Patrick Doyle wrote:
For the last few days or so, when I've attempted to perform a "yum
update", the process has consistently failed. The most common error
is "[Error -1] Header is not complete", which is what I get after I
download all 100% of an RPM from an HTTP site. When yum retries from
another mirror and the mirror happens to be an FTP site, I get other
errors (such as 500 RE*T not understood" or "500 Unknown command").
Poking around, I've seen an explanation for the "Header is not
complete" error that indicated that there is a problem with a proxy
server between me an the outside world. While I won't rule that out,
I can say that I can use wget to fetch the rpm manually -- from a
cursory look, the RPM seems fine to me.
Any suggestions?
Very busy mirrors ?
Aborted mirroring process ?
Initial mirror selected had old or bad metadata, and hence every correct
download is rejected ?
I have seen timeouts from extras repos, and even dir lists that take
more than 10 minutes to complete.
Packet trace the whole yum process: you can see if the client or server
stops doing the right thing.
yum -d 15 and post the "whole thing", maybe as attach or on a web site
if it is very big.
Yes, I've tried
$ sudo yum clean metadata
and
$ sudo yum clean all
Strangely, however
$ sudo yum clean cache
Returns an error:
Error: invalid clean argument: 'cache'
despite the fact that the man page claims "yum clean cache" should
work (and has worked in the past with older versions of yum).
- clean cache has been split up into smaller parts:
clean metadata {ie downloaded repodata}
clean dbcache {sqlite db cache}
If I were to fetch an RPM manually, could I verify that it is correct
somehow and then place it someplace at which yum would be happy with
it?
rpm --checksig my-dud-downloaded-package.rpm ?
DaveT.
