General

Okay, just got an even higer uptime result. A reader sent me a screenshot of one of his older systems: 3720 days, i assume this system was rebooted because of a Y2K procedure because 3720 days ago we had the 1.1.2000.

The October 2001 is a little bit strange from that perspective. So i'm not sure about the real uptime, perhaps the system booted with the wrong clock.

The developer build 133 of OpenSolaris is now available for installation from the repositories. So start your package updater or - when you want the ISOs - download them at the usual location.

Virtualbox 3.1.4 was released a few days ago. You can download it at the usual location. As i already stated in the articles regarding the beta versions, it isn't a feature release, but a maintainance release.

IPMP vs. Link aggregation

Link aggregation is available on many switches for quite a while now. With link aggregation it is possible to bundle a number of interfaces into a single logical interface.

"Nichts ist so beständig wie die ständige Veränderung", "Planung heisst den Zufall durch den Irrtum zu ersetzen" ... es gibt eine Vielzahl ähnlicher Grundsätze im Leben. Und genau das ist jetzt auch bei mir eingetreten: Mein Umzug nach Bremen findet nicht statt. Die Zeit zwischen den Jahren hat Gründe für Bremen/Oldenburg verschwinden lassen und somit deutet wieder alles auf Hamburg. Also: Ich bleibe Hamburger.

Genauso lieb sind mir ja jene Menschen, die einen Wagon betreten, uns sich beim Anblick der "bahn.comfort"-Zeichen erstmal darüber echauffieren, es gäbe Menschen 1. und 2. Klasse in der 2. Klasse. Nein, Verehrteste, es gibt nur Menschen die der Bahn viel oder wenig Geld für Transportleistungen in den Rachen schmeissen. Und damit die Vielschmeisser weiterhin werfen ... ja ... dafür gibt es diese Schilder.

There is the cool command line option for the transaction group rollback feature. This feature was integrated at the end of October. It's useful when you are a victim of lying hardware. For example, a hardware lying about the CACHE FLUSH or about the sequence it uses to write blocks to disks.

But what do you do if you ran into such a zpool corruption and don't have an recent developer build of OpenSolaris in use or you use Solaris 10? Well, Victor Latushkin described in a mail at zfs-discuss a nice solution. When your hardware killed your ZFS, just use a LiveCD with OpenSolaris 128 or better, import the pool as described by Victor, export it and you are done. Now you should be able to use the pool again (given the disk was haunted by a failure mode that is solvable by a transaction group rollback).

PS: Don't even think about using the zfs upgrade command while using the LiveCD. Don't do this or you can upgrade you system to the developer build.

PS2: Recent builds of the LiveCDs are available at genunix or as a Torrent at dlc.sun.com.