Hello everyone,
I'm running into to weirdness syncing repositories that I'm looking for direction / advice on. I'm currently using Katello 3.1 and TFM 1.12. All of my repositories are set to "On Demand" for their sync type. I have the following repos on my server:
CentOS 7 os/updates/scl/extras
OEL 7 latest and 7.2 "base"
(so there are a lot of packages with the same NVRE but different checksums)
The OEL "latest" repo contains all OEL 7 packages (not just updates).
The OEL "base" repo contains only the 7.2 release packages.
I have the OEL "base" repo syncing from a local loop-mounted ISO, and the repo url is file:///path/to/iso. The reason for syncing this and the "latest" repo is because Oracle's public yum repos are not kickstart trees, and I need that for provisioning.
The C7 and OEL7 latest repos are behaving as expected, but the OEL7 "base" repo says "0 packages" after syncing. However, if I look in /var/lib/pulp/published/yum/http/repos/MyOrg/Library/custom/oel/o7base I see all of the symlinks - albeit broken (which I think is correct because of the "On Demand" setting).
If I attempt to kickstart a host using the "base" repo/installation media and the "Kickstart default" provisioning template the install starts (url --url appears to be set correctly) but when Anaconda goes to actually install the packages it seems to fail to download them. I can't find any better error message than:
The following software marked for installation has errors. This is likely caused by an error with your installation source.
…and then it lists all of the packages.
If I point a yum client directly at the repo it appears to work fine.
If I browse to an RPM in a web browser I'm redirected to a huge URL such as:
http://katello/streamer/var/lib/pulp/content/units/rpm/cc/8145fbdf9be82f7fe0fdddd59baa9928848f385bdc30312d15128032c639ea/glib2-2.42.2-5.el7.x86_64.rpm?policy=eyJleHRlbnNpb25zIjogeyJyZW1vdGVfaXAiOiAiMTcyLjE2LjI0Ni4yNDIifSwgInJlc291cmNlIjogIi9zdHJlYW1lci92YXIvbGliL3B1bHAvY29udGVudC91bml0cy9ycG0vY2MvODE0NWZiZGY5YmU4MmY3ZmUwZmRkZGQ1OWJhYTk5Mjg4NDhmMzg1YmRjMzAzMTJkMTUxMjgwMzJjNjM5ZWEvZ2xpYjItMi40Mi4yLTUuZWw3Lng4Nl82NC5ycG0iLCAiZXhwaXJhdGlvbiI6IDE0NzQ0MDY3NjJ9;signature=eBFhBnj5ho-RAndCApPJfMGkvsfesjoDBOnjCLk8F7awaD1wE5Edv86j6tvtxXHhIAAXVhdonEGIHZJlDjOOVEMpu4-XF9I6pGPXa-Ez4NNN56_ud615QeC71kyLtkp0TeAZ8y_fZrQ-BoIWVmLyNlfLY5lkZbaF-q7hZasdCKDKsGqBbJIdeg2v11QKW2ivo9RwMGha3OMOyNYUYi0qCw6HjEs37GpWc9Jf_X8W0B5hDPVtReLpDAC2eSTCTxAoYvtP74wBCh0csxFNYAreK-bI2TR2H64MaJjNPHwmIs46YNFOrxDRCzxRUwUf7_9_q7Hsg4wdUBf81cShRKby4w%3D%3D
…and the file downloads. The interesting thing is that that file doesn't exist on disk (/var/lib/pulp/content/units/rpm/cc/8145fbdf9be82f7fe0fdddd59baa9928848f385bdc30312d15128032c639ea/glib2-2.42.2-5.el7.x86_64.rpm in this case).
I'm really at a loss here and I'm sure that part of it is just a lack of understanding of the innards of Pulp. Does anyone have any ideas / suggestions / insight?
Thanks,
j
···
----- Original Message -----
From: "Jason B. Nance"
To: foreman-users@googlegroups.com
Sent: Friday, September 16, 2016 4:02:51 PM
Subject: Re: [foreman-users] Adding tftp / pxe target
It has become apparent to me that the bigger issue here is that the Yum repos I’m syncing for OEL from Oracle aren’t kickstart trees, so they are missing files such as squashfs.img. My first thought after that realization was to just slap the missing files into /var/lib/pulp/…, but that felt like a recipe for future problems. Now the path I’m headed down is to mount the installation DVD and sync it.
If anyone has had to deal with OEL or just has another idea in general I’m open.
Big thanks to stbenjam in IRC for the info and ideas.
j
----- Original Message -----
From: “‘Jason B. Nance’ via Foreman users” foreman-users@googlegroups.com
To: foreman-users@googlegroups.com
Sent: Friday, September 16, 2016 2:30:53 PM
Subject: [foreman-users] Adding tftp / pxe target
Hello,
After I installed and configured (all via the web UI) Katello 3.1 / Foreman 1.12 and synced CentOS 7 repositories it seems that all of the tftp/pxe stuff necessary to do pxe installs of CentOS was automagically setup for me.
For example, /var/lib/tftpboot/boot contains:
CentOS-7-x86_64-initrd.img
CentOS-7-x86_64-vmlinuz
…and /var/lib/pulp/published/yum/http/repos/My_Org/Library/custom/centos/c7/LiveOS/squashfs.img exists.
I’m trying to add pxe boot/install support for OEL 7. I created similar products, content views, media, and OS as I have for CentOS. I used host discovery to create a new host entry and selected the OEL stuff I setup. /var/lib/tftpboot/boot/OracleLinux-7-x86_64-initrd.img and OracleLinux-7-x86_64-vmlinuz where created for me, but they were 0 byte files. So I manually replaced those files with the appropriate ones from the OEL “boot” CD.
However, the installer now fails because /var/lib/pulp/published/yum/http/repos/My_Org/Library/custom/oel/o7/LiveOS/squashfs.img doesn’t exist.
What is supposed to create those files? Is there a hammer command I should run or something?
Thanks,
j
–
You received this message because you are subscribed to the Google Groups “Foreman users” group.
To unsubscribe from this group and stop receiving emails from it, send an email to foreman-users+unsubscribe@googlegroups.com.
To post to this group, send email to foreman-users@googlegroups.com.
Visit this group at https://groups.google.com/group/foreman-users.
For more options, visit https://groups.google.com/d/optout.
