mirabilos's questions

Host Europe assigned a VM I just rented from them an IP address ending in all zeros (2xxx:xxxx:xxxx:xxxx::/128).

I claimed this as possible interop problems as addresses with the last 64 bits all zeros are normally anycast addresses for the subnet router, and I fear other nodes may treat it specially.

Their support says: “It’s only syntactically a subnet anycast address, but you can use it no problems because we use a /128 subnet size.” (translation mine)

I’m only semi-convinced. Is there anything that allows a node to handle addresses of that kind specially, or can I really just use this like a regular address?

I’ve got virtual machines (with Debian or occasionally *buntu guests, GNU/Linux) running in libvirt/qemu/kvm, and I have discovered the feature to pause/unpause a VM.

After unpausing (virsh resume), the guest clock is, obviously, off. How can I trigger a call to, say, /usr/sbin/rdate when the VM unpauses to do a one-shot update (openntpd will do the rest later)?

I have acpi-support-base installed on the guests so virsh shutdown works cleanly, but neither dmesg nor syslog show any events being triggert from pausing/unpausing.

I’ve found diverging instructions on the ’net about this.

To recap: SATA with the 4-pin Molex (white) power plug is not hot-pluggable, but either the wide connector or the separate (15-pin power and 7-pin data) connectors are.

However, in which order do I plug the cables in? Asrock says to connect first the data cable to the mainboard, then to connect, in this order, the power cable, then the data cable, to the drive. Another hardware guide says to connect the data cable first.

It’s surprising that a definitive answer on this is so hard to find.

I’ve just installed a Windows 2012 server for the first time, and would like it to join our Univention Corporate Server domain, so we have the same LDAP users as on all systems.

The error message (translated – unfortunately, the install CD did not permit switching the language to English) roughly said that the domain was unavailable / not found.

We’re running Univention 3.2 currently.

I’m often rebuilding software packages (mostly Debian) for older versions (of both Debian and Ubuntu, including oldoldstable). Some of these packages use the lsb_release command to figure out the current distribution, in order to change behaviour, Build-Depends, etc.

I’ve got a multiple-distribution and CPU architecture cowbuilder setup, but to prepare the source packages (which are then built in a clean environment that matches the target distribution), I still need to do something like:

dpkg-source -x openjdk-7_7u55-2.4.7-1~deb7u1.dsc
cd openjdk-7-7u55-2.4.7
dch --bpo
dpkg-buildpackage -S

The dpkg-buildpackage -S step creates a source package (*.dsc and assorted files) after running the clean target (which regenerates debian/control in many packages). But for this, the lsb_release output needs to be, for instance, squeeze or even lenny.

From my work on Debian/m68k I know I can create an /etc/lsb-release file containing magic, underdocumented lines to control the output of lsb_release, which otherwise looks into, for example, the APT sources.list file, preferences, etc. to determine the distribution. But there is no easy way to make the output mirror another existing distribution.

Has anyone a collection of such files, to mirror the target distribution?

I’ve just freshly installed virt-manager and libvirt-bin (and related packages) on a Debian sid system, and have a strange problem I’ve not yet seen on another similar installation:

When creating a new VM, I get presented one with an IDE disc and an RTL8139 NIC. I can change the type of the NIC to virtio, no problem, but I cannot add a virtio disc. All I get offered are IDE disc/CD, Floppy, SCSI disc, USB disc.

How to fix this? Am I unexpectedly missing any supplementary packages?

Clarification: I’ve got a disc image which I can add to the VM, no problem, just I’d like to add it as virtio disc, not as IDE disc. No filesystem permission problems or anything.