I've been spinning my head for 2 hours now trying to understand how yum goes through a repo. I'm getting a very strange behavior.
- Local epel-7 CentOS repo on repo server
- The repo was created using
reposync(reposync, createrepo, repomanage)
- I can see the repo fine using
- I can also see the files using a shell on repo server
- The client is configured and has network access to repo
Epel-7 repository structure is a little different than Epel-6 structure. It looks like this:
tree -d . ├── 0 ├── 2 ├── 3 ├── a ├── b ├── c ├── d ├── e ├── f ├── g ├── h ├── i ├── j ├── k ├── l ├── m ├── n ├── o ├── p ├── q ├── r ├── repodata ├── s ├── t ├── u ├── v ├── w ├── x ├── y └── z 30 directories
With all the packages under their respective letters/numbers
I'm saying this because the heart of my problem is that when I do a:
yum --disablerepo="*" --enablerepo="epel-7" list available
I get all the packages except the
l directory and everything under it.
I've tried clearing everything:
yum clean all
I've tried deleting the repodata on the repo server and recreated it but nothing works.
That is why I am trying to understand the internals of the
yum search command. What files/data bases does it call, so I can find which one is corrupted/broken?
Could it be cache on the http server side?