There is one truth (somewhere)

There is a bug in Zfs which i seemed to hit quite often with Solaris Live Upgrade and Zones, but only recently found.

I hope it gets resolved soon….

Here ist some workaround.

another toy.

no scsi but udma 33, means not that bad as the u10. auction said it would come with 256mb ram, 40gb hd and a dvd rom.

it came with 2x512mb and 1x256mb modules, 20gb hd and a cdrom…. this wasnt so bad after all!

they guy i bought it from shipped it in this box:

 

 i should fill a complain on ebay and use buyers protection or something….

ps: there was also a lsi scsi card inside o0

from the "Sun Open Source Summit" Santa Clara California, USA October 15-16, 2007

some notes (not mine):

10 Ways to Kill Your Community: Josh Berkus

• Be Silent
• Document Nothing
• Stop outside commits
• Obfuscate Governance
• Hold closed meetings
• Legalese
• Encourage Poisonous People
• Make Tools Difficult
• Change Licenses

 "In many ways we’ve all done all of these things"

 Well, in many ways you are still doing this kind of things:

- Indiana name dession

- Last submited CIFS server

- The whole Indiana thing ofc

- To consider even remotly to put opensolaris under two licenses

…

I wish i could truely belive you are going to change the path. Atm this f*cking politics are the major show stoper for me to get involved.

I think i read somewhere Mandriva considers upgrades between releases the bigest issue in linux atm. You have the tool but dont push it to be usefull with your new technology (ZFS/Zones).

Hack. People need to lose the thing they love the most before realise that it was important to them.

ps: maybe im at the tech days in frankfurt on 3. dez (germany) 

since b66 there is a bug that prevents sx running on non sse x86 machines.
this means it does not work on anything older than intel p4 (p3 etc) (yeah, no one wants to use something older than an p3.. or? i think there are a couple embedded system out there) or amd 64 athlon-xp ( K7 thunderbird etc) and ofc older via cpus. this sux!
even more, the bug is not fixed yet although "Commit to fix" is snv_70. now some one please get there ass up and fix this.
seems like sun does not use older machines for testing anymore… now sun, get down in your basement and get some older systems for testing!
such a bug since 8 releases (still no fix in b73 b74 b75 b76 if im right) ? thats horrible! 8 10 releases - thats ~16 20 weeks.
on which machines do you think people will install (open)solaris for testing? on there brand-new workstation? i dont think so.

rant over

PS: some people of the dopsy would love to test sxce on there systems (embedded stuff), but they cant….

Update: Menno Lageman points me at a workaround, which is also described in the bug, he explains how apply the workaround via LiveUpgrade which is cool if you have a

system up and running - in my case this prevents people from doing a fresh (first time) install, which might  cause them to never  look at solaris again…

i maybe will take the pain and edit a iso to fix this, depends on if i find time and how long it will take to get a final fix back in nevada.

Update 2: elfedit in snv_75? dont get me started….

for i in `zfs list|grep $pool/$fs/$fs@$date/description+wildcard|cut -d" " -f1 -`;do zfs destroy $i;done

28 min Vortrag, 16 min Q&A - dafür das ich ganz gut fertig war (so 27 std wach oder so ) wars ganz gut. Also bitte verzeiht mir meine Fehler.
Große Teile der Folie sind geklaut, war halt alles sehr knapp. Aufgrund des positiven Feedbacks werde ich aber ganz sicher noch einmal Arbeit reinstecken.
Folie und mp3 / ogg-vorbis vom Vortrag gibts hier.

Vielen Dank an geht an die netten Zuhörer und Constantin Gonzales, von dem ich so schamlos geklaut habe.

Ps: Ihr könnt mir gern schreiben was ich alles falsch gemacht hab, bitte!

simply cause no one needs them…

svcadm disable wbem cde-calendar-manager cde-ttdbserver basicreg cde-printinfo webconsole:console ppd-cache-update:default name-service-cache:default