Am 11.10.2007 schrieb Marco d'Itri: > On Oct 11, Anders Henke <[EMAIL PROTECTED]> wrote: > > > This link is automatically regenerated within the postinstall script: > It's not. > > > E.g. after changing a box due to a failed on-board SCSI HBA, > > the kernel (!) spews out warnings like this: > Surprise: you have a buggy kernel.
The Problem occured both using kernel-image-2.6-em64t-p4 (2.6.18+6etch1) as well as an own built 2.6.22.6 and 2.6.23-rc8-mm2. Well, in this case I think that first udev "freed" eth0/eth1 for the reserved MAC-adresses by renaming eth0/eth1 to eth2/eth3, while write_net_rules afterwards created new entries for the freshly-created eth2/eth3. The variable "NAME" isn't yet set, so persistent-net-generator.rules first tried to create persistent entries for eth2/eth3 and then tried to rename the devices e.g. from eth2 to eth2 (which made the kernel warn). But that's just an idea. > > However, there are clearly situations where the "old" behaviour is much > > more appreciated, and I've given a few examples for this. > And are a small-enough number that persistent interface names are going > to stay. Fine. My idea was to simply make this optional or at least document it, as there are known problems with this behaviour. However, it's your decision and I've filed this only as a "wishlist" bug, so you're free to close this issue. Anders -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
