[sorry for late reply] On Mon, Jan 15, 2007 at 12:41:51AM -0200, Henrique de Moraes Holschuh wrote: > > > > problem is, that fixes need to survive upgrades. > > It survives upgrades just fine, or at least it is supposed to by design.
hope so - we'll see on next round. > Now, how the heck you had a disk-full upgrade that did not abort is ... > We need to double-check if we are handling all error paths fine in all the ... > Are you sure you didn't have a disk corruption and a subsequent fsck at boot > (the files with size zero are in the root partition, which gets fixed > automatically) didn't truncate them while trying to repair? he, that's been a while ago, and since netinstall many upgrades have occurred. I've mix-used both aptitude, dselect, plain apt-get, sometimes had to do 'install' more than once but always landed to EOprocedure without forcing anything special, as far as I remember. Can't tell for sure if there has been a full fsck on ext3 error on any reboot - it's a notebook used by somebody else and I'm doing stuff mostly remotely. An hard fsck on boot would likely have been reported to me, though. I can tell for sure the disk has no bad sectors. Also, I saw misbehaviour of apt-* which I reported. At least last couple upgrades went smoothly. Installer was: DISTRIB_RELEASE="3.1 (installer build 20060904)" X_INSTALLATION_MEDIUM=cdrom and I guess the disk full was on a massive upgrade which involved also OOo2. I've used same netinstall build on another notebook, using 1st one as local .deb and did not have same problem - didn't have diskfull though, using lvm2 there. Indeed I dbl-checked that one for the rcS.5 issue. thanks -- paolo -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
