50]<ajc_z0>Gets fastr cheaper processors?
[19:50]<rrfv>and faster, cheaper users as well
[19:51]<ajxrpr>Ha! I like that: faster and cheaper!
[19:51]<ajc_z0>The users seem to be getting contiually slower to me
[19:51]<pnzvnzb>I have to agree with AJC_Z0 on that one
[19:51]<lzr1>Is there a way to make changes to /etc/system happen without a reboot? (specifically set rlim_fd_max)
[19:51]<ajc_z0>prg3: Of course. Use the debugger on the running kernel
[19:52]<ajxrpr>Thanks for the opinions guys. Later.
[19:52]<pnzvnzb>I suspect if you increase rlim_fd_max using adb -k, a reboot will follow pretty shortly afterwards anyway
[19:53]<gnpzzyus>how come solaris and sparc hardware are not common in top500.org if it's so fast?
[19:53]<ajc_z0>See mdb()
[19:53]<lzr1>AJC_Z0, Got any docs on that process? or just man?
[19:53]<gnpzzyus>i mean, i know from personal experience i'd much rather have, say, a massive e-commerce site running on solaris with sparc hardware and the comfort zone of a support contract with sun
[19:54]<lzr1>nefarius, because linpack is a useless benchmark of anything useful?
[19:54]<pnzvnzb>unless you happen to be simulating atomic bombs and such
[19:55]<ajc_z0>kadb (-> adb) -> kmdb -> mdb
[19:55]<lzr1>Not even there.. it's a very very specific benchmark, that has pretty much no useful implementation in real HPC
[19:55]<pnzvnzb>ok
[19:55]<gnpzzyus>http://www.mpa-garching.mpg.de/galform/virgo/millennium/ or that or genomic research or pretty much any scientific computing...???
[19:55]<gnpzzyus>i hardly think that's "anything useful".
[19:56]<gnpzzyus>google is not useful?
[19:56]<lzr1>noob-saibot, linpack isn't useful
[19:57]<lzr1>linpack is the benchmark that the top500 list is using for it's ranking
[19:57]<lzr1>umm.. no idea how noob-saibot got in there..
[19:59]<ajc_z0>Lies, damn lies, statistics, then benchmarks
[19:59]<lzr1>Bingo..
[19:59]<lzr1>More useful information about comparing Sparc to others would be the Spec numbers..
[19:59]<gnpzzyus>yes, specint
[20:00]<lzr1>Top500/Linpack goes to the one with the most money, and half of the equation there is the interconnect between processors, not the processors themselves. BlueGene/L is top 5, with 800mhz processors
[20:00]<lzr1>fp_rate, and jbb should show sparc in a good light
[20:01]<lzr1>In the end, the only good benchmark you can have, is your application, running in your production setting
[20:13]<gnpzzyus>http://en.wikipedia.org/wiki/LINPACK
[20:16]<ggjmygmgz>it's so nice to have a real sysadmin job again.
[20:20]<rnnc2jy>man is the best friend of the system administrator, front
[20:20]<rnnc2jy> same the dog
[20:22]<rrmnsm>geekboi, good sysadmins have no friends, just groupies ;-p
[20:28]<gnpzzyus>lol
[20:28]<gnpzzyus>jamesd: s/groupies/users ? ;>
[20:31]<rrmnsm>nefarius, no.. most likely not ;-p
[20:32]<gnpzzyus>:)
[20:48]<rsruvxjp>jamesd: I did another try with pxe boot, I extracter the x86.multiboot image and changed NFS_CLIENT_VERSMAX, gzipped it to the tftpd again - the install worked than from a linux nfsd ...
[20:49]<rrmnsm>gsauthof, okay but did you get a usable install? are there any missing bits?
[20:49]<rsruvxjp>missing bits, yes - but i don't suspect nfs for that
[20:50]<rsruvxjp>the install does not configure grub right
[20:50]<rrmnsm>gsauthof, what the cookie monster took them? solaris nfs install is extremely well tested... because many many sun customers use jumpstart, where the system is net booted and installed automtaticly
[20:51]<rsruvxjp>the installer is not perfect, perhaps?
[20:51]<rrmnsm>nope.. works perfectly.. slow as heck but works perfectly
[20:52]<rsruvxjp>but the grub code is very new, or not?
[20:52]<rrmnsm>its been in solaris express for over 6 months... very well tested including nfs installs
[20:53]<rsruvxjp>so it is perfect?
[20:54]<rrmnsm>the resulting installs allways work perfectly, nfs installs are quite common... linux nfs server is broken... linux does claim to have nfsv4 btw... but its not compatible with solaris nfsv4... and since sun designed nfs .. guess who did it right?
[20:54]<rsruvxjp>what is your theory, how the nfsd bother only the grub config and every other package get installed
[20:55]<rsruvxjp>i used nfsv3 ...
[20:55]<rrmnsm>gsauthof, how can you be sure every other bit was installed correctly? the nfs server is broken...
[20:55]<rsruvxjp>how can you be so sure with that - only because it is not a solaris server ...
[20:56]<rrmnsm>because i did it my self.. and it was broken.. changed to a solaris nfs server, same isos same everything except for the server and it all worked
[20:56]<rsruvxjp>ok, one case proves it
[20:58]<rrmnsm>gsauthof, dude i have been in this channel a long time helping users, everytime there is an linux nfs server used in the install the results are the same.. broken installs.. it could also be the combination of broken nfs server plus bugs in the loopback filesystem, when linked they test things that aren't normally used on linux.
[21:00]<fafv>hi jamesd
[21:00]<fafv>just read your whu solaris, i love it
[21:00]<fafv>s/why/whu
[21:01]<rrmnsm>gsauthof, and most likely i have used linux longer than you have, my first kernel was 1.0.9... so i'm just not trying to bash linux here... its a bug, no one wants to fix it because linux to linux nfs works perfectly the linux developers don't care that the nfs server is broken
[21:01]<rsruvxjp>that is not about how long someone uses linux
[21:02]<gnpzzyus>jamesd: yah, there is a terrible kernel bug with hanging nfs mounts where, if you do a df, that hangs as well and you cannot kill the df process without totally restarting the system
[21:02]<rrmnsm>gsauthof, okay its over 12 years
[21:02]<gnpzzyus>i think solaris' nfs implementation is much cleaner
[21:03]<rrmnsm>nefarius, that really isn't a bug, its by design.. nfs is supposed to handle terrible networking environments, it can handle the loss of a router for days, remember it was designed when the inernet was a collection of 9600 baud lease lines and modems.
[21:04]<ajc_z0>If you want your NFS mounts to be soft or interruptable, then you have to _tell_ mountd

Back to #solaris
Or go to some related logs:
#gentoo
#gentoo
#css
#gentoo
#ubuntu