Hi, (this sounds indeed a bit weird, but I have a think for weirdos and this sounds strangely interesting as well, so lets talk…) On Fri, Apr 10, 2015 at 03:47:14PM -0700, Corbin Champion wrote: > are being done on these files to make sure they are as expected? Is there We check filesize and the hashsums of the file content at various stages to ensure we are dealing only with good valid files nobody has tempered with. You prominently mention /etc/timezone. We aren't actually using this file at all. I presume libc6 does, which leads me to saying that e.g. http assumes it can set the modification times of files it deals with – and read them in later runs. > anyway to get more verbose information about this failure from apt-get? Lets just add some debug options and hope for the best as I have no particular idea what could be wrong here just yet. -o Debug::pkgAcquire::Worker=true ^ the communication between the acquire system and the individual workers like 'http' and 'xz' doing the hard work. Its a text interface, so if its a worker acting strangely you can run them standalone with relative ease. -o Debug::Acquire::http=true ^ (no, this isn't a typo) insights into the inner working of the worker. In this case http. Some others have this as well, but might not be as talkative as http is. It will mostly hashout requests and responses it send/got from the server it is talking to. I would guess you are looking for an 'URI Failure' message in the output of the former. apt-get has the tendency to fallback from xz to bzip to gzip to uncompressed on failure and the publicly only announces the uncompressed failure while the rest is more or less silently ignored. Best regards David Kalnischkies
Attachment:
signature.asc
Description: Digital signature